The Summary

UK Healthcare: An Opportunity to Inspire a Nation and Beyond — a living testament to the lives harmed, and to the staff who dedicate their lives to acting for the good of others. It tells a story in three parts.

The NHS is a story of humanity and of extraordinary achievement. Somewhere along the way it lost that identity, at a terrible human cost.

And it is fixable, largely without great sums of money, because the places that get this right almost always have humanity at their core. Ten volumes plus sources, eighteen annexes and a glossary; this one is the whole case, and the rest is the evidence.

The story in three parts

1. The NHS is a story of humanity, and one of the great achievements of the last century.

It began with a son holding his dying father: in February 1925, just over a century ago, Aneurin Bevan watched his father die of miners' lung disease, uncompensated because the law did not count it as an industrial disease.

He grew up in Tredegar, where since 1890 the town's workers had paid a few pence from every pound into a medical aid society so that no neighbour faced illness alone — by the 1920s it covered some 20,000 people.

In 1948, in a country close to bankrupt after the war, he scaled that town's answer up to a nation.

So the idea at the heart of the NHS is not seventy-eight years old but nearly 300: Tredegar's society dates from 1890, and its roots go back to the Wales of the 1730s.

It was not invented by a government: a community of working families proved it first, and went on to shape how a whole nation would be cared for.

Bevan is reported to have put it simply: "All I am doing is extending to the entire population of Britain the benefits we had in Tredegar for a generation or more. We are going to 'Tredegarise' you."

He stood in a long tradition: in 1840 Elizabeth Fry's nurses were already attending the poor "with the same kindness and care bestowed upon the more wealthy". Annex 4 · The founding vision of 1948

On 5 July 1948 the fear of the doctor's bill was lifted from everyone at once — "It was just unbelievable," one radiographer remembered — and in its first year the service supplied some four and a half million pairs of spectacles to people who had lived for years unable to see properly.

Since then it has cut infant mortality to less than a third of its 1980 level, delivered the world's first IVF baby and the world's first Covid vaccination outside a trial, and it is still kept going by some 1.5 million people in England who go to work every day to look after strangers.

2. But somewhere, the NHS and the wider care system lost that identity — and Chester shows how far.

In January 2009 the NHS wrote its founding values into a Constitution: compassion, respect and dignity, everyone counts. Six years later the opposite happened.

Between June 2015 and June 2016 a nurse on the neonatal unit at the Countess of Chester murdered seven babies and tried to murder seven more; when consultants raised the alarm, the hospital's leaders treated them as the problem.

The Thirlwall Inquiry, reporting on 15 September 2026, found "a profound failure of management, governance and safeguarding" and recorded "a total failure of safeguarding at every level".

The same pattern runs through twenty-five years of inquiries, through Durham and Nottingham, and through a social and children's care system where the most vulnerable meet the least oversight. The values were written down. They were never built in. Annex 1 · Organisational failure: the worked cases · Annex 5 · The care emergency: social and children's care

3. It is fixable, and the fix is not mainly money.

Bevan built the NHS when Britain could least afford it; what he built it on was a way of treating people.

The best-rated care in England today comes from its least-funded part — a quarter of hospices rated outstanding against about one in sixteen NHS acute hospitals.

Leeds and Manchester took their children's services from Inadequate to Outstanding under the same funding pressure as every other council, by putting relationships back at the centre.

Attentive care halves delirium in frail older patients; person-centred care in care homes costs less than usual care.

And money without humanity fails: in a private London hospital where money was no object, an 80-year-old woman died of a clot that routine care would likely have prevented. Capacity matters, and England is short of it. But the thing every bright spot shares is not money. It is humanity, built in.

History offers both a promise and a warning.

The promise: from 1731, when Griffith Jones's schools began teaching the Welsh poor to read, ordinary people in Wales chose to act for the good of others, drew others in and built what lasted — chapels they governed themselves, libraries, a medical aid society, and finally the NHS.

The warning: whenever the powerful have decided about people without them — from the commissioners who judged Wales in 1847 to the long-stay hospitals the NHS inherited — the result has been harm.

That spirit is weaker now, but it has not gone — and it can be regained. Annex 10 · Why we all love the Welsh · Annex 11 · Why a top-down model fails

Research synthesis

What. Patients are being harmed and killed by the way hospitals are run — not by the limits of medicine. Fourteen babies murdered or attacked on one neonatal unit. Over 300 patients harmed in one breast surgery service.

More than 2,500 families in one maternity service. Almost nothing in the inquiry record is the limit of medical knowledge being reached.

Nor is the one case with a deliberate perpetrator an exception: care is at bottom protection, and what failed at Chester was protection — for a year after the alarm had been raised.

And the inquiry record covers the best-watched part of the system — the rest of it tells you two more things.

In adult social care, which employs more people than the NHS hospital workforce, England has no professional register for frontline staff at all — the only UK nation without one — and 83.5% of community services have no current CQC rating, more than a third never assessed. The failures there are not fewer; they are less visible.

Which makes healthcare the jam in the sandwich.

When social care, children's care and prevention fail, people end up in hospital — and hospitals then cannot discharge them.

There are no quick fixes. But there can be no serious review of the NHS that does not take social care, children's care and prevention with it.

Meanwhile the best-rated care in the country comes from its least-funded part: when the CQC last compared sectors, in October 2017, 25% of hospices were rated outstanding against roughly 6% of NHS acute hospitals and 2% of care services — charities run by local trustees, working alongside 89,000 volunteers, measured on how a person's last weeks go. Neither fact is about money.

And none of it is new: the same diagnosis has been written, in the same words, for twenty-five years. Bristol set out the pattern in 2001; Francis restated it in 2013, Kirkup in 2015, Gosport in 2018, East Kent and Shrewsbury in 2022.

Thirlwall, on 15 September 2026, found "a profound failure of management, governance and safeguarding" — and reported that of three decades of inquiry recommendations most "have not been implemented", with still "no mechanism to track progress or enforce implementation": "It is inexcusable to ignore such failings."

What is new is the rate of reporting, not the diagnosis. Eight major reports landed in twenty-nine months, between April 2024 and September 2026, and each describes a failure that had been running for years before anyone reported it.

Chester's babies were attacked in 2015 and 2016, and the report came eleven years after the first death.

County Durham's breast surgery service lost its GMC trainees in 2012 and was still being warned in 2025; of 514 cases reviewed so far, 315 patients were harmed. Ockenden's Nottingham review covered care from 2012 to 2025.

Infected blood reaches back to the 1970s, with four decades of institutional denial after it. The reports are recent. Almost nothing in them is.

The mechanism is identical every time, and it is human rather than technical. Concerns raised, reframed as conduct, candour delayed, regulators shown a curated picture, and the person who spoke managed out.

Chester's own note from May 2017 lists it as a plan: "1. GMC 2. Actions from grievance 3. Mitigation from SoS/whistleblowing 4. Action plan to manage out".

And when the internal mechanism was exhausted, the external one failed as well.

Chester's answer to concerns it could no longer dismiss was to buy a review from a royal college — peer review sold for a fee, "not a regulatory process or a forensic investigation" — which could not answer the one question that mattered, and whose reassurance was then used to close the question down.

Every system of assurance rests on someone from outside being trusted to call it as it is. When that person gets it wrong, everyone downstream inherits a false reassurance and acts on it. Annex 16 · When the sheriff gets it wrong

What corrected it, in the end, was people holding the line at a cost to themselves.

Seven consultants said the same thing for eleven months, through an apology they were required to make to the person they had reported and a written plan to "manage out" the doctors who raised the alarm; an NHS England official who was in the room remembered "the clinicians trembling", and that they "felt like they were putting their careers on the line, but they were prepared to do this given the seriousness of the matter".

The same is true of every case here: the truth arrived because a family would not stop asking, or a clinician would not stop being a nuisance.

That is the thing this document is ultimately asking to be built — and the people who do it get targeted, worn down and sidelined, which is why champions need champions. Annex 17 · Holding the line

And it did not stop once the organisation was under scrutiny: in 2022 four of Chester's own senior leaders ran a covert campaign, "Project Countess", to remove the chief executive who had exposed the earlier failure — a tribunal found hundreds of documents deleted, the chair resigned within days, and it cost the Trust about £3m ($4.05m).

Durham owed candour in May 2024 and began it in February 2025.

In every case the truth emerged because a family would not stop asking — they were the warning bell, which is precisely what the designed mechanisms were supposed to be.

And the same facts describe an opportunity — which is the other half of this document, and the more important half.

The NHS was founded on a moral proposition and still carries one in writing; what it has never had is that proposition built into how the place actually runs.

Build it — humanity engineered into appraisal, appointment, measurement and training rather than hoped for on the day — and the prize is specific: the clinical performance currently lost to how people treat each other, recovered; a negligence bill of which a quarter is legal costs, reduced; staff who stay; families who no longer have to campaign for the truth; and a standard of care no developed system has yet managed.

That is what is actually on the table — not a repaired reputation, but the first health service deliberately designed to stay humane under pressure, built on the founding vision of 1948 and fit for the hundred years after it.

So what. This is negligence, not misfortune. The failures are organisational rather than medical, which means they sat within the gift of boards and managers to prevent.

The diagnosis has been settled since 2001 and the remedies are published, costed and in several places already proven.

To know the problem, to know the remedy, and to act on neither is the ordinary definition of negligence — and at this scale it is negligence against the health and wellbeing of a whole nation.

And the conditions that produce it are still fully in place.

The service runs at over 90% bed occupancy with the fewest scanners in the G7, which removes every margin that would otherwise absorb a lapse; 30,323 nurses left the register in the year to March 2026; and Nottingham is removing around 1,279 posts while implementing Ockenden and asking staff to speak up more freely.

Psychological safety is easiest to build in calm conditions and needed most in stressed ones.

The option being paid for is the most expensive one available.

The NHS carries a £60.3bn ($81.4bn) provision for harm already caused — second only to nuclear decommissioning on the government balance sheet — pays out £3.2bn ($4.3bn) a year with a quarter of it going to legal costs, and loses an estimated £12.1bn ($16.3bn) a year to staff ill-health and cover.

Thirlwall alone has cost £18.6m ($25.1m). Not one pound of that buys any patient any care.

The cost does not stay inside the NHS. Councils' care overspends now squeeze the prevention that would have reduced them, as Annex 5 · The care emergency: social and children's care shows in Lincolnshire and Teesside. Annex 3 · The impact: when people stop caring sets the human, organisational and financial costs side by side.

Meanwhile the thing that would help is the cheapest input in the building, and it is measurable.

Incivility costs its target about a third of their performance and bystanders about a quarter; in a simulated operative crisis 91% of clinicians passed in normal conditions and 64% when exposed to rudeness; 27% of US practitioners knew of a medication error in the previous year in which disrespectful behaviour played a part.

Civility is not a mood — it restores diagnostic accuracy, information-sharing and help-seeking, which in a hospital are the safety system.

And it is not evenly distributed, which is the finding the service has been slowest to act on.

Lord Mann's review, published on 4 June 2026 with 24 recommendations, found racism "normalised in the NHS" — a judgement about the ordinary working environment, not about isolated incidents.

The Staff Survey points the same way: 8.77% of 766,285 respondents reported discrimination from colleagues or managers in 2025, the lowest figure ever recorded and still roughly one member of staff in eleven.

Australia, which measures this openly, records 20% of doctors in training experiencing bullying, harassment, discrimination or racism, rising to 56% among Aboriginal and Torres Strait Islander trainees.

The mechanism is the one above, applied unequally: rudeness degrades a team's clinical performance, and it degrades it most for the people who absorb most of it — the Canadian data find it aimed disproportionately at those who are younger, more junior, female, non-native speakers, or from a non-dominant ethnic group. Volume 6 takes this further, because it is the reason the argument is addressed to the medical profession rather than to managers.

Now what. Fix the foundations — because nothing currently on offer touches them. Everything being done works on the building above ground: detect the failure sooner, punish it harder, count whether the recommendations were implemented.

The foundation is what produced the failure in the first place — organisations rewarded for finance and throughput, run with no capacity margin, staffed by people with nothing holding their humanity in place when the pressure arrives.

Hold people to account by all means, but accountability is not a cure — and the system cannot lay its own foundations. Barring a chief executive removes a person, not the conditions that made his conduct rational.

The wounded rarely recover unaided and the self-harmed almost never do; twenty-five years of internally generated reform is the evidence. As Rabbi Lord Jonathan Sacks put it, drawing on the Talmud, "a prisoner cannot release himself from prison".

And the deeper reason accountability cannot carry this is that you cannot regulate people to care.

It has to come from within, and to last it has to be hard-wired into the foundations. This is about identity: own it, model it, live it — make it true.

When you lose your humanity, you lose everything — and that is most true under pressure, which for the NHS is the permanent condition.

So the capability has to be imported — and everything that works came from outside healthcare: Toyota, policing, a district nurse who rejected management, and institutions built by outsiders.

Importing is not the same as copying. England imported Buurtzorg's self-managing teams twice and Sweden's Esther question once, and each faded because the form crossed and the function did not.

What travels is the principle, redesigned for the place by the people who have to run it — and only when whatever sits above the team changes too. Annex 15 · Humanity in healthcare: examples

But the idea being imported is not foreign to the NHS. It is the NHS's own. Bevan's case in 1948 was moral before it was administrative, and the NHS Constitution still carries compassion and respect and dignity among its six values.

What Bevan built into structure — free at the point of need, funded collectively — has survived seventy-eight years. What he left to chance is the half that appears in every inquiry report.

Humanity-first healthcare completes the founding settlement rather than replacing it.

And Britain is the country best placed in the world to build it.

No developed system has solved this, and Britain holds four things no one else holds together: the failures documented to a standard no other country can match, the values already written into a constitution, the measurement to see whether anything changes (766,285 people surveyed a year), and one national system in which a change can land everywhere.

A model built and proven here would travel — which is what this country has historically done with ideas forged in adversity.

The one-sentence version. Diagnosed, not treated: England has built an unmatched capacity to investigate its own healthcare failures and has not yet built the capacity to prevent them — and humanity under pressure is the cheapest source of clinical performance it is not yet buying.

Fixing the foundations: humanity-first healthcare

So here is the missing foundation, and it is buildable. Humanity-first healthcare is the deliberate building of the capability that fails first under pressure. Not a values statement, not a training day, not another framework.

The capability is specific: the capacity of a person under acute stress to regulate their own state, see past their own distress, and still act for the good of someone else — and the capacity of a team to make that the normal thing to do, including when the difficult thing to say is upward.

This is the unfinished half of Bevan's settlement, and the country has changed in the one way that matters.

When Bevan built the NHS, humanity did not need building in. Britain had just come through two world wars: people had queued, rationed, taken in evacuees and pulled neighbours out of the rubble.

What the country lacked was capability — hospitals, doctors, medicines, reach — and that is the half he wrote into law.

Seventy-eight years on we have the capability, and the thing he could take for granted has thinned. One marker will do: the share of young people with no close friends has risen from about one in 100 in 2014 to one in 20 today.

Every failure in this document ran through people who lost that capacity in organisations that rewarded something else.

It is trainable, on Staub's evidence that "learning by doing can change people in either positive or negative directions", and it is a performance intervention: civility restores diagnostic accuracy, information-sharing and help-seeking, which in a hospital are the safety system.

The order matters — identity first, then values, then structure: who we have decided to be, what that commits us to, and the arrangements that hold us to it on a bad night.

In one line. The foundation is the capacity to act for the good of others; the goal is that the patient will be at the heart of everything; and three pillars stand between them — every person counts, caring for the carers, and it's who we are.

At its best, humanity becomes instinctive: the last of what Champions calls its Four Stages of Humanity, "unconscious humanity", where being human under pressure is simply another day at the office. The executive summary and Volume 10 set them out.

There is a second three, and it is how the first three get built rather than what they stand for. The individual, trained and practised in the capacity to act for the good of others.

The community, in small circles that belong to a network beyond the employer, so that concerns and praise both have a route management cannot close.

And leadership that answers for both: trust, so people tell the truth early; forgiveness, so an honest mistake does not have to be hidden; and being held to account, so forgiveness is not mistaken for indifference.

None of the three needs a reorganisation, new legislation or a heroic appointment, which is the point of them. Bevan wrote the money into an Act of Parliament and it held for seventy-eight years; the humanity was left to chance.

Building these three is what finishing his settlement looks like, and the test of whether they are there is one question: what here would still be true six months after the chief executive left? Annex 12 · Humanity-First Healthcare: applying lessons from the past · Annex 14 · Healthcare: why the Art of Enablership is critical · Annex 13 · Radical leadership

All three exist for one moment: when doing the right thing carries a personal cost — the job, the reference, the training number, the family's security — and every respectable alternative says look away.

Holding the line is a choice, but it is almost never held alone: it takes choice, connection and creation, which is to say it is a team effort. Annex 17 · Holding the line

It is also not a new idea — it has a lineage, and the lineage is the evidence.

Elizabeth Fry founded the Institution of Nursing Sisters in 1840 on what reads as the first humanity-first standard in healthcare: "the poor should be attended in their houses with the same kindness and care bestowed upon the more wealthy part of the community".

She did not leave it as sentiment — she built an institution, secured clinical training at Guy's from 1841, and drove the Gaols Act 1823 through Parliament; the BMJ called her "the founder of nursing" in 1897.

Mayo gave away a profitable practice in 1919 and wrote its values into a salary policy and its bylaws.

The Aga Khan network seeks accreditation no regulator requires, scoring 98.4% in Nairobi in December 2025, "because it is the right thing to do for our patients".

Intermountain — handed fifteen hospitals by the Mormon church in 1975 and secular ever since — took about $50m a year out of the cost of care in Utah while cutting only $41m from its own costs, losing roughly $9m a year in revenue by improving obstetric care, and did it anyway.

Nuka transferred ownership to the people it serves; Buurtzorg made supervision arithmetically impossible; Virginia Mason put a clock on every safety concern and set a target of more alerts.

Three faiths and three secular institutions, spanning three centuries, with no shared funding model — and almost every one of them an outsider to the establishment of its day. What they have in common is not a creed and not a mechanism.

It is a settled identity, an owner whose purpose is something other than margin — which is what lets an organisation absorb a loss to do the right thing — and arrangements built to protect both.

That is why England's abundance of mechanisms without that identity underneath produces compliance instead of care.

And the early evidence that it moves quickly is encouraging. Champions ran its first healthcare programme at Union Health Service in Chicago — six sites, around 250 staff, 40,000 members — from June 2025.

The starting point was not a skills gap: chronic emotional exhaustion, collapsed trust between teams, a 1.6 Yelp rating, and a workforce the pack described as "one setback away from catastrophic failure: not from clinical skill, but from emotional depletion".

After 90 days, an independent evaluation found every interviewee reporting greater emotional awareness and self-regulation, around 90% improved cross-team collaboration, 85% reduced emotional reactivity and 70% positive change in patient interaction — with behavioural detail more persuasive than the percentages: "People stop now, they don't just walk past"; "before I'd ask, 'What's your problem?' Now I stop and wonder what they're going through."

Its chief executive: "I've done leadership training before, but this felt fundamentally different."

The honest bound, which the evaluation states itself: one non-randomised outpatient case study at 90 days, directional rather than statistical. Read it as indicative — and note that it points the same way as the wider literature, in a real organisation, fast. Volume 10 has the full case.

And the opportunity, which is larger than the problem

Read this document as an indictment and you will miss the bigger half of it. Britain is, on the evidence, the country best placed in the world to build a humanity-first model of care — and it would not be the first time it built something everyone else adopted.

The heritage is not decoration; it is the reason to believe this is possible here. Parliamentary democracy and the rule of law.

The Quaker movement's demonstration that enterprise could be built on trust, plain dealing and equality of persons — by people excluded from the universities and the professions, which is exactly why they had to build their own institutions.

The industrial revolution. Vaccination (Jenner, 1796) and epidemiology (John Snow, 1854). Modern nursing — Elizabeth Fry's Institution of Nursing Sisters in 1840, then Nightingale.

Penicillin, the jet engine, the structure of DNA, the first IVF birth, the World Wide Web. The NHS itself in 1948, copied in principle across the world. The Oxford–AstraZeneca vaccine, supplied at cost.

Even the object in most people's pocket: the iPhone's industrial design came from a Londoner, Jony Ive.

The pattern in that list is that British ideas have usually arrived from adversity and from outsiders — a bankrupt post-war country, a cholera epidemic, a wartime workshop, a dissenting minority shut out of the establishment. This moment has both of those conditions in abundance.

And four assets sit here that no other health system has together. The failures are documented — eight major inquiries in twenty-nine months, with a forensic record of how organisations behave under pressure that no other country possesses.

The values are already written down — six of them, in a constitution, agreed with patients and staff.

The measurement exists — the NHS Staff Survey asks over 766,000 people every year how they are treated, one of the largest workforce datasets in the world, so a change in culture would actually be visible.

And it is one system — what works can be adopted once and land everywhere, with four nations running as natural experiments alongside it. Most countries attempting this would first have to build the evidence base, the values statement and the measurement.

England already has all three, and has been using them only to describe the problem.

The prize is not only domestic.

Every developed system reports the same incivility, the same burnout and the same silence, and none has solved it; the United States spends four to five times what the UK does on long-term care with roughly four times the staff turnover; and social care worldwide is larger, less regulated and less scrutinised than hospital medicine everywhere.

A model that demonstrably restores humane behaviour under pressure — built and proven in the most heavily documented health system on earth — is exportable in a way that almost nothing else in this field is.

So the honest framing of this moment is two-sided.

A country that wrote the first humanity-first standard for nursing in 1840, and the first universal health service in 1948, is now sitting on the most complete record of institutional failure ever assembled in healthcare.

It can treat that record as a scandal to be managed, or as the raw material for the third thing — a standard for how people are treated inside the system, built into its structures rather than its statements.

That is what is actually on the table, and it is why this document ends in a proposition rather than a complaint.

The whole argument, in four moves and sixteen lines

Four moves, four lines each. The pattern (how harm actually happens) → how it becomes normal (and who is left to notice) → why money and skill are not enough → what actually holds, and what happens next. If you remember nothing else, remember the four.

Move one · The pattern

1. Culture fails before care does. Across eight major inquiries, the proximate cause of catastrophic harm is the same, and it is never technical.

2. The mechanism is always the same. Concerns raised, reframed as conduct, candour delayed, regulators shown a curated picture, the person who spoke managed out.

3. It isn't bad people. Half of those who admit to rudeness say it is because they feel overloaded; 4% because it is enjoyable.

4. But small acts license bigger ones. Rudeness is contagious — people treated badly pass it to third parties who did nothing, and witnesses are primed to read the next borderline act as normal.

Move two · How it becomes normal — and who finally notices

5. So nobody in the room is neutral. Bystanders' own performance drops 22–28%. Silence is a position, and it is information: this is tolerated here.

6. How unacceptable behaviour becomes normalised.

The threshold moves until the behaviour that would have been unthinkable becomes normal, and the person who finally objects is the one who looks unreasonable.

7. Families have been the safety valve, not the governance. In every case the truth emerged because someone bereaved would not stop asking. They were the warning bell — which is precisely what the designed mechanisms were supposed to be.

8. And what they were warning about was organisational failure, not medical mistake. Healthcare is like aviation: when coordination fails it does not harm one patient, it destroys many lives at once, and the families alongside them.

Almost nothing in the inquiry record is the limit of medical knowledge being reached. These were known risks, with known remedies and known routes to raise them, not taken.

That is what makes them unforgivable rather than merely tragic — not that they happened, but that after everything already documented, they were still allowed to happen.

Move three · Why money and skill are not enough

9. Money is not the whole story. In a private hospital where money was no object, a patient died of a clot that routine blood-thinning medication would likely have prevented. Resources are necessary and not sufficient; what fails is human coordination.

10. But capacity removes every margin. A system at over 90% occupancy with the fewest scanners in the G7 has no room for the pause, the second opinion or the investigation. Which leaves three options and only three: add capacity, build the human capability to hold the line under strain, or keep commissioning inquiries.

11. Technical excellence does not produce humanity. 1930s Germany led the world in medical science, and its physicians joined the Nazi party in greater proportion than any other profession.

The danger is not peculiar to a regime — nor is it inseparable from science. It belongs to any human system not rooted in values of humanity; medicine's power only raises the stakes.

12. Pressure does not erode values — it exposes them. Where nothing structural is holding them up, the wall breaches and what follows is a flood of behaviours nobody intended. Every institution in the inquiry record was under pressure; the ones that failed had nothing holding the line.

Move four · What actually holds — and what happens next

13. Values hold when they are rooted in identity and expressed in structure — both, in that order. Mayo, the Aga Khan network and Fry each first settled what kind of institution they were, then built arrangements to hold themselves to it under pressure.

England has values statements and a great many mechanisms, and little settled identity underneath either — which is why its mechanisms become compliance.

14. The reforms of 2026 are the first with teeth, and still address the wrong half.

Detection and punishment, not the conditions that produce the failure or the capacity of people to withstand them.

Accountability is necessary and it is not a cure: no system that produced a failure of this kind has ever diagnosed and repaired itself from the inside.

15. Northern Ireland is the next major inquiry already in the pipeline.

England has demonstrated repeatedly — Mid Staffordshire, Nottingham, Durham — that running a health service on financial and operational targets, without being rooted firmly in clinical and human priorities, reliably produces human harm.

Northern Ireland is now being told to do exactly that, in a system already failing its patients, with the thinnest culture data in the UK to notice the consequences. Annex 9 · Joining up health and social care: what integration can and cannot do

16. And humanity has to be at the heart of medicine, starting with the carers. Technical excellence is settled, and on its own it has never been enough.

The open question is whether medicine chooses to be the best of the best at being human — towards its patients and towards its own people — and it already has every tool it needs to do it.

Everything in the volumes that follow is the evidence for those sixteen lines, in that order — ending with what a humanity-first response would involve, and why humanity is a performance multiplier rather than a soft extra.

A preventable catastrophe

To know the problem, to know the remedy, and to act on neither is the ordinary definition of negligence. What follows is that charge, at national scale, with the health and wellbeing of a whole nation as the thing at stake.

Every element of it is already established, and most of it admitted.

Twenty-five years of inquiries have produced one diagnosis, restated with growing precision: harm follows when the people who can see a problem are not heard, and the institution protects itself instead.

The remedies are known, published, costed, and in several places already proven to work. The failures are organisational rather than medical, which means they sat within the gift of boards and managers to prevent.

And they keep happening anyway — Thirlwall, reporting on 15 September 2026, describes the same disease Bristol described in 2001 and Francis described in 2013, and says so in terms.

Fourteen babies at the Countess of Chester. More than 2,500 families at Nottingham. Over 300 patients harmed in a single breast surgery service in County Durham. None of it at the frontier of medical knowledge. All of it preventable.

The shorthand used throughout for that state of affairs is diagnosed, not treated: the analysis is settled and the treatment is not. It recurs deliberately, because the gap between the two is where the harm lives.

English healthcare is clinically capable and institutionally unwell. The failures documented between 2024 and 2026 are not, in the main, failures of medical knowledge or technical skill. They are failures of governance, candour and culture.

The corollary matters as much: English clinicians are good, the system around them is not, and because the binding constraint is human and organisational, so is the remedy.

Three things define the present moment

The volume of inquiry is unprecedented.

Between May 2024 and September 2026 England absorbed the Infected Blood Inquiry, the Cass Review, the Covid-19 Inquiry's healthcare module, the Lord Mann review into NHS racism, the Ockenden review of Nottingham maternity services, the Amos national maternity investigation, the Aubrey review of breast surgery at County Durham and Darlington, and the Thirlwall Inquiry into the Countess of Chester.

Several more are live. This is a step change in scale and in legal force, not the historic baseline.

The findings converge. Different specialties, different decades, different trusts — and the same mechanism each time.

Clinicians raise concerns; managers experience those concerns as reputational threat; the concerns are reframed as a personnel problem; families are told nothing; regulators are shown a curated picture and record the trust as adequate.

Thirlwall found a "profound failure of management, governance and safeguarding" at Chester. Francis found materially the same thing at Mid Staffordshire thirteen years earlier.

Accountability has finally become asymmetric in the other direction. For the first time, senior NHS managers face statutory regulation, a barring scheme, and — at Chester and Durham — active criminal investigation.

Against that, the only sanctions actually imposed on Chester's leadership so far have come from an employment tribunal and a professional body rather than from anywhere inside the NHS.

Whether the new machinery changes behaviour or simply adds another layer of defensive compliance is the central open question of the next three years.

The uncomfortable conclusion running through all of it: the NHS has become extremely good at investigating cultural failure and has not yet become good at preventing it.

How this document is organised, volume by volume and annex by annex, with a line on what each one covers: Contents

Not from the UK — or want the system in numbers? Everything needed to orient yourself sits in its own annex: Annex 2 · UK healthcare in numbers. It opens with ten numbers, a map of the parts and who pays for each, and a timeline of the last seventy-eight years, then covers money, workforce, structure, the health and social care divide, dentistry, hospices, community care and unpaid carers, the US comparison, and what is changing.

The case, in six points

The NHS is a story of humanity and of extraordinary achievement. Somewhere it lost that identity, and patients, families and staff are paying for it.

It is fixable, largely without great sums of money, because the places that get it right almost always have humanity at their core. Six points, each standing on its own numbers, built to be said out loud.

Everything in the volumes that follow is the evidence for these.

1. The NHS is a story of humanity, and one of the great achievements of the last century. It began with a son holding his dying father.

In February 1925 — just over a century ago — Aneurin Bevan watched his father die of miners' lung disease, uncompensated because the law did not count it as an industrial disease.

He grew up in Tredegar, where since 1890 the town's workers had paid into a medical aid society so that no neighbour faced illness alone, and in 1948, in a country close to bankrupt after the war, he scaled that town's answer up to a nation.

The idea at the heart of the NHS is nearly 300 years old, rooted in the valleys of Wales, and a community of working families proved it long before any government did. As Bevan is reported to have said: "We are going to 'Tredegarise' you."

He stood in a long tradition — from 1731 Griffith Jones's schools had taught the Welsh poor to read, and in 1840 Elizabeth Fry's nurses were attending the poor "with the same kindness and care bestowed upon the more wealthy". Annex 10 · Why we all love the Welsh · Annex 4 · The founding vision of 1948

On 5 July 1948 the fear of the doctor's bill lifted from everyone at once: "It was just unbelievable," a radiographer remembered. In its first year the service supplied some four and a half million pairs of spectacles to people who had lived for years unable to see properly.

Since then it has cut infant mortality to less than a third of its 1980 level, delivered the world's first IVF baby and the world's first Covid vaccination outside a trial.

And it is kept going by some 1.5 million people in England who go to work every day to look after strangers.

2. But somewhere it lost that identity — and nothing shows it more starkly than Chester. On 21 January 2009 the NHS wrote its founding values into a Constitution: compassion, respect and dignity, everyone counts. Six years later the opposite happened.

Between June 2015 and June 2016 a nurse on the neonatal unit at the Countess of Chester murdered seven babies and tried to murder seven more (the Inquiry proceeded on those convictions; an application to the Criminal Cases Review Commission is pending).

Consultants raised the alarm in the summer of 2016; the police were not called until May 2017, and a note of the chief executive's plan that month reads: "1. GMC 2. Actions from grievance 3. Mitigation from SoS/whistleblowing 4. Action plan to manage out".

The families learned the truth from the police or at the trial — one mother: "to not inform us at all until somebody is arrested is unforgivable."

The Thirlwall Inquiry, reporting on 15 September 2026, found "a profound failure of management, governance and safeguarding" and recorded "a total failure of safeguarding at every level".

And the failure of governance and leadership continued, with different people: in 2022, with the leaders of 2016 gone and the hospital under national scrutiny, four of its new senior leaders, including the new chair, ran a covert campaign to remove the chief executive brought in after the scandal, once she had raised a whistleblowing complaint.

New leadership, the same behaviour — not only a few bad apples, but a system that reproduces the same failure in whoever runs it.

And none of it is new: Bristol set out the pattern in 2001, Francis in 2013, Kirkup in 2015, Gosport in 2018, East Kent and Shrewsbury in 2022. Thirlwall's verdict on three decades of recommendations is that most "have not been implemented" — "it is inexcusable to ignore such failings".

3. The failures are organisational, not medical — which means the people running the organisations could have prevented them. Eight major reports landed in twenty-nine months.

At County Durham, 315 of 514 patients reviewed so far were harmed — 61% — including 20 unnecessary mastectomies, in a breast surgery service that lost its GMC trainees in 2012 and was still being warned in 2025; through nearly all of that time the regulator rated the trust well-led, and it owed patients candour from May 2024 but began it only in February 2025.

Ockenden's Nottingham review covered more than 2,500 families and care from 2012 to 2025.

Not one of those failures was the limit of medical knowledge being reached, and Chester's deliberate perpetrator is not an exception: care is at bottom protection, and the gap she worked in stayed open for a year after the alarm was raised.

In every case the truth came out because a family would not stop asking. Annex 1 · Organisational failure: the worked cases

And the outside check failed too.

When the consultants' concerns could no longer be dismissed, Chester bought an external review from a royal college: peer review sold for a fee, "not a regulatory process or a forensic investigation", and unable to answer the one question that mattered.

The reassurance it produced was then used to close the question down. Annex 16 · When the sheriff gets it wrong

What stopped it was seven consultants who held the line for eleven months, at a cost to themselves: an apology they were required to make to the person they had reported, a reference to avoiding a GMC referral, and a written plan to "manage out" the doctors who raised the alarm.

The system did not correct itself. It was corrected by people who refused to stop being a nuisance and who paid for it — which is why the people who hold the line need someone holding it for them. Annex 17 · Holding the line

And this is the watched half of the system: in adult social care, which employs more people than the NHS hospital workforce, England has no professional register for frontline staff at all, and 83.5% of community services have no current CQC rating. Annex 5 · The care emergency: social and children's care

4. The behaviour underneath it is everywhere, and it measurably degrades clinical work.

In England, the NHS Staff Survey 2025 — 766,285 respondents — records 17.0% suffering bullying, harassment or abuse from colleagues, 9.1% from managers and 25.3% from patients and the public; the colleague and manager figures are the lowest recorded, and a quarter of staff are still abused at work.

In the United States, 79% of practitioners had personally experienced disrespectful behaviour and 27% knew of a medication error in the past year in which it played a part (ISMP, 2021).

In Australia, across seven tertiary hospitals, 38.8% experienced incivility or bullying at least weekly, and 49.8% said unprofessional behaviour had a moderate or major negative effect on patient care (MJA, 2021).

Pooled international prevalence of bullying among medical residents is 51%.

And the effect on the work is experimentally established, not asserted: rudeness costs its target about a third of their performance and bystanders about a quarter, makes colleagues 74–84% less likely to help, and took clinicians in a simulated operative crisis from a 91% pass rate to 64%.

5. The bill is already one of the largest liabilities the British state carries. The NHS clinical negligence provision stood at £60.3bn ($81.4bn) at 31 March 2026 — second only to nuclear decommissioning on the government balance sheet.

Cash paid in 2025/26 was £3.2bn ($4.3bn), a quarter of it legal costs, and the Government Actuary projects £4.1bn ($5.5bn) a year by 2029-30. On claims of £25,000 ($33,750) or less, legal costs run at 3.7 times the damages.

Alongside it, poor staff health and wellbeing is estimated to cost £12.1bn ($16.3bn) a year, and the Thirlwall Inquiry alone has cost £18.6m ($25.1m) so far.

The cross-sector precedents say where that money ends up: the coal health schemes paid £4.1bn ($5.5bn) in compensation against £2.3bn ($3.1bn) in administration, £1.3bn ($1.8bn) of it to solicitors; Post Office Horizon has paid £1.7bn ($2.3bn) in redress and counting. (Sterling figures throughout carry a US dollar equivalent at $1.35 = £1.) And the cost spills beyond the NHS: councils overspent on adult social care by £715m ($965m) in 2025/26, and those overspends squeeze the prevention, such as youth services, that would have reduced them. Annex 3 · The impact: when people stop caring · Annex 2 · UK healthcare in numbers

6. It is fixable, and the fix is not mainly money: the bright spots almost always have humanity at their core. Bevan built the NHS when Britain could least afford it; what he built it on was a way of treating people.

The best-rated care in England comes from its least-funded part — when the CQC last compared sectors, 25% of hospices were rated outstanding against roughly 6% of NHS acute hospitals.

Leeds and Manchester took their children's services from Inadequate to Outstanding under the same funding pressure as every other council, by keeping the same people around the same children.

Buurtzorg runs community nursing on 8% overheads against a 25% sector average; Intermountain in Utah gave up about $9m a year of revenue to improve obstetric care, and did it anyway. At the bedside, attentive care roughly halves delirium in frail older patients, and person-centred care in care homes costs less than usual care.

And money without humanity fails: in a private London hospital where money was no object, an 80-year-old woman died of a clot that routine care would likely have prevented.

Capacity matters, and England is short of it — but the one thing every bright spot shares is not money. It is humanity, built in. Annex 7 · Why being human is good for health

The whole case in one sentence

The NHS is one of the great stories of humanity. Somewhere it lost that identity, and the cost is counted in babies, in billions of pounds and in families left to fight for the truth.

The problems are known, the remedies are known, and most of the remedies are cheap. To know all of that and act on none of it is the ordinary definition of negligence — here, at the scale of a nation's health.

And one more thing, because it decides what happens next

Holding this system to account will not fix it. The two obvious responses — hold those responsible accountable, and implement what the inquiries have already established — are both necessary and neither is a cure.

Barring a chief executive removes a person, not the conditions that made his conduct rational; an audit of recommendations counts compliance, it does not create capability.

What is missing is capability, and capability can be built: people trained to act for the good of others under pressure, circles that belong to a network outside the employer so that concerns and praise both have somewhere to go, and leadership that can be trusted, can forgive an honest mistake and can still be held to account. Annex 13 · Radical leadership · Annex 14 · Healthcare: why the Art of Enablership is critical

And the reason neither response reaches it is that you cannot regulate people to care.

It has to come from within, and to last it has to be hard-wired into the foundations. This is about identity: own it, model it, live it — make it true.

When you lose your humanity, you lose everything — and that is most true under pressure, which for the NHS is the permanent condition.

And history is clear about the alternative that tempts every troubled system: deciding about people from above, without relationship with them, has failed again and again — from the Blue Books of 1847 to the long-stay hospitals the NHS inherited. Annex 11 · Why a top-down model fails

And the conditions for the next failure are being assembled right now.

Northern Ireland has 504,537 people waiting for a first outpatient appointment, 53.7% of them for more than a year, emergency department performance around 31%, and a health budget baseline falling from £8,898m ($12.0bn) to £8,481m ($11.4bn) — with no staff survey in place to notice the consequences.

Northern Ireland has run health and social care as one system since 1973, which is why merging them is not, on its own, the answer. Annex 9 · Joining up health and social care: what integration can and cannot do

And healthcare is the jam in the sandwich: when social care, children's care and prevention fail, people end up in hospital, and hospitals then cannot discharge them. There can be no serious review of the NHS that does not take all three with it.

Nottingham University Hospitals is removing around 1,279 posts while implementing Ockenden and asking staff to speak up more freely.

The 2026 reforms are real and are the first with consequences attached — manager barring, regulation by 2032, a National Audit Office audit of implementation from 2027 — but every one of them addresses detection and punishment rather than the conditions that produce the failure.

And a system cannot be the source of its own repair. The wounded rarely recover unaided and the self-harmed almost never do — which is why every serious recovery starts with something brought in from outside.

Rabbi Lord Jonathan Sacks put it with a Talmudic maxim: "A prisoner cannot release himself from prison" (Brachot 5b). Knowing where the door is is not the same as being able to open it.

Twenty-five years of internally generated NHS reform is the evidence: the internal market, foundation trusts, two Acts, integrated care boards, the abolition of NHS England, and the same failure described in the same words throughout.

So the fix has to be imported — and the things that work came from outside healthcare altogether. Stop-the-line reporting came from Toyota's production system.

Active-bystander training came from policing, built on research into how ordinary people take part in atrocity, and is now mandatory in more than 430 US law-enforcement agencies. Buurtzorg came from a district nurse who rejected management.

Mayo, the Aga Khan network and Elizabeth Fry's Institution were built by outsiders — religious minorities and women excluded from the professions they went on to reshape.

That is the case for humanity-first healthcare: not another internal programme, but a capability brought in from outside a system that has proved, over twenty-five years, that it cannot generate this one for itself.

Importing is not the same as copying. The English attempts on Buurtzorg and on Sweden's Esther faded because the form was imported and the function was left behind. What travels is the principle, redesigned for the place by the people who have to run it. Annex 15 · Humanity in healthcare: examples

And the same facts describe an opportunity — which is the other half of the case

What is being imported is not foreign to the NHS. It is the NHS's own idea.

Bevan's argument in 1948 was moral before it was administrative — "no society can legitimately call itself civilised if a sick person is denied medical aid because of lack of means" — and the NHS Constitution still carries compassion, and respect and dignity, among its six core values.

And the country has changed in the one way that matters. When Bevan built the NHS, humanity did not need building in: Britain had just come through two world wars, and people had queued, rationed, taken in evacuees and pulled neighbours out of the rubble.

What the country lacked was capability, and that is the half he wrote into law. Seventy-eight years on we have the capability, and the thing he could take for granted has thinned — the share of young people with no close friends has risen from about one in 100 in 2014 to one in 20 today.

What Bevan built into structure, free at the point of need and funded collectively, has survived seventy-eight years and every reorganisation since. What he left to the professions and to chance is the half that appears in every inquiry report.

What was built into structure survived; what was left to exhortation did not — and humanity-first healthcare completes that settlement rather than replacing it.

And no country is better placed to finish it.

Four things sit here that exist nowhere else together: failures documented to a standard no other health system can match; the values already written into a constitution; the measurement to see whether anything changes, with 766,285 staff asked every year how they are treated; and one national system in which a change can land everywhere.

Most countries attempting this would first have to build the evidence, the values statement and the measurement. England already has all three, and has so far used them only to describe the problem.

The proposition in one line. Humanity-first healthcare rests on the capacity to act for the good of others, serves one goal — the patient will be at the heart of everything — and stands on three pillars: every person counts, caring for the carers, and it's who we are.

And it is built through three things that do not depend on who is in charge: the individual, trained to act; the community, in circles that belong to a network beyond the employer; and leadership that answers for both.

The test is what would still be true six months after the chief executive left. Annex 12 · Humanity-First Healthcare: applying lessons from the past

And all three exist for one moment: when doing the right thing will cost you something and nobody would blame you for looking away. Holding the line takes choice, connection and creation. It is a team effort, and champions need champions. Annex 17 · Holding the line

Both sentences are true at once, and that is the whole case. To know the problem, to know the remedy and to act on neither is negligence at the scale of a nation's health.

To know the problem, to know the remedy and to act on it is the chance to build the first health service deliberately designed to stay humane under pressure — on the founding vision of 1948, and fit for the hundred years after it.

The argument of this document is that the second is available now, and costs a fraction of the first.

The inquiry record

Eight major inquiries in twenty-nine months, into different hospitals, specialties and decades, found the same thing: concerns raised and explained away, candour delayed, regulators misled, whistleblowers isolated.

This is not a run of bad luck. It is a pattern with a recognisable mechanism, and Durham shows it working in real time. The one thing to take away: the failures are organisational, they are predictable, and they keep being described in the same words.

Research synthesis

What. Not eight scandals — one condition, observed eight times. The record starts in 2001, not 2024; what changed recently is the rate of reporting. Eight major reports landed in twenty-nine months, April 2024 to September 2026: infected blood, Cass, Covid Module 3, Lord Mann on racism, Ockenden at Nottingham, Amos nationally, Aubrey at County Durham, Thirlwall at Chester.

Nine more are live, including Lampard into roughly 2,000 mental health inpatient deaths in Essex, reporting 2027.

Read alongside Bristol (2001), Mid Staffordshire (2013), Morecambe Bay (2015), Gosport (2018), East Kent and Shrewsbury (2022), they describe one sequence, not eight stories.

So what. Every stage of that sequence is a management decision, not an accident. Concern raised, treated as reputational risk, reframed as a personnel dispute, whistleblower managed out, families not told, regulator shown a curated picture, trust rated adequate, harm continues until an external trigger forces an inquiry. Seven mechanisms recur in almost every report.

Durham is the cleanest single illustration because it was written for the Trust's own board rather than for the public, and still finds every element — a service that lost its GMC training accreditation in 2012 and was still being warned in 2025.

Now what. Treat the pattern as the primary risk, not each scandal as a novelty. The Aubrey verdict is the one-line test any board can apply to itself: did we accept reassurance, or seek genuine assurance?

Detailed worked cases — Chester after Letby, and Durham — are in the annex on organisational failure. Annex 1 · Organisational failure: the worked cases

The inquiry landscape, 2024–2026

Eight major reports in twenty-nine months, covering neonatal murder, maternity collapse, surgical harm, racism, mental health and a historic treatment disaster. Listed newest first.

Still running, and what is due in the next twelve months. Expected by September 2027: the Covid-19 Inquiry's report on adult social care (Module 6, autumn 2026); the National Maternity and Neonatal Taskforce action plan following Amos (December 2026); Phase 2 of the Southport Inquiry (spring 2027); and Baroness Casey's commission on adult social care, which the government now wants completed by summer 2027 (its terms of reference still say a first report in 2026 and a final one by 2028).

Later or not yet dated: the Lampard Inquiry into roughly 2,000 mental health inpatient deaths in Essex, 2000–2023, whose final hearings are in June 2027 and whose report is now expected in 2028; the Nottingham Inquiry into the Calocane attacks and psychiatric discharge decisions, whose hearings closed in September 2026 with no report date announced; the statutory inquiry into Tees, Esk and Wear Valleys, which formally began on 21 September 2026; and Donna Ockenden's maternity reviews at Leeds Teaching Hospitals and University Hospitals Sussex, with timescales "to be confirmed".

Why so many, so fast. Three shifts explain the surge rather than a sudden collapse in safety. The Inquiries Act 2005 gave inquiries power to compel evidence under oath, making them effective against cover-up.

Institutional deference has collapsed, and families organise and litigate far faster than they did. And commissioning an inquiry has become the default political response to a crisis — it demonstrates action and defers decision simultaneously.

The recurring pattern

The striking feature of the inquiry record is not variety but repetition. Read Bristol (2001), Mid Staffordshire (2013), Morecambe Bay (2015), Gosport (2018), East Kent (2022), Shrewsbury (2022), Nottingham (2026) and Chester (2026) in sequence and the same sequence of events recurs, in different clinical settings, across twenty-five years.

Each stage is a management decision, not an accident. That is the significance of the diagram: at every node, someone chose reputation over disclosure.

The seven mechanisms that appear in almost every report.

1. Concerns reframed as conduct. At Chester, consultants who reported Letby were pressed to apologise to her and threatened with referral to the GMC. At Bristol, the anaesthetist who raised the alarm was out of practice for years. The pattern is consistent: the person reporting harm becomes the problem to be managed. The external review commissioned at Chester made that easier rather than harder, and Annex 16 · When the sheriff gets it wrong sets out how.
2. Reputational management as the governing instinct. Thirlwall described board presentations as "an exercise in spin" and found an intention to stall the police investigation. Mid Staffordshire pursued foundation trust status while basic care collapsed.
3. Duty of candour honoured late or not at all. Families at Chester were kept unaware for years that deliberate harm was suspected. At Durham, candour was owed from May 2024 and began in February 2025 — a nine-month breach, with patient safety leads excluded from the decision.
4. Regulatory blindness. The CQC rated trusts on data the trusts themselves supplied, and a "Good" rating reduced the frequency of on-site inspection. Thirlwall found CQC's approach insufficiently "investigative and inquiring", and that greater curiosity in February 2016 "may well" have uncovered the rise in mortality. The failure mode resembles the pre-2008 credit ratings agencies: a rating derived from the rated party's own disclosures.
5. Recommendations implemented in form, not substance. Duty of candour became statute. Freedom to Speak Up Guardians were appointed — roughly 1,200–1,500 of them, mostly part-time and without investigative power, handling over 35,000 cases a year. Safe-staffing ratios, the single most concrete Francis recommendation, were quietly shelved in 2015. Medical examiners, recommended after Shipman in 2003 and again after Mid Staffs in 2013, were not fully rolled out until 2024.
6. The revolving door. Executives who presided over failure retired on full pension, moved to another trust, or left with a settlement. Thirlwall put it as a system finding: a "revolving door" that "allowed the movement of failed managers into new roles, irrespective of the reasons for their failure", oiled by what one witness called "a vanilla reference". Until now there was no register from which a failed non-clinical manager could be barred.
7. Hierarchy suppressing challenge. Junior staff, nurses and — repeatedly — mothers and families were not listened to. Nottingham found minority-ethnic mothers denied interpreter support; Amos found the same nationally.

The conclusion the inquiries themselves increasingly draw is that these are not eight separate scandals. They are one systemic condition, observed eight times.

The worked case: Durham

The pattern documented in one report. The Aubrey review of breast surgery at County Durham and Darlington (November 2025) is the most useful single illustration, because it is a governance review written for the Trust itself rather than a public inquiry, and it still finds every element of the sequence above.

Its thirteen key issues are listed in priority order, and the first two are governance and leadership, not clinical practice. The fuller chain of events is in the annex on organisational failure; what follows is the shape of it. Annex 1 · Organisational failure: the worked cases

The formulation worth remembering is its verdict on the board: it "accepted reassurance rather than seeking genuine assurance". That is the mechanism of regulatory and governance blindness stated in seven words.

What the review found, in sequence:

• The early warning was ignored for thirteen years. The service lost its training status in 2012, with trainees withdrawn by the GMC for inadequate supervision and poor clinical governance. "No substantive action was taken to restore training accreditation", and registrar posts had still not been reinstated by 2025.
• Warnings kept arriving and kept not landing. Internal audits, external reviews (GIRFT 2019 and 2024, Northern Cancer Alliance 2024, Royal College of Surgeons 2025), national benchmarking — the Trust was a regional and national outlier on re-excision rates as early as 2018 — and the National Cancer Patient Experience Survey for 2022 and 2023 all flagged serious concerns. The Trust Board never received the full GIRFT 2019 report, its key findings or an action plan, and as at the review neither the 2024 NCA report nor GIRFT 2024 had been formally presented to a board meeting either.
• Outsourcing without oversight — and into the clinical lead's own companies. Contracts with an independent provider from 23 April 2015, and with two private companies from November 2022, ran outside the assurance framework: monitoring was informal meetings with no agenda, discontinued in 2019 and never reinstated. About half of all patients referred into the Trust were seen outside standard Trust clinics. Both the insourced evening and weekend clinics and the Saturday theatre lists were delivered by the same consultant clinical lead in breast surgery, who was a director of both private companies. Aubrey found "conflicts of interest where consultants held directorial roles in provider LLPs", multiple breaches of standing financial instructions, no board-approved risk assessment, a partnership operating without CQC registration of its own, and indemnity cover that was never confirmed. Nearly £6m ($8.1m) was paid by the Trust over six years to the diagnostic clinics and surgery companies the clinical lead ran privately.
• The money was paid per patient, so volume paid. The contracts came to about £1.8m ($2.4m) a year for breast care in 2024/25, and payment was per patient, which the review found created "a clear financial incentive". Clinics ran at 18–20 patients in a four-hour Friday evening session and 30–42 in a full weekend day, against the Association of Breast Surgery's benchmark of up to 14 in four hours and 28 in eight; Saturday clinics at North Durham held 42 ten-minute slots from 09:00 and often overran to 18:30. Theatre lists ran 10–12 cases where an NHS list takes two or three. Aubrey concluded that "patient selection may have been influenced by throughput and performance metrics ... rather than being driven solely by clinical need". It did not even pay: the partnership delivered £1.604m ($2.17m) of activity against £842,460 ($1.14m) of income — a cost-to-income ratio of 133% and a £761,555 ($1.03m) deficit.
• Self-audit instead of scrutiny. "Breast surgeons audited their own practice without independent validation against national standards", and internal reviews "were carried out by the same consultants whose practice was under scrutiny". Explanations offered by the service were "accepted at face value" and "not subject to timely, critical interrogation of the underlying data".
• Equipment refused while harm accrued. Contrast-enhanced mammography and related diagnostic tools were requested from May 2021 onward; staff were told the Trust was "massively underfunded" and the items "too expensive". Some equipment was finally bought in 2025, nearly four years after the first request, and only after external review.
• Duty of candour breached for nine months. The Northern Cancer Alliance gave verbal feedback on 3 May 2024 and reported on 21 June 2024. The instruction to prepare large-scale candour was given on 3 May and reversed four days later, "following a collective decision which was influenced by external advice" for which no documentation was produced. Patient safety leads — the people whose job it was — were excluded. The CQC and the ICB both wrote in December 2024. Candour began in February 2025, after an incident form was finally submitted on 18 February.
• Staff had said so, repeatedly. They reported they could not "speak up to their manager" and described a "chilling effect" in raising concerns with senior leaders: "everything seems to stall with management", "it fell on deaf ears", and after a wellbeing briefing, that "they would never go to HR again". Some were warned that speaking up "could adversely affect their careers".
• And the response to being told was to look for who told. The review records that "attempts by some Executive Directors to identify contributors" to it breached the anonymity staff had been promised, reflecting "a leadership approach perceived to prioritise control over accountability". That happened during the governance review itself.
• And through nearly all of it, the regulator judged the Trust well-led. The CQC asked about breast services in April 2018, after eighteen months of concerns; the Trust's response omitted an incident in which a patient presented with bone metastases after a delayed diagnosis, and a separate complaint about a delay. The inspection that followed, in July 2019, rated the Trust "Good" overall and "Good" for well-led, on the back of a well-led review that also concluded "Good" — only the safety domain was "Requires improvement". That stood as the regulator's standing judgement for the next six years, across most of the period now under review. It changed only after the reviews had landed: a warning notice served on 23 December 2025 over the Trust's governance systems, its management of identified risk and its processes for learning from incidents and complaints; then, with the inspection reports published on 12 June 2026, well-led downgraded from Good to Inadequate, surgical services rated Requires improvement, and findings of failure to oversee complaints and to manage clinical work subcontracted to third parties.

On why quality lost, the review is blunt. It found staff consistently reported a culture "predominantly performance-driven, with an emphasis on capacity, performance targets, and financial targets", in which "quality was deprioritised unless it directly impacted performance indicators". That is the root-cause argument of Volume 4, evidenced from inside one organisation.

Aubrey's own summary: "Governance structures were bypassed, escalation pathways were ineffective, and clinical leadership was allowed to operate without oversight. This resulted in a culture of complacency, at least 27 missed opportunities, and outdated practice, which led to avoidable harm."

The scale of harm, and where the figures come from. Two different sets of numbers are in circulation and they should not be conflated. The Aubrey report recorded that at the time of writing, 39 patients had been identified for full duty of candour disclosure, with a further 96 reviews concluding no or minor harm and 128 incidents still open.

The Trust's own look-back, running under new leadership, has since gone much further. As reported on 18 September 2026, of 514 cases reviewed, 315 patients were found to have suffered harm — around 61% — including 77 significantly harmed, 20 who underwent unnecessary mastectomies, and one known death. The helpline set up for former patients has taken around 640 contacts.

That review currently covers January 2023 to February 2025. The Trust has been considering extending it back to 2015, which would bring up to 10,000 records into scope, with a decision expected at a public board meeting on 24 September 2026. If it proceeds, this becomes one of the largest clinical look-backs the NHS has undertaken.

Chief executive Steve Russell, appointed after these events, has said "what happened was utterly unacceptable" and that he is "deeply sorry for the harm caused". His admission that "we recognise there was a conflict of interest that wasn't managed properly" connects directly to Aubrey's finding about consultants holding directorial roles in the partnerships the Trust was outsourcing to.

The former clinical lead for breast services, in post from 2013 to 2024, has been suspended from clinical practice while investigations continue and remains employed; a criminal investigation is open. Nothing here has been determined against any individual.

The sequencing is the point. Aubrey reported in November 2025 on why this was allowed to happen. The scale of what happened was still being counted ten months later, and may yet be counted at twenty times the volume.

A governance failure of this kind does not resolve when the report lands; it generates years of subsequent work, which is itself a cost of not having acted in 2012.

And it is worth recording what did eventually work, because it was not the regulator. The trust's outlier status on re-excision and benign lesion excision was flagged by the Northern Cancer Alliance in 2018; surgeons at another trust escalated formally in 2021; the alliance's own peer review of 2023 operations reported avoidable harm in June 2024; and the Royal College of Surgeons' invited review in January 2025 was serious enough, on informal feedback before any report, to force the governance review that followed.

The check that worked was the clinical network — people who could see across organisations. It still took thirteen years from the first warning. Annex 16 · When the sheriff gets it wrong

And read the other way round, this record is an asset. No other health system has documented its own failure mode to this standard — eight major reports in twenty-nine months, each naming the same mechanism, on top of twenty-five years of the same findings.

A diagnosis this complete is what makes a remedy designable, and it is the reason this document ends in a proposition rather than a complaint.

The families

This volume is what the failures did to people, told as far as possible in their own words: Chester, Nottingham, infected blood, Durham, and those who paid for speaking up. Every family asked for the same three things — to be told the truth, to be listened to, and to know it would not happen to anyone else.

The one thing to take away: care is, at root, protection, and the families became the safety system the organisations failed to be, while being treated as the threat.

Research synthesis

What. The people this happened to found out last, and are still living it a decade later. Fourteen babies at the Countess of Chester were murdered or attacked between June 2015 and June 2016; one survivor needs 24-hour care for a life of unknown length.

No family was told by the hospital what was suspected — they learned from the police in 2018 or at the trial in 2023. Eleven years on, parents describe separations, seizures, nightmares every other night, and children who stopped speaking.

At Nottingham, 2,500+ families; at Durham, over 300 patients harmed including 20 unnecessary mastectomies; in infected blood, four decades of state denial.

So what. The complaint is never technical error — it is not being believed, and then not being told. Institutional defensiveness is a second injury, and families name it as such. In every case the truth emerged because someone bereaved would not stop asking, which means families have been the safety valve and the warning bell that the designed mechanisms were supposed to be.

And what they were warning about was organisational failure, not medical mistake: known risk, known remedy, known route to raise it, not taken. And because care is at bottom protection, Chester belongs in this volume rather than outside it: an organisation that has stopped protecting leaves a gap, and a malevolent actor only has to find one.

The families held a line that the institutions did not. They had no authority, no access to the records and every reason to accept what they were told by people who were kind to them, and they kept asking anyway, for years, while grieving.

That is the same act this document asks of staff — done by people with far less protection and far more to carry. Annex 17 · Holding the line

Now what. Test every reform against whether it would have changed what happened to these specific people. This volume exists so the rest of the document cannot be read as an abstraction. On that test, Thirlwall's own line does the work: "Nurses do kill. It is not unthinkable. All doctors, nurses and managers should know that."

What every family in this record says

Every family in this record says the same thing, and it is not about clinical error. It is about not being believed. What the rest of this document calls governance failure and defensive culture was experienced by specific people as something much simpler than that. Read the analysis against their words.

Chester: healthcare's worst nightmare, and it ran for a year

Everything else in this volume is a failure of care. So is this — and the fact that it also involved a deliberate perpetrator is the reason it gets filed outside that category, wrongly. Between June 2015 and June 2016, on the neonatal unit at the Countess of Chester Hospital, babies were deliberately killed and injured by the nurse looking after them.

Lucy Letby was convicted in August 2023 of the murders of seven babies and the attempted murders of six more, and at a retrial in June and July 2024 of the attempted murder of a seventh: fourteen babies and fifteen offences, one baby having been the subject of two attempted murders. The Thirlwall Inquiry states it as "murdering seven babies and of attempting to murder another seven babies (eight offences)".

Where the report records the methods, they were these: insulin added to feed bags (Baby F's blood held insulin at 4,657 picomoles per litre, Baby L's at 1,099); a dislodged breathing tube; massive overfeeding with milk — Baby G's vomit reached the chairs three to four feet from her cot, and 100 millilitres of milk was aspirated from her stomach afterwards.

The unit had recorded one to three deaths a year for the previous five years. In 2015 there were eight, from 468 admissions. There has been no neonatal death on that unit since September 2019.

The babies who survived did not all survive intact. Thirlwall's finding on Baby G: "Baby G suffered profound brain injury from her collapses with lifelong consequences." Her father: "her brain scans changed and clearly showed a brain injury which had wiped out most of her brain … She can't be left unattended due to the risk of her choking.

My wife and I provide the majority of her care and do so lovingly. She is our little girl, our fighter and our star. I have been besotted with her ever since the day she was born."

Her mother: "I feel that Lucy Letby has ruined our lives. She has ruined everything. Our daughter needs 24-hour care because of Letby, we don't know how long she will live and it affects every single minute of all of our days."

Baby G had arrived stable and was nearly ready to go home; the report records that "the nurses on the Unit had prepared balloons, cake and a banner to celebrate our daughter's 100th day", and that "the collapses came, apparently, from nowhere".

Baby F survived with learning difficulties. His mother: "We had to grieve for the life that we thought we were going to have with Child F, with his learning difficulties."

What the parents saw, in their own words. Mother E and F, arriving on the unit: "I'd never heard a baby cry like that. And then I walked into the room and I realised it was my baby. And I went to him, and he had blood around his mouth and I was just shocked."

She was told it would be the feeding tube rubbing his throat, and sent back to the ward. Her own conclusion, years later: "Stable babies don't bleed."

Parents G were told their daughter had vomited and had sepsis; the Inquiry finds flatly, "They were not told that Baby G had collapsed."

The father of the triplets, watching his second son collapse the day after the first: "It was almost an exact repeat of the day before. Everyone was running around like headless chickens looking like they had no idea what was wrong.

I asked one of the doctors what was going on and said 'it's happening again, isn't it'. Nothing was said to me; no one could explain it, again." Their mother, while that son was being resuscitated: "I felt entirely forgotten and ignored."

Father N: "We now know that Child N's oxygen saturations dropped very low overnight … We were not told about this deterioration … I find this disgusting. As parents we have an absolute right to know what was happening to and with our son."

And on the nurse's conduct towards them, Father C on her plugging in the cold cot while his son was still alive: "I believe she wanted to savour my son's dying moments for herself, which fills me with both emotion and anger. Had I not challenged her, she would have further intruded on our private goodbye to Child C."

None of them found out from the hospital. Mother C learned an investigation existed when a friend sent her husband a photograph of a newspaper article. She went to the bereavement office, and "the first question that I was asked was whether I was from the press, which was an awful question to be asked at that point".

Told the Trust had tried her landline once, and that staff "didn't know whether parents would want to know", she replied: "That is not your choice to make, who would want to know, when you're talking about the death of people's children."

Most learned the substance from the police in 2018 or at the trial five years later. Mother C: "to not inform us at all until somebody is arrested is unforgivable." The triplets' mother: "Everything I have since found out about what really happened I have learned through the police, through the trial and through my solicitors.

Within the NHS there is supposed to be a duty of candour. Nobody at the Countess of Chester Hospital was candid with us." Mother G: "I feel that the Countess of Chester Hospital have covered up what happened to our daughter for years, telling us all this time that our daughter suffered neonatal sepsis, despite there not even being a marker for sepsis in her blood tests at the time."

One mother was handed the redacted external review by a taxi driver at her door, thirty minutes before it went online: "I was absolutely mortified. I was panicking. I didn't know what was going on … it beggars belief."

The Chair's verdict on that episode: "The families were failed by the Countess and by the RCPCH, from the commissioning of the report to its publication."

They were handled as a reputational problem. The Inquiry's own finding: where executives "focus on the effect of a situation upon reputation rather than the effect of the situation on patient safety, the strong impression given is that reputation has become more important than patient safety".

A note of the pre-inquest discussion about one baby recorded the desired outcome as "Narrative Verdict 'Unascertained' No negative comments No press".

On the parents of the triplets, left without any support after two of their sons were killed: "The absence of any support in these circumstances was inexcusable." One mother said it without the Inquiry's restraint: "They shouldn't have been concentrating on saving their own skins and jobs and reputations.

Babies died because someone in an office being paid hundreds of thousands of pounds didn't want the hospital to look bad if they shut."

Eleven years on, this is what it has done to them. The report's foreword opens on the arithmetic of anniversaries: 2026 marks the eleventh anniversaries of the deaths of five of the babies and the injuries to another, and the tenth of the deaths of two more.

Mother K: "you don't only just grieve your daughter, you're grieving who you were. I grieve who we were as a husband and a wife. It just completely destroys what's around you and you have to pick yourself up and find out who you are again in this new world and it just doesn't stop, it doesn't go away and we live with it every single day."

Mother I: "I wore sunglasses constantly to hide the pain and tears from my other kids as I didn't want to upset them"; her husband "wished that he was dead instead of our baby. We even separated for a while as neither of us could deal with what happened.

Our other kids also suffered. They gave up things they enjoyed and my older daughter stop[ped] speaking."

Father L and M had the first seizure of his life in front of his children as the trial approached, which his doctors attributed to the stress: "Even to this day I get flashbacks to what I saw on the unit."

The triplets' father: "I have visions and nightmares approximately every other night, of being back at the hospital … the colour of their veins and their veins pulsating is something which comes up a lot."

Mother H, on what it did to her relationship with healthcare itself: "you put your whole trust into them to take care of your baby and that has been completely abused … I really do find it very traumatic to have to go back to a hospital." Mother D: "Since [she] passed away I live beside my own shadow."

And the sentence that ends any argument about urgency. It comes from the mother of Baby Q, the last baby harmed, months after the consultants had first raised their concerns: "One thing that I really struggle with is, this had been going on for such a long time and Child Q was the last one if she would have been caught before, this wouldn't of happened to Child Q." The delay between the first concern and the police being called was not an administrative fact. It has names attached to it.

Thirlwall's instruction to the whole service is one line: "Nurses do kill. It is not unthinkable. All doctors, nurses and managers should know that." And the concession she records as the one made by almost everybody who gave evidence: "there was a total failure of safeguarding at every level".

What the hospital did after all this — between 2016 and 2022 — is set out in the annex on organisational failure. It is the part of the story that is hardest to read. Annex 1 · Organisational failure: the worked cases

Care is protection — which is why this is still a failure of care

The deliberate act does not put Chester outside the category of care failure; it puts it at the centre of it. The usual way to set this case aside is that it was a crime rather than a system — one person did something monstrous, and no amount of culture work would have stopped her wanting to.

That reading holds only until you ask what care is for. Care is not only kindness and communication. At its foundation it is protection: a neonatal unit exists to keep alive babies who cannot protect themselves, and to stand between them and anything that threatens them, including one of its own staff.

Measured against that, the failure is not a side-issue to the crime. The unit failed at the first thing it existed to do, and went on failing for a year after it had been told.

What a malevolent actor needs is not skill but an unguarded gap. Atrocity-prevention research has a name for that gap. In 1979, in Accounting for Genocide, the sociologist Helen Fein described a society's universe of obligation — the people "toward whom obligations are owed, to whom rules apply, and whose injuries call for amends".

Harm becomes possible at scale not when a population turns malevolent, but when a category of person falls outside that circle, so that nobody in a position to act feels obliged to act.

Fein was writing about nations; the same boundary can be drawn inside a single organisation, and the inquiry record shows where Chester's was drawn. Inside it were the unit's reputation, the executives' standing and the institution's account of itself.

Outside it were the babies, their parents, and the consultants who raised the alarm. That is how the response to a suspected killer came to be the management of the doctors: the obligation people felt ran to the organisation, not to the patients.

Which is also the answer to the objection that culture work cannot stop a murderer. It is not meant to. What it changes is whether the gap is there to be used — whether a colleague can say what she has noticed without weighing her career against it, whether a pattern of unexplained collapses is treated as data rather than as disloyalty, whether a parent's account counts as evidence rather than as distress.

What stopped it in the end was two consultants who would not let go, and the police. Everything between June 2015 and May 2017 is the measure of how wide the gap had been left open. Volume 5 sets out the continuum along which it opens, the small permissions that make larger ones unremarkable; Chester is what stands at the end of that continuum when someone arrives who is willing to use it.

Nottingham, in mothers' own words

These are from the Ockenden review evidence, as reported at publication:

"They then spent 12 hours not believing me that I was in labour, denying me of pain relief." (2020)

"I felt coerced to accept an oxytocin drip … the midwife performed a membrane sweep without notice or consent." (2019)

"My waters were broken against my will."

"The midwife told my partner to put his hand over my mouth."

"My immediate reaction was to terminate the pregnancy because I was too frightened." (2014)

Donna Ockenden's own words to those families: "Many shared the most painful chapters of their lives, speaking with extraordinary courage about experiences of trauma and grief … To every family who came forward, I want to say this: we have listened." The significance of that last sentence is that it had to be said at all.

Infected blood: a different shape of failure, and worth citing precisely

The treatment itself was given from the late 1970s to the early 1990s, with the critical years in the 1980s. It is not recent practice and it is not the same animal as a unit failing now; read across from it carelessly and the comparison can be dismissed.

What makes it belong in this document is not the original harm but the four decades that followed it: a sustained institutional and state refusal to admit what had been done, records lost or destroyed, patients not told what they had been given, and the line that the matter was finished business held across successive governments.

That part is not historical in character at all — it is the same reputational reflex the other inquiries describe, sustained for longer than the lifetime of many of those it harmed.

Sir Brian Langstaff opened publication in May 2024 by handing the report back to the people it concerned: "This is your report. I may have done the words, certainly the errors, the omissions they're all mine … But the words come from you and your stories."

His central finding was put in the same register: "This disaster was not an accident. People put their faith in doctors and in the government to keep them safe and their trust was betrayed." And on what the institutions added to the original harm: "The NHS and successive governments compounded the agony by refusing to accept that wrong had been done."

Note the division in that last sentence — the harm was historical, the compounding was continuous, and it ended only when the state stopped denying it.

Durham: what irreversible means

The breast surgery harm at County Durham and Darlington has a particular quality, which the former breast surgeon Liz O'Riordan named in six words while reviewing patients' photographs for reporters: "You can't put a breast back."

Reviewing one woman's case, where a follow-up letter acknowledged that a second lumpectomy could have been offered instead of the mastectomy she had, her verdict was: "That's not acceptable at all. I'd have brought her back for revision surgery."

Much of the harm in this document is measured in delay and distress. Some of it removed parts of people's bodies that did not need to be removed, and no inquiry, apology or barring scheme returns them.

The cost of speaking up, borne by families too

Julie Bailey founded Cure the NHS after her mother Bella died at Stafford Hospital, running the campaign from her café and securing the public inquiry that produced the Francis reports. She was abused in the street by local people who blamed her for the threatened downgrading of their hospital, and eventually left the town.

She was appointed CBE in 2014. The pattern in which the person who raises the alarm pays the price is not confined to staff.

And it reaches the top of the organisation. Dr Susan Gilby became medical director at the Countess of Chester three weeks after the arrest and chief executive after that; she is the person who established internally that the concerns had been real and that the focus had stayed on the people who raised them.

In July 2022 she made a whistleblowing complaint about the conduct of the Trust's new chair. Four members of the board — the chair, the deputy chair, another non-executive and the chief people officer — then ran a covert campaign to remove her, which they called "Project Countess".

She was offered a secondment on condition that she withdrew the allegations, declined, and was suspended on 2 December 2022.

She won at tribunal on 13 February 2025 — constructive unfair dismissal, automatic unfair dismissal, and detriment for protected disclosures — in a judgment Thirlwall describes as containing "excoriating criticism of the leadership of the Countess at that time", and which found that hundreds of messages and documents had been deleted.

The chair resigned within days; the case settled in January 2026 for £1.4m ($1.89m) in damages and about £3m ($4.05m) all in. The full account is in the annex on organisational failure. Annex 1 · Organisational failure: the worked cases What the harm costs families, staff and the public purse is set side by side in Annex 3 · The impact: when people stop caring.

What belongs in this volume is the symmetry: the price of speaking up was paid by a bereaved daughter running a campaign from her café and by a hospital chief executive, and the mechanism used against them was the same.

Why this volume comes before the analysis

Three things follow from the testimony that do not follow as clearly from the analysis. First, the harm was relational before it was clinical — the recurring complaint is of not being listened to, not of technical error.

Second, institutional defensiveness is itself a second injury, and families name it as such. Third, in every case the truth eventually emerged because patients and families would not stop asking. They were the safety valve and the warning bell — which is precisely what the designed mechanisms were supposed to be, and were not.

Organisational failure, not medical mistake. This is the distinction that matters most, and the one most often blurred. Healthcare belongs with aviation, rail and nuclear power in the class of systems where failure is not contained: when coordination breaks down it does not injure one patient, it takes many lives, and it damages the lives of the families left behind alongside them. Almost nothing in the inquiry record is the limit of medical knowledge being reached.

Mid Staffordshire was not a diagnostic puzzle. Morecambe Bay, Shrewsbury and Nottingham were not the frontier of obstetric science. Durham was not an unrecognised surgical technique. Chester was not a clinical puzzle either: the deaths were unexplained only for as long as the explanation was unwelcome.

In each of them the risk was known, the remedy was known, and there was a route by which someone could have raised it — and the organisation did not act.

Aviation's answer to this class of failure was to make the system itself the object of scrutiny: mandatory occurrence reporting, protected candour, crew resource management, findings implemented as a condition of continuing to operate. Healthcare has the reports and not yet the implementation.

Which is what makes these failures unforgivable rather than merely tragic. Not that they happened — every complex system produces failure — but that they happened after everything already known. Each inquiry since Bristol has described the same pattern with greater precision, and each time another organisation reproduced it.

By 2026 no board can claim the mechanism was hidden from them. That is the difference between an accident and a decision not to act on knowledge already held, which is why Langstaff's finding on infected blood — "this disaster was not an accident" — reads as the sentence the whole inquiry record has been building towards.

And the families left the service something it can still act on: a design standard. Their test is one question, and it is free to apply — would this reform have changed what happened to us? Hold every mechanism in this document against it.

Barring a chief executive after the fact would not have changed it. Being believed the first time would have. That is the whole of what follows, put in the words of the people who paid for the knowledge.

Why it happens

The causes are not mysterious. Thirty years of reform taught the NHS to manage money and targets and left humanity to chance; the workforce is stretched and churning; capacity is thin; and the service's values were written down in 2009 but never wired into anything.

Johnson & Johnson's Credo shows what values built into an organisation look like — and what happens when even they stop being renewed. The one thing to take away: a system produces the behaviour it is built to produce, so change what it is built to do.

Research synthesis

What. The NHS measures what is easy to count, and manages only that. Thirty years of reform built organisations judged on finance and throughput, run by a mobile executive class. Dr Chris Turner, who worked in the emergency department at Mid Staffordshire from 2007: "in mid staffs the exec team was measured on finance, and pretty much nothing else."

The service now runs at over 90% bed occupancy with among the fewest beds and scanners of comparable health systems, misses every constitutional target — 75.0% four-hour A&E performance against 95%, 7.33m elective pathways, 72.8% cancer 62-day against 85% — and saw 30,323 leavers against 41,542 joiners — churn rather than shrinkage — in the year to March 2026, with burnout consistently among the top three reasons.

So what. The instruments built to make the NHS accountable are substantially what made it blind. An organisation that measures success by whether a box is ticked loses the ability to notice that a patient has lost their dignity, or that babies are dying in a pattern.

Capacity does not cause governance failure, but it removes every margin that would otherwise absorb it — no pause, no second opinion, no investigation. And until 2026 there was no regulator, and no barring power, for the non-clinical managers making those calls.

Nor were the values that should have counterweighted all of it wired into anything: the NHS Constitution's six values appear in no appointment, appraisal, payment or rating that carries a consequence — which is the difference between a poster and Johnson & Johnson's Credo, which its own management had to argue out in 1975, which employees have rated since 1985, and which its chief executive said gave him "the ammunition I needed" to order a $100m recall in 1982.

Now what. Three options, no fourth — and only one of them is affordable this year. Add capacity, build the human capability to carry the strain, or keep commissioning inquiries. England has committed to the first in principle, delivered it slowly, treated the second as a soft extra, and paid for the third — which is the most expensive and the only one that buys nothing. The capability option is the only one available this financial year.

Root causes

The mechanisms in Volume 2 have a common origin: a thirty-year reconstruction of the NHS around measurable performance, delivered through locally autonomous organisations whose leaders were judged on finance and throughput.

New Public Management and the tyranny of the target. The internal market arrived in 1990; foundation trusts followed from 2004, with earned autonomy conditional on hitting financial and operational targets.

National oversight concentrated on what was easy to count — four-hour A&E waits, referral-to-treatment times, the balance sheet. Francis concluded that Mid Staffordshire's board became so fixed on foundation trust status and cost control that it lost sight of basic care.

The mechanism is not that targets are wrong; it is that what is measured becomes what is managed, and dignity, candour and safety culture are not on the dashboard.

Professionalisation of management. Hospitals once run by medical superintendents and matrons — people who had worked the wards — came to be run by a mobile executive class moving between the NHS, local government and the private sector.

The gain was financial and operational discipline. The loss was proximity: safety concerns reached boards as risk items rather than as clinical facts, and whistleblowers appeared on the risk register rather than in the clinic.

Professionalisation of nursing. Moving nurse education into universities produced technically stronger clinicians capable of managing far more complex medicine. Critics argue it also devalued the hands-on, relational core of nursing — feeding, washing, noticing — and substituted documentation for presence.

This is a contested reading rather than a settled finding, and it is worth holding lightly: the same period saw enormous gains in clinical capability.

Accountability without consequence. Clinicians are regulated by the GMC and NMC and can be struck off. Until 2026 there was no equivalent for non-clinical managers. The Fit and Proper Person Test introduced after Francis had no barring power behind it.

Unitary board responsibility, intended to make directors jointly accountable, functioned in practice to make no individual accountable. Thirlwall named the consequence as a system fact: a "revolving door" that "allowed the movement of failed managers into new roles, irrespective of the reasons for their failure".

Consequence on its own is not the answer either. What has to be written down in advance is the line: an honest mistake, reported promptly, is protected; silence, downgrading an incident and retaliation against the person who raised it are not. Trust, forgiveness and being held to account only work as a set. Annex 13 · Radical leadership

And the founding values were never hard-wired — they were put on the wall. The NHS Constitution sets out six values, and on their own terms they are the right ones. What was never built is the machinery that makes them bind.

They do not appear in the terms on which executives are appointed, appraised or paid; no regulator has ever rated a trust on them as such; no board has ever failed a test because of them. Annex 4 · The founding vision of 1948

Hard-wiring is the whole of the alternative, and it has a name. Enablership is the work of making an organisation able — building the capability into the rules, the habits and the fabric, so that it survives the departure of whoever built it.

The founders of the American republic faced the same problem in a harder form and answered it by design: a constitution written for the officeholder you might actually get, not the one you hope for. Bevan did exactly that with the money in 1948, and left the humanity to chance. Annex 14 · Healthcare: why the Art of Enablership is critical

Bevan built the financial half of the 1948 promise into an Act of Parliament and left the humane half to chance — and what was written into structure survived seventy-eight years of reorganisation while what was left to exhortation did not survive the first sustained squeeze.

Values do not fail because people stop believing in them. They fail because nothing in the organisation depends on them.

The clearest worked example of the alternative is corporate rather than clinical. Johnson & Johnson's Credo was written in 1943 by General Robert Wood Johnson and presented to the board that December, a year before the company went public.

It ranks its responsibilities: patients, doctors and nurses and all others who use the products first; employees second; communities third; and then, explicitly last, "Our final responsibility is to our stockholders. Business must make a sound profit."

That ordering was not left to hang in reception. In 1975 James Burke, about to become chief executive, ran a series of Credo Challenge meetings at which the company's management had to argue out whether the values still held — with abandoning the document on the table as the honest alternative to quietly ignoring it.

In 1985 he added the Credo Survey, asking employees worldwide to rate how well the company was actually living up to each of the responsibilities in turn. Between them, those two mechanisms made the values contestable and measured rather than decorative.

Then they were tested. In the autumn of 1982 seven people in the Chicago area died after taking Extra-Strength Tylenol capsules that had been laced with cyanide. Tylenol held about 35% of a $1.2bn analgesic market.

Burke recalled 31 million bottles at a cost of about $100m, before any regulator required it, and relaunched two months later in tamper-evident packaging; market share was back to 35% within about a year.

His own account of why he was able to do it is this volume's argument in one sentence: the Credo, he said, "gave me the ammunition I needed to persuade shareholders … to spend the $100 million on the recall".

The document was usable in the worst week of the company's history because someone had spent the previous seven years making it explicit, arguable and measured. A framed poster would not have moved a shareholder.

And the counter-lesson is in the same company, which is why this is a warning rather than a reassurance. J&J's recent record is not its 1982 record. In February 2022 it agreed to pay up to $5bn towards US opioid settlements, on terms it says are "not an admission of any liability or wrongdoing".

In July 2026 it proposed a $5.5bn settlement of roughly 76,000 remaining ovarian-cancer claims over talc, after a bankruptcy court rejected an earlier $9bn plan; the company maintains it would have prevailed at trial.

Whatever the merits of any individual claim, the trajectory is the one this volume describes: values hold while they are stress-tested and measured, and erode when performance is the only thing anyone is actually measured on.

A constitution does not protect an organisation. Maintaining one does — and the NHS has never built the maintenance.

A first-hand account of the mechanism. Dr Chris Turner, now an emergency medicine consultant in Coventry and co-founder of the Civility Saves Lives campaign, joined the emergency department at Mid Staffordshire as a junior doctor in 2007 — the grade now called a resident doctor — and gave evidence to the public inquiry.

His summary is the shortest statement of this volume's argument available: "I worked in mid staffs, and in mid staffs the exec team was measured on finance, and pretty much nothing else."

His inquiry evidence described a "culture of bullying and harassment", staff "utterly demoralised", nurses threatened "on a near daily basis" with losing their jobs, and "no significant medical leadership".

And it cost him the job: "First I stopped sleeping and eventually eating"; he lost two and a half stone in nine weeks, slept three hours a night, and "after two years I broke under the pressure and resigned".

The person best placed to describe what that culture does to clinical work is someone it removed from the department.

The paradox is exact. The instruments introduced to make the NHS accountable — targets, ratings, audits, inspection regimes — are substantially what made it blind. An organisation that measures success by whether a box is ticked loses the capacity to notice that a patient has lost their dignity, or that babies are dying in a pattern.

The workforce is leaving, and leaving early

This is where the professionalisation argument stops being historical and becomes a live operational problem. In the year to 31 March 2026 the NMC register held 867,935 professionals, with 41,542 joiners against 30,323 leavers — a net gain, but a very large churn.

The NMC's own finding is that the top three reasons for leaving have been consistent over time: retirement, physical or mental health, and burnout or exhaustion.

The early-career trend is the one that matters for culture, and it has to be stated with its base, because a percentage increase on its own tells you nothing about scale. RCN analysis of NMC data (November 2024) found that the number of UK-educated nursing staff in England leaving the register within five years of registering rose from an estimated 697 in 2021 to an estimated 1,166 in 2024 — the 67% increase usually quoted, on a small base — while those leaving within ten years rose 43%, from an estimated 1,607 to 2,295.

The RCN projects a further 11,307 leaving within their first ten years by 2029 on current trends. Nicola Ranger: "Nursing is an incredible career, and it should be one for life, but thousands are quitting early because they are burnt out, underpaid and demoralised."

For context, registered nursing vacancies in England stood at 23,046, or 5.3%, in June 2026 (NHS Vacancy Statistics, published 27 August 2026), and the RCN puts real-terms nursing pay down 23% since 2010/11.

The NHS Staff Survey 2025 records 29.55% of staff often thinking about leaving, a three-year high, and 16.75% saying they would go as soon as they found another job, an all-time high.

What those figures do and do not show. Leaving the register and leaving the NHS are two different measures, and the gap between them is wide enough that conflating them is the main way this argument gets discredited.

On the register: Nuffield Trust analysis of NMC data (Waste not, want not, September 2023) found 98% of UK-trained nurses and midwives were still registered three years after joining — early attrition from the profession itself is in the low single digits.

On employment: the same report found around 18% of nurses had left NHS hospital and community settings within two years, alongside radiographers at 17% and occupational therapists and physiotherapists at 21%.

So the honest version is three-part. Very few newly qualified nurses leave nursing quickly. Nearly a fifth leave the part of the NHS they started in. And the small number who leave the register outright is rising fast, which is why it is worth watching rather than because it is yet large.

The mechanism is what the literature calls transition shock. Under the apprenticeship model, student nurses were hospital employees embedded in a ward's social fabric from the start; they acquired the pacing, the hierarchy and the coping before they carried accountability.

Under graduate entry they are supernumerary — legally extra, protected, not counted in staffing — and that protection vanishes on the first day of registration, when a full patient load and full legal accountability arrive together.

Clinical knowledge is better than it has ever been; practised floor resilience is thinner. England is now partly reversing course through Registered Nurse Degree Apprenticeships, which put people on the wards as paid employees while they qualify.

Why this belongs in root causes rather than a workforce annex. The inquiries describe organisations where nobody with enough standing and confidence challenged what was happening. An early-career workforce that churns every few years, mentored by people who are themselves two or three years qualified, is structurally less able to do that.

Psychological safety depends on people who have been there long enough to know what normal looks like, and to be unafraid of the consequences of saying so.

People can break, because they are human. Like a muscle, the right kind of stretch, with support and time to recover, makes people stronger. Relentless strain without it tears them: they tire, burn out and leave. And, as in any team sport, every player lost puts more weight on everyone still on the pitch.

The annex on doctors and nurses shows what that looks like from the inside: teams broken up, a broken career ladder, and organisations that stopped treating their staff as people. Annex 6 · Doctors and nurses: when the organisation stops caring

A caveat on two figures you will encounter. First, "30–50% of new nurses leave within two years" is not a UK figure, and it is not traceable to a primary US source either. The actual US primary figures are 22.7% of newly hired registered nurses leaving within a year (NSI National Health Care Retention & RN Staffing Report, 2026 edition) and 17.5% leaving their first job within a year (Kovner et al., Policy, Politics, & Nursing Practice, 2014) — and both measure changing employer, not leaving the profession.

No equivalent UK cohort figure is published at all; the nearest is the Nuffield Trust's 98% three-year register retention. So do not state a UK early-attrition rate: state the counts, and name which of the two measures you mean.

Second, student attrition is a different thing again — leaving the course, before ever registering. Nuffield Trust, using HESA data, found around one in eight nursing students in England and Wales did not gain their intended degree between 2014 and 2020 (about one in six in Scotland); the RCN's higher 21% figure is its own calculation on three specific intakes, and the NAO's 33% is a completion-on-time measure, not a dropout rate. Three numbers, three different questions.

Performance and capacity

Performance is off the floor but nowhere near standard. Every headline constitutional target in England is being missed, and has been for years. The waiting list has come down from its September 2023 peak of 7.77 million but rose again through summer 2026.

Sources: House of Commons Library NHS key statistics, BMA backlog analysis, Nuffield Trust NHS performance dashboard, NHS Vacancy Statistics (27 August 2026).

For scale: the list stood at 4.57 million in February 2020. The government's commitment is to restore the 92% eighteen-week standard by March 2029. Workforce has grown substantially over the same period — doctors up around 24% and nurses around 22% since February 2021 — which makes the persistence of the backlog a productivity and flow problem as much as a headcount one.

The four nations side by side, and the social care figures, are in Annex 2 · UK healthcare in numbers.

Against comparable countries, the King's Fund assessment is that the NHS is "neither a leader nor a laggard" among eighteen similar systems. It spends slightly below average per head, and it has substantially fewer CT and MRI scanners, fewer hospital beds, and strikingly low numbers of doctors and nurses per capita than most peers, with heavy reliance on internationally trained staff.

It is strong on financial protection — people are not bankrupted by illness — and on administrative efficiency. It lags on life expectancy, on deaths avoidable through timely treatment, and on cancer survival.

Capacity as the silent multiplier. None of the failures in this document happens in a system with slack. A system running permanently at the edge of its capacity has no room in which to pause, escalate or investigate.

Capacity does not cause the governance failures described above, but it removes every margin that would otherwise absorb them — and every incentive points towards keeping the flow moving.

The three-way choice, and there is no fourth option

Project management has long had its iron triangle — cost, quality, time: pick two, because under pressure something always gives. Healthcare under sustained pressure has its own version, and a service can only decline to choose by choosing the third by default:

1. Add capacity. Beds, scanners, theatre lists, staff, time — so the strain itself comes down and the margins come back.
2. Build the human capability to carry the strain well. The civility, escalation, speaking-up and active-bystander capability that determines what actually happens in the moment when there is no slack and someone has to say the difficult thing anyway.
3. Keep commissioning inquiries. Because if neither of the first two is done, the difference is absorbed by patients, and they absorb it literally. The mistakes are not hypothetical, and the harm does not distribute itself thinly across a population: it lands on individual people in individual beds, and on the families in Volume 3.

England has spent fifteen years committing to the first in principle, delivering it slowly, treating the second as a soft extra, and paying for the third. The third is the most expensive of the three and the only one that buys nothing — an inquiry is purchased after the harm, and none of its cost reaches the patient it was meant to protect. Annex 3 · The impact: when people stop caring

The honest framing is not humanity instead of capacity. Capacity is necessary, and it takes years and money the service does not have this year. The capability to hold the line under pressure can be built now, at a fraction of the cost, and without it the capacity that does eventually arrive will still be run by people with no margin left for each other. That is the whole case for sequencing the second option first rather than last.

Incivility and the continuum

Rudeness is not bad manners. It measurably degrades clinical performance, it comes mostly from good people under pressure, and it spreads: each act that goes unchallenged makes the next easier, and research on how ordinary people come to do terrible things describes the same mechanism at its far end.

The one thing to take away: small acts are signals of a system's health, and the cheap place to act is at the small end, by someone speaking up early.

Humanity and inhumanity both run along a continuum, and both spread. A single act — a kind word to a frightened patient, a sneer at a junior nurse — does not stay where it lands. It changes how the person who received it treats the next person, how the people who saw it read the next ambiguous moment, and what everyone learns is normal here.

The ripples travel further than anyone intends, for good or ill. Rudeness passes to people who were never involved, and each act that goes unchallenged makes the next one easier.

The same mechanism runs the other way: care practised again and again builds the capacity for care, and one person who speaks up early can stop a spiral that would otherwise end in a public inquiry.

Small acts are also signals: they tell you the state of the whole system. Safety science has worked this way for nearly a century. Herbert Heinrich's accident triangle (1931), and Frank Bird's later version drawn from about 1.75 million accident reports, pictured serious harm as the tip of an iceberg resting on a much larger base of minor incidents and near misses — the things that are cheap to notice and cheap to fix.

The ratios have rightly been criticised: they vary by industry, and major disasters often have different causes from everyday slips. BP's Texas City refinery, where fifteen people died in 2005, had good personal-injury figures; the Baker Panel found the company had "mistakenly interpreted improving personal injury rates as an indication of acceptable process safety performance".

The lesson that survives is narrower and more useful: watch the small signals that share a cause with the harm you fear.

In healthcare, how people treat each other is one of those signals, and it has been shown to predict harm. At Vanderbilt University Medical Center, Gerald Hickson's team found that unsolicited patient complaints predicted which doctors would face malpractice claims (JAMA, 2002).

Across 13,653 patients at Vanderbilt and Stanford, surgeons with more reports from coworkers about unprofessional behaviour had more complications — about 18% more with one to three reports, and 32% more with four or more (JAMA Surgery, 2019).

Vanderbilt acts on those signals at the small end: the first response is an informal, private conversation with a peer, most doctors respond to it, and only a tiny fraction ever reach disciplinary action.

In the NHS, trusts that report more safety incidents have better safety cultures, not worse (Hutchinson and colleagues, 2009) — a high count of small signals is often a sign that people feel safe to raise them.

That is why incivility, and civility, are among the most important flags an organisation can watch. They appear early, they are cheap to see, and they share a cause with the harm that inquiries uncover years later.

Research synthesis

What. Rudeness is not bad manners — it measurably degrades clinical performance, and it comes from good people under pressure. Half of those who admit to rudeness say it is because they feel overloaded; 4% because it is enjoyable.

In Riskin's 2015 neonatal simulation, mild rudeness significantly worsened diagnostic accuracy, information sharing and procedural performance; in Porath and Erez's experiments, targets performed 33% worse and produced 39% fewer creative ideas, witnesses 22–28% worse, and recipients were 74–84% less likely to help a colleague.

More than 75% of healthcare employees have witnessed physician incivility (CMAJ, 2024). In the United States, 79% of practitioners had personally experienced disrespectful behaviour and 27% knew of a medication error in the previous year in which it played a part (ISMP, 2021); in Australia, 38.8% of hospital staff experienced incivility or bullying at least weekly and half said unprofessional behaviour harmed patient care (MJA, 2021).

So what. Each small act licenses the next, until nobody can see the threshold has moved. Andersson and Pearson's model has tipping points; Foulk's work shows rudeness transmits to people who were not involved and primes witnesses to see hostility a week later; Porath shows bystanders pay a performance cost, so no one in the room is neutral.

Which is why the line has to be held early, when holding it still looks like an overreaction. Every step along this continuum is easier to stop than the one after it, and harder to notice. The person who objects first is objecting to something small, in a room where nobody else has, and pays the full social price for it. Annex 17 · Holding the line

Ervin Staub, who studies the extreme end, states the mechanism in one line: "Once perpetrators begin to harm people, the resulting psychological changes make greater harm-doing probable." James Waller's account of ordinary people is the same situation logic. That is the state every inquiry in this document describes on arrival.

Now what. Break it at the small end, because that is the only cheap place to break it. The spiral is stopped by someone naming the first thing — or expensively, years later, by a public inquiry.

Staub's own mechanism runs both ways: "learning by doing can change people in either positive or negative directions." He turned that into ABLE, mandatory active-bystander training now used by more than 430 US law-enforcement agencies.

Healthcare shares every characteristic of that profession and has nothing at that scale. Single training events do not work; multi-component, leadership-led programmes do.

Incivility is contextual, not a property of bad people

Start here, because everything else depends on it. Dr Chris Turner, now an emergency medicine consultant in Coventry, who joined the emergency department at Mid Staffordshire as a junior doctor in 2007, resigned after two years under the pressure of it, and gave evidence to the public inquiry, puts it about himself rather than about others: "Yeah, I've been rude many times… I got it wrong last week at work and I've had to go and apologise to somebody.

She was really hurt and I feel absolutely terrible about it." He describes the spread as "like an emotional muscle — once we start to display rudeness, incivility, it gets easier and easier to accept."

That framing matters for three reasons. It is almost certainly true — Porath's own data show that half of people who admit to rudeness say it is because they feel overloaded, and only 4% because it is enjoyable and they can get away with it.

It is the only framing that leads anywhere useful, because a theory of bad apples generates disciplinary processes, while a theory of good people under pressure generates changes to the pressure and to people's capacity to withstand it.

And it is the one that survives contact with the inquiries, none of which found villains so much as ordinary professionals in environments that made the wrong behaviour easy and the right behaviour costly.

So there are two distinct failure points, and they need different remedies.

The first is the environment. Sometimes that is irreducible — an emergency department at three in the morning is genuinely hostile, and no amount of culture work makes a resuscitation calm. More often it is organisational: the rota gap, the unrepaired equipment, the target that cannot be met honestly, the manager who treats a concern as a threat.

Turner's own diagnosis of Mid Staffordshire is a one-line summary of the organisational version: "in mid staffs the exec team was measured on finance, and pretty much nothing else." That is fixable, and it is what most of this document is about.

The second is individual capacity — whether a person under acute stress can regulate their own emotional state, see past their own distress, and still act for the good of someone else. This is the part the NHS has consistently tried to solve with policy and has consistently failed to solve, because it is not a policy problem.

It is a practised human capability, and like any capability it can be built, eroded, or never developed at all. A workforce that is exhausted, churning and mentored by people two years qualified is a workforce in which that capacity is thin, however good its protocols.

The practical implication is that an organisation needs both: conditions that do not routinely overwhelm people, and people equipped to stay human when conditions do. Addressing only the first leaves staff with no resources on the bad night.

Addressing only the second is asking individuals to absorb a system's failures, which is both unfair and unreliable.

The wider version of this idea is very old. A formulation often attributed to the twelfth-century Andalusian philosopher Ibn Rushd (Averroes) runs: "Ignorance leads to fear, fear leads to hatred, and hatred leads to violence."

The chain is worth taking seriously even though its attribution is doubtful — see the sources volume — because it identifies the root as a failure of understanding rather than an excess of malice. What is not understood becomes frightening; what frightens is easier to despise; what is despised is easier to harm.

The same insight is inscribed at Rwanda's genocide memorials, in the words of survivors: "If you had known me, and you had really known yourself, you would not have killed me." Note the two halves of that sentence.

Knowing the other person is not sufficient; the speaker also requires the killer to have known himself. That is exactly the distinction above — understanding of others, and the self-understanding that makes it possible to act on it under pressure.

The compounding: why small acts are not small

This is not a comparison between a stressed ward and a genocide. It is a claim about direction of travel, and about a mechanism. Inhumanity is a continuum, and the danger is not only that small acts sit at the thin end of it — it is that each one licenses the next.

That is the part usually missed, and it is the part with the strongest evidence behind it. Three findings, in sequence:

It starts small and low-stakes by definition. Andersson and Pearson's foundational model defines incivility as "low-intensity deviant behaviour with ambiguous intent to harm the target, in violation of workplace norms for mutual respect" (Academy of Management Review, 1999).

The ambiguity is the point: nothing at the start is clearly a breach, so nothing triggers a response. Their model then describes a spiral with tipping points, in which exchanges escalate in intensity, and they position incivility explicitly as a precursor capable of escalating toward aggression.

It transmits to people who were not involved. Foulk, Woolum and Erez — "Catching Rudeness Is Like Catching a Cold", Journal of Applied Psychology, 2015 — found that people treated rudely by one partner went on to behave rudely toward a different partner who had done nothing.

Merely witnessing rudeness primed observers to detect rudeness faster afterwards, and the effect persisted a week later. Trevor Foulk: "When you experience rudeness, it makes rudeness more noticeable.

You'll see more rudeness even if it's not there… Rudeness will flavour the way you interpret ambiguous cues." So one act does not merely land on its target; it reshapes how everyone present reads the next ambiguous exchange, which makes the next uncivil act more likely and more likely to be perceived as normal.

Bystanders pay a performance cost too, which is why neutrality is not available. Porath's experiments found that people who merely watched rudeness performed 22% and 28% worse on the two tasks. There is no uninvolved position in a room where this is happening — everyone's judgement is degraded, including that of the person who said nothing.

Put together, the compounding works like this. One small act, ambiguous enough that challenging it would seem disproportionate. It goes unremarked, which is itself information: this is tolerated here.

The target, now depleted, is measurably more likely to pass it on to someone uninvolved. Witnesses are primed to see hostility where it may not exist, and to read the next borderline act as normal. The threshold for what counts as a breach moves.

And each act that goes unchallenged licenses something slightly further — until the behaviour that would have been unthinkable at the start is unremarkable, and the person who finally objects is the one who looks unreasonable.

That is precisely the state every inquiry in this document describes on arrival.

It is also why don't be a bystander is not a moral slogan but the operative intervention: the spiral is broken at the small end, cheaply, by someone naming the first thing — or it is broken expensively, years later, by a public inquiry.

The Active Bystander approaches now used in parts of the NHS work on exactly this logic, and the evidence above is the reason they are worth more than they look.

The scholarly foundation: Staub, Waller and Fein

The continuum argument is not an analogy borrowed from atrocity studies and stretched to fit a hospital. It is the same mechanism, described by the people who study its extreme end, and two of them are worth knowing by name.

Ervin Staub named this. In The Roots of Evil (Cambridge, 1989) and subsequently, he set out the continuum of destruction: perpetrators do not arrive fully formed. They are changed by what they do. His formulation in The Psychology of Good and Evil (Cambridge, 2003, ch. 23, "Steps Along a Continuum of Destruction: Perpetrators and Bystanders") is as close to a one-line statement of the compounding problem as exists: "Once perpetrators begin to harm people, the resulting psychological changes make greater harm-doing probable."

He identifies the enabling mechanism as devaluation, often via the just-world belief — the assumption that "victims have earned their suffering by their actions or character". Anyone who has heard a ward describe a difficult patient, or a board describe a persistent complainant, will recognise the move.

Britain has walked this path before: the 1847 judgment of the Welsh, the eugenics of the early twentieth century and the long-stay hospitals the NHS inherited all began with contempt for a group before they became harm. Annex 11 · Why a top-down model fails

Two things make Staub indispensable rather than merely apposite.

The first is that his mechanism runs both ways: "learning by doing can change people in either positive or negative directions." Repeated acts of care build the capacity for care exactly as repeated acts of harm erode it.

That single sentence is the strongest available warrant for treating humanity as a trainable capability rather than a fixed trait — and it comes from a man who survived Nazi-occupied Hungary as a child, protected by Maria Gogan, whom he describes as "a hero in every possible way". His conclusion from that: "even in the worst of times, people can be caring and helpful."

The second is that he identified the counterforce, and then built it. On the spiral, his finding is that "early public reactions can counteract these changes and inhibit further violence" — which is the evidential basis for intervening at the small end.

He then turned it into professional training. After the Rodney King beating in 1991 he developed what became EPIC (Ethical Policing Is Courageous) in New Orleans, now the ABLE Project run through Georgetown Law, described as "a training that is also a vehicle for cultural change" and mandatory from leadership to rank-and-file.

More than 430 US law-enforcement agencies now use active bystandership training built on his framework, with related work in Rwanda, Burundi, Congo, Amsterdam and Serbia.

That precedent matters for healthcare more than any statistic in this volume. Research on how ordinary people come to participate in atrocity has already been successfully translated into a scalable behavioural intervention for a high-stress, hierarchical, high-consequence profession with a defensive culture and a history of closing ranks.

Policing has done it. Healthcare, which shares every one of those characteristics, largely has not — Active Bystander training exists in pockets of the NHS, but nothing of ABLE's scale or mandate. Aviation and industry have gone further again, making protection for the person who speaks conditional on speaking early: Annex 13 · Radical leadership.

James Waller supplies the other half: the perpetrators are ordinary. In Becoming Evil: How Ordinary People Commit Genocide and Mass Killing (Oxford, 2002; second edition 2007), Waller — now Christopher J. Dodd Chair in Human Rights Practice at the University of Connecticut — rejects the explanation that atrocity requires monsters, and builds a model around the cultural construction of worldview, the psychological construction of the "other", and the social construction of cruelty, in which professional socialisation, group binding and the merging of role and person do the work that individual pathology cannot explain.

Read his chapter titles next to an NHS inquiry and the overlap is uncomfortable. "Social Construction of Cruelty: The Power of the Situation" is, in substance, what Francis found at Mid Staffordshire, what Kirkup found in East Kent, and what Aubrey found at Durham: not cruel people, but a situation in which decent people behaved in ways they would not otherwise have contemplated, and stopped noticing.

Helen Fein supplies the boundary condition. Writing in Accounting for Genocide (Free Press, 1979) about why some nations protected their Jewish populations and others did not, the sociologist Helen Fein described a society's universe of obligation: the people "toward whom obligations are owed, to whom rules apply, and whose injuries call for amends".

Where Staub explains how harm escalates and Waller explains how ordinary people come to take part, Fein explains who it happens to — whoever has fallen outside the circle.

The organisational version of this is exact and testable: in any hospital some interests sit inside the circle and some outside, and you can tell which by watching what happens when they conflict. When the unit's reputation is inside it and the patient's own account of what happened is outside it, harm has somewhere to go and no one who could stop it feels obliged to.

Volume 3 applies this to Chester, where the gap was wide enough for a deliberate perpetrator to work in it for a year after the alarm was raised.

The practical conclusion from both, and the reason this volume sits where it does: if harm at the far end of the continuum is produced by situations acting on ordinary people, then so is harm at the near end — and the near end is where it is cheap to intervene.

Staub's counterforce is a colleague who says something early. Everything else in this document is what happens when nobody does.

The clinical evidence: civility is a performance variable

Incivility measurably degrades clinical performance. The foundational study is Riskin et al. (Pediatrics, 2015), in which neonatal intensive care teams in simulation were exposed to mildly rude comments.

Diagnostic accuracy, information sharing, help-seeking and procedural performance all deteriorated significantly — by 15–17% on the team's information-sharing and help-seeking measures.

Related work reports clinicians performing at expected levels 91% of the time in a normal environment, falling to 64% when exposed to rudeness (Katz et al., 2019). This is the crucial finding for anyone making the business case: civility is not etiquette, it is a performance variable.

Exposure is high, normalised, and it reaches patients. The Institute for Safe Medication Practices has surveyed US practitioners on this for two decades. Its 2021 survey (1,047 respondents) found 79% had personally experienced disrespectful behaviour and 60% had witnessed it, with only 5% reporting neither; 64% had been on the end of condescending or demeaning comments; and 75% said their organisation does not deal with it effectively.

The figure that matters most for safety: 27% were aware of a medication error in the previous year in which disrespectful behaviour played a part — up from 11% when the same question was asked in 2013. That is the mechanism by which culture becomes harm, measured directly.

(Note on a figure in wide circulation: the "88% experienced intimidating behaviour" line is not from the 2013 ISMP survey, whose own figures were 77% for reluctance or refusal to answer questions and 68% for condescending language. Use the 2021 numbers.)

The cognitive step in that chain is experimentally established rather than inferred: rudeness impairs working memory, and the impaired working memory degrades the task. The step from incivility to patient harm is strong association, and this document says so.

In Porath and Erez's third study (Academy of Management Journal, 2007), participants exposed to rudeness performed significantly worse on a paired-associate word task — a standard measure of working memory capacity — recalling 11.02 items against 13.06 in the control condition.

Crucially, that impaired recall mediated the effect of rudeness on anagram solving, creative output and cognitive flexibility. Rudeness degrades cognitive processing, and the degraded processing is what degrades the work.

Porath states the effects plainly: those belittled before a task "performed 33 percent worse" and produced "39 percent fewer creative ideas", while people who merely witnessed the rudeness fared "22 percent and 28 percent worse".

In the underlying studies, recipients were also 74–84% less likely to help a colleague — the figure that matters most in a hospital, where help-seeking and information-sharing are the safety mechanism.

The peer-reviewed position is now settled enough to cite directly. CMAJ, 2024: more than 75% of healthcare employees have witnessed uncivil behaviour from physicians, and 31% of physicians report receiving rude, dismissive or aggressive communication from other doctors weekly or daily.

Disruptive behaviour is associated with medication errors, patient neglect, surgical complications and death.

Incivility is directed disproportionately at trainees who are younger, junior, female, non-native speakers, or from a non-dominant ethnicity — which is the same population whose concerns the inquiries found were not heard.

The NHS measures this better than almost anyone — and the direction is finally right, from a bad place. The NHS Staff Survey covers around 766,000 respondents from a workforce of roughly 1.5 million. In 2025 it recorded 17.01% experiencing bullying, harassment or abuse from colleagues and 9.11% from managers — both the lowest figures on record, against 23.2% on the colleague measure in 2023 — with discrimination from colleagues or managers at 8.77%, also the lowest recorded.

Two things should be said about that honestly. It is real movement, and it deserves to be acknowledged. And it still means roughly one in six staff was bullied, harassed or abused by a colleague in a year, 25.25% were abused by patients, service users, relatives or the public, and 29.55% often think about leaving — a three-year high. The culture measures are improving; the pressure measures are not.

This is not a uniquely British disease — which is the point, not the excuse. Australia: across seven metropolitan tertiary hospitals (5,178 respondents), 93.6% had experienced at least one unprofessional behaviour, 38.8% experienced incivility or bullying at least weekly, and 49.8% said unprofessional behaviour had a moderate or major negative effect on patient care, error rates or quality of service (MJA, 2021).

Its national Medical Training Survey 2025 (17,622 doctors in training) records 20% experiencing and 26% witnessing bullying, harassment, discrimination or racism — rising to 56% among Aboriginal and Torres Strait Islander trainees.

New Zealand: 37.2% of senior medical officers self-reported being bullied and 67.5% had witnessed a colleague being bullied (ASMS, 2017). The Royal Australasian College of Surgeons found 38.7% of surgeons and trainees had been bullied (2015).

And the pooled international figure for medical residents is 51% (95% CI 36–66) across 13 studies and 44,566 participants (Journal of Occupational Health, 2023).

What distinguishes England is the quality of its measurement and the visibility of its inquiries, not the severity of its culture — and that means every health system reading this document has the same problem, whether or not it has the data to see it.

Why it spreads. Turner's account of the dynamic explains why incivility propagates rather than staying put: "It's like an emotional muscle. Once we start to display rudeness, incivility, it gets easier and easier to accept."

His framing of the evidence is worth borrowing, because it pre-empts the obvious objection — he describes himself as "a dyed-in-the-wool, evidence-based practitioner" who does not expect to be taken on trust, and concludes that "every bit of evidence that we have at this moment says the same thing: it says the way we behave towards each other counts."

What works, and what does not. Single training events have little durable effect. The interventions with evidence behind them are multi-component, organisation-wide, leadership-led and sustained — the Canadian CREW programme (Civility, Respect and Engagement at Work), run across roughly 2,000 healthcare workers, produced sustained improvements in civility and reductions in supervisor incivility and staff distress.

Organisational leadership is the decisive variable: leadership training, role modelling, clear institutional definitions, workable reporting, and a consistent response to poor behaviour — escalating to disciplinary action where a pattern proves unresponsive.

The implication for leaders is uncomfortable: incivility is a systems and culture problem, not a personality problem, and what leaders tolerate is what becomes the culture.

Small-scale evidence suggests the effect sizes are worth chasing. The Union Health Service programme in Chicago — roughly 250 staff across six sites serving 40,000 members — recorded after 90 days that all interviewees saw increased empathy and self-regulation, 90% saw improved cross-team collaboration, 85% reduced emotional reactivity, and 70% visible improvements for patients, from a starting point of low morale and a 1.6 Yelp score.

This is a single non-randomised case study in an outpatient setting and should be read as indicative rather than generalisable, but it points the same way as CREW.

Why humanity should be at the heart of medicine: caring for the carers

Technical brilliance does not make medicine humane — 1930s Germany proved it at the limit. Values give way under pressure unless they are deliberately formed and protected, and in today's NHS that pressure falls on the carers.

The one thing to take away: humanity has to sit at the heart of medicine, and it reaches patients only if it reaches the people caring for them first.

Expertise alone cannot make medicine humane. Humanity has to be put at its heart, and that has to begin with how the people who give care are cared for themselves. This volume makes the case in three steps.

History shows that the most technically brilliant medical profession in the world can abandon its humanity entirely. The Lancet Commission that examined that history concluded that medical values are fragile under pressure and have to be deliberately formed and protected.

And the pressure falls first on the carers: a doctor or nurse who is exhausted, isolated and unheard cannot go on giving what they are not given.

Medicine already has the machinery to form and protect a skill. What it has not yet done is treat humanity, including its own people's, as part of the job.

Research synthesis

What. Technical excellence is not enough, and history has run the experiment at the limit. Medicine has the most competitive entry, the longest training and the most elaborate systems of examination, revalidation and audit in working life, and all of that apparatus is pointed at one half of the job.

1930s Germany led the world in medical science, and its physicians joined the Nazi party in greater proportion than any other profession, nearly 50% by 1945, and were seven times more likely to join the SS than other employed German men.

The Lancet Commission on Medicine, Nazism, and the Holocaust (November 2023) concluded that this was "corrupt moral agency" meeting dangers it calls inherent to modern scientific medicine.

Nearer home, the Lord Mann review (4 June 2026) found "racism is deeply pervasive and has reached the stage where it has become normalised in the NHS".

So what. Humanity has to be at the heart of medicine, and it has to start with the carers. Technical excellence does not produce humanity, and can supply the instruments for its opposite. The danger is not inherent to science; it is inherent to any human system not rooted in values of humanity.

The Commission's own finding is that those values are "fragile" and change "particularly when placed under pressure" — and in today's NHS the pressure lands on the people giving care. Half of the doctors surveyed by the BMA in 2021 reported moral injury, and fewer than half of NHS staff feel their organisation values their work.

People cannot keep giving what they are not given: an organisation teaches its staff how to treat patients by how it treats them.

Now what. Form humanity as deliberately as technique, and protect the people who carry it. The Commission's remedy is not compliance but history-informed professional identity formation: a settled sense of what kind of professional one is, able to hold under pressure.

That needs two things medicine has not yet done. It must treat humanity as a clinical competency, taught, practised and assessed like any other. And it must treat the wellbeing, belonging and voice of its staff as a patient-safety measure, not a perk.

A profession that standardised the surgical checklist worldwide and revalidates every doctor on a five-year cycle has the machinery to do both. Being the best of the best technically is settled. The open question is whether medicine puts humanity — its patients' and its own people's — at its heart.

Technical excellence is not enough

Medicine is the best in the world at technical skill and has left its humanity to chance. History shows why that is dangerous, and why expertise alone will never be the answer.

Medicine is, by common consent, the summit of applied human technical skill. Entry is the most competitive of any profession. Training is the longest. The knowledge base is the largest and fastest-moving.

The tolerance for error is the lowest, the consequences of error the gravest, and the systems of examination, revalidation, appraisal and audit the most elaborate anywhere in working life. When a society wants a shorthand for the highest attainable human competence, it reaches for the surgeon.

And all of that apparatus is pointed at one half of the job.

And history has already run this experiment, at the limit. Germany in the 1930s was the undisputed world leader in medical science — its laboratories, its universities, its Nobel laureates, its physicians described by one historian of the period as "already the best in the world".

It was also the setting for the most complete collapse of medical humanity on record, and the medical profession was not a bystander to it. On Robert Proctor's figures, physicians joined the Nazi party "in droves (nearly 50% by 1945), much higher than any other profession", and were seven times more likely to join the SS than other employed German males.

That is the fact that disposes of any assumption that technical excellence implies humane practice. The most scientifically advanced medical profession in the world was also the most politically complicit in genocide.

Excellence in the one did not produce the other, and there is a case that it helped enable it — the authority, the language of science, and the habit of categorising people supplied the instruments.

The Lancet Commission, and one word that needs care

This is the subject of the Lancet Commission on Medicine, Nazism, and the Holocaust, published in November 2023 — the first Lancet Commission devoted to the history of medicine, convened by editor-in-chief Richard Horton, with twenty scholars and a student advisory council drawn from ten countries.

Its central finding is not about monsters, and it should be read alongside Staub and Waller in Volume 5:

"The central insight from the history of medicine during Nazism and the Holocaust is that the atrocities that health professionals committed during the Nazi reign and the Holocaust represent, to a large degree, the outcome of corrupt moral agency in the face of potential dangers that are inherent to modern, scientific medicine as it emerged in the 19th century."

A necessary note on that word. "Inherent" is the Commission's own term, and it is easy to misread as inseparable — as though scientific medicine were permanently and irredeemably dangerous.

That is not what follows from it, and the Commission's own other messages show why: it says the core values of healthcare are "fragile and need to be protected", that they "require constant critical assessment and reinforcement", and that "courage, resistance, and resilience are necessary to prevent and counteract" abuses of trust and power. A danger that can be prevented and counteracted is latent, not inevitable.

The more exact formulation is this. The danger is not inherent to science. It is inherent to any human system that is not rooted in values of humanity. Scientific medicine is a case where the stakes are unusually high, because it concentrates authority, it categorises people as a matter of method, and it acts directly on bodies.

Those features amplify the consequences of value-less organisation; they do not create the flaw, and they do not make it permanent. Which is exactly why the remedy is structural rather than resigned — and why every institution in Volume 8 is an existence proof that the same science, practised inside different arrangements, produces something else.

The Commission's other key messages follow from that, and read as a direct brief for the argument of this document:

• "The core values and ethics of health care are fragile and need to be protected. They require constant critical assessment and reinforcement."
• "Courage, resistance, and resilience are necessary to prevent and counteract potential abuses of trust, power, and authority in health care."
• Values and ethics in healthcare are "not immutable but change over time dependent on cultural, social, economic, and political factors (and particularly when placed under pressure)" — and therefore must be "critically assessed and reinforced to protect them from potentially becoming exclusionary and inhumane".
• "Health professionals have particular responsibilities in fighting against antisemitism, racism, and other forms of discrimination."

The phrase "particularly when placed under pressure" is the whole of this document in five words. But it needs one refinement, because the Commission's wording — values "change over time" — implies a gradual wearing away, and that is not what the inquiry record actually shows.

Pressure does not erode values. It exposes them, and reveals what was holding them up. Where values are genuinely anchored — in a deed of gift, a salary policy, a constitution, a practised habit — pressure meets resistance and the line holds.

Where they are held only by statement and good intention, pressure finds nothing to push against, and the failure is not gradual at all. The wall breaches, and what comes through is a flood: the licensing cascade described in Volume 5, in which each unchallenged act permits the next until behaviour nobody intended has become ordinary.

That distinction matters because it locates the work in the right place and at the right time. If values eroded, the remedy would be periodic renewal — another set of statements, a refresh of the values programme, which is roughly what the NHS has repeatedly done.

If pressure instead exposes whether anything structural was ever there, the remedy is to build the wall before the water rises, and to hold the line when it does.

Elizabeth Fry is the demonstration, and instructively so. When the Anglican Church and Queen Adelaide objected to her naming the body the Protestant Sisters of Charity, she changed the name and kept the institution — conceding the label and holding the substance.

And she did not rely on holding it personally: she put it into a standing institution, a training relationship at Guy's, and an Act of Parliament, so that it would go on holding after she had gone. That is what a wall looks like, and it is the difference between a value that survives pressure and one that is merely professed.

What has and has not changed since the 1930s. It would be wrong to say nothing has changed: the Nuremberg Code, informed consent, research ethics, human rights law and modern professional ethics were all built substantially in response.

But two things the Commission documents have not changed, and they are the ones that matter here. The vulnerability is structural and remains. And the lesson has largely not been institutionalised in how clinicians are formed: "health sciences curriculums, however, rarely cover this topic".

Eighty years on, the most thoroughly documented case of medical complicity in atrocity is still not a standard part of medical education anywhere.

The Commission's first recommendation is that it should be mandatory worldwide: every training programme for health professionals should "include learning about the history of medical involvement in Nazism and the Holocaust", with emphasis on "the responsibility of medical professionals and their institutions to uphold human rights", and accreditation bodies should write it into their requirements.

It is explicit that this is not a German or a British problem: "virtually every debate about health professional ethics can gain from an understanding of this shameful history — from questions regarding the beginning and the end of life, to health professionals' roles as economic actors or as agents of the state."

Its proposed remedy is worth noting precisely because it is a formation argument rather than a compliance one. The Commission calls for history-informed professional identity formation — "the shaping of professional identity through learning about, and reflecting on, historical episodes in which medical values and priorities were challenged" — with the aim of producing "morally conscious and self-critical, yet courageous and resilient health professionals — independent thinkers who are capable of upholding professional values in the face of pressure and who will, when needed, act as agents of change".

Note the word the Commission chose: identity. Not compliance, not governance, not training — professional identity formation. Its judgement, after examining the most extreme failure of medical humanity on record, is that the durable protection is not a rule but a settled sense of what kind of professional one is, formed deliberately and capable of holding under pressure.

That is the same conclusion Volume 8 arrives at from the opposite end, and it is the layer beneath every structure described in this document.

The Commission's own summary of what the history shows cuts both ways, and is the better note to end on: it demonstrates "the potential for health professionals to harm their patients, but also, when necessary, to stand up to power and protect the most vulnerable".

Caring for the carers

Values give way under pressure, and in healthcare the pressure falls on the people giving care. The Lancet Commission's phrase — values change "particularly when placed under pressure" — describes the modern NHS ward as well as 1930s Germany, in a far milder register.

The annex on doctors and nurses sets out what the pressure looks like: teams broken up by rotation, staff asked shift after shift to give care they know is not good enough, and a workforce that has learned it is on its own. Annex 6 · Doctors and nurses: when the organisation stops caring

The evidence on the state of the carers is stark. In a 2021 BMA survey of about 2,000 doctors, 78% reported moral distress and 51% moral injury, with too few staff the leading cause.

In the 2025 NHS Staff Survey, 31% of staff reported feeling burnt out because of their work and only 43% felt their organisation valued them; a record low of 63% would be happy with the standard of care if a friend or relative needed it.

One doctor in training in five is at high risk of burnout (GMC, 2025). And nearly half of the nurses and midwives leaving the register who answered the regulator's 2024/25 survey had left earlier than they planned.

What the carers need has already been written down. The GMC's 2019 review Caring for doctors, caring for patients, by Michael West and Denise Coia, found that "changes in many organisations, such as the loss of team structure, have left them feeling exposed, with increases in bullying, blaming and undermining".

It set out three core needs for doctors' wellbeing and motivation: autonomy, belonging and competence. None of them is expensive. All of them are organisational.

Belonging is the one the modern NHS has taken away most completely, and it matters more than it sounds. Resident doctors are mostly in their twenties, the age group that now reports the most loneliness, and they are moved between hospitals and teams every few months.

No UK survey yet asks NHS or care staff whether they feel alone. Annex 8 · Loneliness, the silent killer

Caring for carers is not a perk. It is how patients get humane care. Across the NHS, trusts with more engaged staff have lower patient mortality (West and Dawson, King's Fund, 2012). Staff who attended Schwartz Rounds regularly saw psychological distress roughly halve.

And laboratory research suggests the answer to burnout is not to care less but to care differently: training in compassion — warm concern and the wish to help — increased positive feelings, where training in empathy alone, feeling another's pain, increased distress (Klimecki, Singer and colleagues).

Some of this evidence is stronger than the rest, and the annex on why being human is good for health weighs it. But the direction is consistent: people who are cared for give better care. Annex 7 · Why being human is good for health

Which is why this cannot be left to individual resilience. Telling exhausted staff to be kinder is the wall poster again. What holds is what holds everywhere in this document: an organisation that decides its own people are inside its circle of obligation, and builds teams, rotas, rest and voice to match. Humanity passes from person to person. It reaches the patient only if it reaches the carer first.

Discrimination is the same problem in another register, and it reaches patients

The Lord Mann review, published 4 June 2026 after being commissioned in October 2025, reached a conclusion in the same language the inquiries use about safety culture: "racism is deeply pervasive and has reached the stage where it has become normalised in the NHS".

Three findings matter for the argument here. First, the source of the discrimination is largely internal: Jewish staff reported discrimination from managers and colleagues at levels identical to those reported by Black African and Indian staff, and higher than what they experience from patients — 11% of Jewish and 16% of Muslim staff had faced discriminatory behaviour from patients or the public in the preceding twelve months.

Second, it produces measurable patient harm, with Jewish patients describing "real fear in seeking care" and considering removing religious symbols before seeking treatment; groups recording their religion as Judaism show "a much lower than average level of trust and confidence in the doctors and staff treating them".

Third, and most familiar, the review found "a sorely lacking response to concerns raised about racism from the NHS as an employer" — the same failure to act on what staff report that Thirlwall, Ockenden and Aubrey each describe.

It also found the regulators wanting, holding that the Medical Practitioners Tribunal Service "has failed to evaluate where it does not have sufficient expertise to determine cases concerning allegations of antisemitism".

The review made 24 recommendations; the government published its response the same day, committing to act on those directed at the Department and NHS England, supporting those aimed at independent regulators in principle, and reporting progress to Parliament by October 2026.

Note the shape of that response — it is the recurring limit of NHS reform, in which the parts of the system that most need to change are the parts government can only encourage.

And this is not a healthcare problem that happens to involve society; it is a societal problem that healthcare reflects. Mann's NHS review followed his wider work with Dame Penny Mordaunt on a commission for the Board of Deputies of British Jews, whose finding was that Jewish people are "tolerated rather than respected" in the UK, with rising incidents across the NHS, universities and the arts — death threats and bomb hoaxes against students and staff, antisemitic tropes taught in some primary religious education classes, artists excluded on grounds of heritage, and inconsistent police enforcement.

By September 2026 the rise in antisemitism was the subject of sustained parliamentary concern.

That context cuts both ways for the argument of this document. It means the NHS cannot be held solely responsible for attitudes its staff and patients bring through the door, and it means the NHS is one of the few institutions positioned to do something about them at scale — 1.5 million employees, and near-universal contact with the population.

"Tolerated rather than respected" is also, precisely, the register in which every inquiry in this document describes how concerns were received: heard, logged, not acted upon, and the person raising them managed rather than valued.

Humanity is a clinical competency, on the profession's own standards

The evidence in Volume 5 says the other half of the job — how people treat each other while doing the technical work — is not a soft adjunct to clinical quality but a determinant of it. Rudeness measurably degrades diagnosis.

Incivility makes colleagues 74–84% less likely to help. In the most recent survey of US practitioners, 27% knew of a medication error in the previous year in which disrespectful behaviour played a part (ISMP, 2021).

Every major inquiry in this document found the culture failing before the care did. On the profession's own evidentiary standards, humanity is a clinical competency.

So the question is not whether medicine can put humanity at its heart. It is why it hasn't. A profession that can standardise a surgical checklist across the world, revalidate every doctor on a five-year cycle, and audit outcomes by individual operator has never seriously applied that same machinery to how people behave towards one another — despite having better evidence for the effect of behaviour on outcomes than it has for many interventions it mandates without hesitation.

There is no reason in principle why this should be the hard part. The profession already accepts that a skill can be taught, assessed, examined, revalidated and audited; it does this with cannulation and with consent.

Staub's finding that "learning by doing can change people in either positive or negative directions" says the same thing about humanity that medicine already believes about technique.

Policing, with lower entry standards and less training time, has taken genocide research and built mandatory bystander training across 430 agencies. Mayo wrote its values into its bylaws and its pay structure.

Aga Khan Nairobi seeks accreditation no regulator requires. None of that was beyond those institutions, and none of it is beyond a profession that routinely does harder things before lunch.

The positive case is stronger than the remedial one. Healthcare is not merely a sector with a culture problem to fix — it is the natural place for a society to demonstrate what humane, high-performing institutions look like.

It has the moral licence, the public trust, the scale and, in the NHS, the reach into nearly every family in the country. If medicine set the standard for how people treat each other under pressure, the standard would travel, in the way clinical standards have travelled into aviation, engineering and beyond.

Which reframes the whole document. The inquiries describe a profession that reached the top of human technical achievement and left its humanity to chance. That is not an indictment of the people in it — the evidence is unambiguous that they are ordinary people in demanding situations.

It is an observation about where the profession has chosen to point its formidable capacity for rigour. Being the best of the best at technical skill is settled. The open question is whether medicine chooses to be the best of the best at being human — towards its patients and towards its own — and it already has every tool it needs to do it.

Comparisons

Every developed health system reports the same culture problems; England's are unusually well measured, not unusually bad. Money and funding models do not settle it — failures happen in private hospitals too — and some systems, like Israel's, do better with less.

The one thing to take away: no country has cured this, so the question for England is not what it can afford but what it chooses to build.

Healthcare is sick in much of the world, and England is no exception — but nor is it a lost cause. Across the developed world the same symptoms appear: staff bullied and burnt out, challenge suppressed, patients harmed by failures of teamwork rather than of knowledge.

About half of medical residents internationally report being bullied or harassed; 79% of American practitioners have experienced disrespectful behaviour at work; Japan's doctors have literally worked themselves to death; and in a private London hospital where money was no object, an 80-year-old woman died of a blood clot that routine care would likely have prevented.

England shares the illness, and on waiting times and capacity it has it badly. But the comparison also shows the way out. Israel gets better results than England with fewer beds and less money. Mayo, Nuka and Buurtzorg have sustained humane care inside ordinary budgets, and Leeds and Manchester have done it in English social care.

No country has cured this, so no country is hopelessly behind — and the places that are well show both that there is hope and what works.

Research synthesis

What. England is not exceptional — it is just better at measuring itself. Across the four nations, between 10% and 25% of staff report bullying or harassment and considerably more witness it; England simply publishes most.

Northern Ireland is the domestic worst case — 504,537 waiting for a first outpatient appointment at June 2026, 53.7% of them over a year, ED performance around 31%, on a budget baseline falling from £8,898m ($12.0bn) to £8,481m ($11.4bn).

It has also run health and social care as one system since 1973, the clearest proof that merging the organisations does not, on its own, make care better. Annex 9 · Joining up health and social care: what integration can and cannot do

Internationally, no system has solved this: the US combines world-leading capability with 79% of practitioners reporting they have personally experienced disrespectful behaviour; Japan buys excellent access with karoshi; Germany outspends the UK with roughly triple the bed capacity and different perverse incentives.

So what. Money does not settle it, and neither does the funding model. Hierarchy plus high consequence plus professional tribalism is associated with the same suppression of challenge everywhere, whoever is paying.

And when money is genuinely no object it still happens: Lady Joan Branson died in a private London hospital in November 2025 of a pulmonary embolism after routine clot prophylaxis was not given — the coroner found her death "would have likely been prevented".

No beds, scanners or waiting lists were the constraint. What failed was the reliability of human coordination.

Now what. Stop treating this as a British disease with a funding cure. England's advantage is that it knows: twenty years of the best culture data in world healthcare. The indictment is that measurement has not been matched by improvement — and the thing that distinguishes the genuine exceptions, in Volume 8, is not their funding model.

The four nations

Direct comparison is difficult because each nation measures culture and performance differently. England has by far the most mature dataset; the others are not necessarily healthier, only less visible.

Across the UK as a whole the picture is consistent: somewhere between 10% and 25% of staff report bullying or harassment, considerably more witness it, and a significant minority do not believe raising a concern will change anything. Every major inquiry identifies failures of culture before failures of competence.

The reasonable reading is that England's apparent exceptionalism is partly an artefact of measurement. It investigates more, publishes more, and therefore looks worse. That is not a defence — twenty years of measurement without commensurate improvement is its own indictment — but it should temper any conclusion that English healthcare culture is uniquely bad.

On the available evidence it is unexceptional among Western systems, and unusually well documented.

International context

England is not an outlier. Set against other systems, its combination of clinical competence and institutional defensiveness is ordinary — which is either reassuring or damning, depending on your view of what ordinary should be.

What the comparison does reveal is that the exceptions, the institutions that genuinely sustain humane care at scale, have something specific in common, and it is not their funding model.

The good, the bad and the ugly, by country.

The honest summary: no system has solved this. Every developed system reports a substantial minority of staff experiencing bullying or intimidation — the pooled international figure for medical residents is 51%, and hierarchy plus high consequence plus professional tribalism is associated with the same suppression of challenge everywhere. What differs is whether anyone measures it.

Israel, and the limits of what it proves

Israel is the comparator England should be studying, because it is short of precisely what England is short of and gets better results anyway. England has among the fewest beds and scanners of comparable systems and treats that as the explanation for its performance.

Israel has fewer: 1.76 acute beds per 1,000 people against an OECD average of 3.18, and 5.57 nurses per 1,000 against 9.5. It spends 7.3% of GDP on health against an OECD average of 9.3%. And it reaches a life expectancy of 83.8 years, fourth among high-income countries. Whatever is carrying that, it is not capacity and it is not money.

What Israel does not prove. It does not prove that culture substitutes for capacity, and nothing here should be read as saying so. Israel's population is younger than the UK's, which flatters life expectancy and demand alike; about a third of health spending is out of pocket, which is high for a universal system and shifts cost onto households rather than removing it; and the headline figure conceals a substantial gap between the Jewish and Arab populations — 84.3 years against 80.7, with infant mortality of 1.9 against 5.2 per 1,000.

Bed occupancy was 89% in 2021 against an OECD average of 82%, which is high, but England's general and acute occupancy averaged 90.2% in 2019/20, so this is not a difference in pressure.

Nor do hospital rankings settle anything: Sheba Medical Center tops Newsweek's 2026 Israeli table and the hospital says it placed seventh in the world, but 53.5% of that score is expert opinion and patient-experience survey, and the patient-reported-outcomes component measures whether a hospital has implemented such measures, not what they show. It is a reputation index, not an outcomes study.

The structural feature worth taking seriously is who owns the hospitals. No single owner dominates and almost nobody owns them for profit: of acute beds, government holds about 45%, the largest health fund 30%, non-governmental organisations 22%, and private for-profit providers about 3%.

Among those non-governmental owners are Jewish-founded charitable institutions — Hadassah, Shaare Zedek — and Christian mission hospitals, side by side. This is the arrangement Volume 8 identifies as the one thing the bright spots have in common: an owner whose stated purpose is something other than margin.

And the founding ethic is real, but has to be quoted accurately. Jewish law treats the saving of a life, pikuach nefesh, as overriding almost every other religious obligation, including the Sabbath — the principle is derived in the Talmud at Yoma 85b from the statutes which a person "shall do and live by them, and not that he should die by them", with the practical rulings at 84b. The three cardinal prohibitions are the exception.

The familiar English maxim that whoever saves a single life saves a whole world is from Mishnah Sanhedrin 4:5, and it comes with a genuine textual variant: the standard printed Babylonian Talmud reads "one soul from the Jewish people", while the Mishnah tradition reads simply "a single life".

The universalist form is real, but it is a reading, and this document says so rather than flattening it.

One further caution before this is used as an argument. No institution's mission statement found in preparing this volume invokes pikuach nefesh by name, and there is no evidence that faith-founded hospitals outperform comparators — see Volume 8 and the Sources volume for the one recent study, which points the other way and cannot bear the weight.

What the Israeli case supports is narrower and more useful: a system can be built on an explicit obligation to the person in front of you, and it can then hold up under capacity constraints worse than England's.

When money is no object

One more comparison belongs here, because it removes the most common explanation. If underfunding and capacity were the whole story, private healthcare would not fail in the same way. It does.

Lady Joan Branson, 80, fell while abroad and sustained a fractured pelvis and right hip. She was admitted to the Lister Hospital in Chelsea, a private HCA Healthcare facility, on 17 November 2025, and died there on 24 November 2025 of a pulmonary embolism arising from a deep vein thrombosis.

At the inquest concluding in September 2026, the Senior Coroner for Inner West London, Professor Fiona Wilcox, found: "I do find her death would have likely been prevented and the lack of anticoagulation did cause or contribute to her death probably."

She had not been given prophylactic anticoagulants despite being at high risk of clotting after a fall and immobilisation, and the required bleeding risk assessment was not completed. There is no plausible reading of this in which money, beds, scanners or waiting times were the constraint.

Two things make it instructive rather than merely sad. First, UK private hospitals are largely built around elective, planned, single-specialty work; complex trauma in an elderly patient with multiple interacting risks is not what the model is designed around, and the routine prophylaxis that a busy NHS orthopaedic ward performs reflexively was missed.

Second, the failure was one of teamwork and shared responsibility rather than knowledge — the doctor involved reflected afterwards on the importance of "good communication between team members". Nobody in that hospital was unaware that immobilised elderly patients need clot prophylaxis. The system around them did not ensure that somebody owned it.

That is the same mechanism as every case in this document, stripped of the excuse of scarcity. Resources are necessary and not sufficient; what fails is the reliability of human coordination.

And there is a hopeful reading of the same comparison, which Volume 8 takes up. If no developed system has solved this, then nobody is behind — and the institutions that have sustained humane care at scale did it inside ordinary funding constraints, not above them.

Mayo gave away a profitable practice; Buurtzorg removed its own management layer; Intermountain deliberately forwent about $9m a year of revenue to improve obstetric care. What they did is copyable rather than purchasable, which is the only reason this document can end in a proposal.

The comparison also carries a warning about borrowing, which is worth taking before any of it is copied. England has already imported several of the models admired in this volume — Buurtzorg's self-managing teams, Sweden's Esther question — and the imports faded where the form crossed and the function did not.

The mechanism rarely transfers; the principle usually does. Annex 15 · Humanity in healthcare: examples

Bright spots

Humane care at scale has been done, repeatedly, on ordinary budgets — at Mayo, Nuka and Buurtzorg, and in English councils such as Leeds and Manchester. None of them did it by telling staff to care. Each began with a settled answer to who they were for, and then built structures that hold that answer on a bad night.

The one thing to take away: identity first, then values, then structure — and what they did can be copied, not bought.

Research synthesis

What. Humane care at scale has been done, repeatedly, and it is not a funding story. The places that sustain it share no funding model, country or religion: Mayo Clinic (85,000 staff, ~$21.5bn revenue, salary-only pay for over forty years, bylaws reserving leadership to physicians); Nuka in Alaska (Baldrige winner twice, emergency visits down ~50%, staff turnover 37% to 17%); Buurtzorg (70,000+ clients, one coach per 45 teams, 8% overhead against a 25% sector average); the Esther model; Virginia Mason's Patient Safety Alerts, where reporting is compulsory and the line stops; Martha's Rule (12,301 calls to February 2026, 2,310 changes of treatment); Learning from Excellence; Schwartz Rounds; Aga Khan Nairobi, scoring 98.4% on accreditation no regulator requires; Elizabeth Fry, 1840.

And bright spots can go out, which is the other half of the lesson. Salford Royal was rated Outstanding in 2015 and again in 2018; it is now part of a trust under a regulator's warning notice, a few years after the chief executive who held the standard in place retired.

What was written into structure survived, and what was carried by people left when the people did. Annex 12 · Humanity-First Healthcare: applying lessons from the past

And it has been done in England, in the part of the system with the least money: Leeds and Manchester took their children's services from Inadequate to Outstanding, and the government has paid to spread three English council models to others.

So what. None of them works because someone was told to care. Each is an identity — a settled answer to who are we — made operational in arrangements that keep holding under pressure. Mayo gave away a profitable practice; Fry used an Act of Parliament; the Aga Khan network seeks a standard nobody demands of it.

Structure without identity becomes compliance, which is England's actual failure mode: it has a statutory duty of candour, Freedom to Speak Up Guardians, fit and proper person tests and a leadership framework in abundance, and no answer to who are we that staff inhabit.

Now what. Identity first, then values, then structure — in that order. Identity without structure dies with the founder; structure without identity gets ticked off or turned against the people it was built to protect; values without either are the wall poster.

And MSF is the honest warning: strong values do not immunise anyone. What MSF did differently was publish its own failure rather than wait for an inquiry — exactly what Chester, Nottingham and Durham did not do.

Bright spots: what good looks like, and why

The inquiry record tells you what failure looks like. It does not tell you what to build. These are the places where humane, reliable care has been sustained at scale — and they share no funding model, country or religion.

What they do share is an identity: a settled sense of what kind of institution they are, held by the people in them, and then written into arrangements that keep holding it under pressure. Read the mechanisms below as the second half of that sentence, never the whole of it.

The structures matter enormously — they are what stops good intentions depending on a good day — but every one of them started as an answer to who are we, and a structure imported without that answer becomes compliance, which is the thing this document spends most of its length describing.

The eight arrangements, before the examples. Each is an identity made operational — a way of ensuring the institution keeps being what it says it is when it is tired, stretched and under scrutiny:

1. Remove the incentive to do the wrong thing rather than exhorting people not to (Mayo's salary policy).
2. Put clinical judgement in charge by constitution, not convention (Mayo's bylaws).
3. Give the served population real power over the design (Nuka's customer-owners).
4. Push decisions to small, named, accountable teams and starve the supervisory layer (Buurtzorg's coach ratio).
5. Reduce one question to something a junior can ask and use it to override process (Esther).
6. Make raising a concern mandatory, logged and answered on a clock (Virginia Mason's Patient Safety Alerts; Martha's Rule).
7. Study what goes right, not only what goes wrong (Learning from Excellence).
8. Protect time to be human about the work (Schwartz Rounds).

Mayo Clinic — excellence with the incentives removed

Mayo is the most instructive counter-example because its distinctiveness is structural, not exhortatory. Its primary value is "the needs of the patient come first", from William J. Mayo's 1910 formulation that "the best interest of the patient is the only interest to be considered". Three things give that sentence teeth:

• Salary, not volume. A structured salary-only compensation model has run for over forty years with "no incentives or bonus pay", explicitly "to remove financial incentives to do more than is necessary or less than desired for the patient". All cardiologists are paid the same. (The fully flat structure applies after five years on staff.)
• Clinical leadership written into the constitution. Mayo's bylaws require the President to be a physician of at least five years' voting staff, and at least two-thirds of the Board of Governors to be physicians. Compare England's mobile generalist executive class.
• Ownership surrendered. In October 1919 the Mayo brothers gave the practice's entire assets and most of their life savings — over $10m at the time — to convert a profitable partnership into a not-for-profit trust whose deed stipulated that they and all future personnel would take a salary and not profit personally.

Founded by a frontier doctor and the Franciscan Sisters of St Francis, who offered to build a hospital if the Mayos would staff it. Mother Alfred Moes: "The cause of suffering humanity knows no religion or sex; the charity of the Sisters of St. Francis is as broad as their religion." Now roughly 85,000 staff and about $21.5bn of operating revenue — not a boutique model.

Intermountain Health, Utah — and the $9m a year it lost by improving care

The Church of Jesus Christ of Latter-day Saints transferred its fifteen hospitals to a new independent non-profit in 1975 and retained no administrative powers afterwards. Say that precisely, because it matters to this volume's argument: Intermountain has been a secular non-profit for fifty years. The faith connection is in its founding capital, not its governance.

From 1988, under Dr Brent James, it did something almost no health system does — it measured its own clinical variation and treated the result as a question rather than a verdict. James's description of the method is the part worth keeping: "we couldn't determine who was right or who was wrong, but we could sure show that they were different", and what changed behaviour durably was not management instruction but peer normative pressure — "coming from the other obstetricians, that is normative".

On elective induction of labour, inductions not meeting clinical criteria fell from 28% to under 2%.

And here is the fact that earns this case study its place. That protocol took roughly $50m a year out of the cost of care in Utah — but only about $41m out of Intermountain's own costs. Improving care therefore cost the organisation around $9m a year in lost revenue, and it did it anyway (James and Savitz, Health Affairs, June 2011).

That is what an owner whose defined purpose is something other than margin can absorb. It is also, exactly, what an organisation rewarded on activity and finance cannot — which is the whole argument of this document, stated as an accounting entry.

Three caveats, because the rule in this document is to state them. The savings figures are Intermountain's own accounting, published by its own quality officer, who notes the experience "might not be generalizable to all care delivery organizations"; there is no independent peer-reviewed evaluation of its system-wide cost and quality performance.

Utah's unusually low health spending has demographic causes — a younger population, less smoking and drinking — that have nothing to do with management.

And the mission-versus-money tension is not abolished: in June 2026 Intermountain cut 41 jobs in Montana and closed a family medicine clinic citing "long-term financial sustainability", and since its 2022 merger with SCL Health it includes Catholic hospitals operating under the US bishops' Ethical and Religious Directives.

External validation exists but is thin — Fortune/PINC AI ranked it the top large US health system in 2024, its first such ranking.

The transferable feature: an ownership structure that can absorb a financial loss in order to do the right clinical thing, and clinical variation data used as professional inquiry rather than as a managerial weapon.

Nuka System of Care, Alaska — hand the system to the people it serves

Southcentral Foundation's Nuka system is "owned, managed, directed, designed, and driven by Alaska Native people, referred to as 'customer-owners'". It is the only healthcare organisation to have won the Malcolm Baldrige National Quality Award twice, in 2011 and 2017. Each customer-owner has a named primary care team with a panel of around 1,200 people and direct access to it.

The externally examined 2011 Baldrige figures, against a roughly 2000 baseline:

Two honest caveats. Southcentral's own published figures are mutually inconsistent across documents — hospital admissions appear as −53%, −36%, −63% and −31% depending on baseline and source — so treat any single percentage as indicative.

And the peer-reviewed evaluation, which has SCF co-authors, documents real pain in the transition: one physician is quoted saying "I've seen providers cry", and only 67% of interviewed customer-owners supported the coordinated care system.

The transferable part: ownership determines whose preferences count as the design brief, and empanelment with same-day access removes the rationing that pushes people into emergency departments.

Buurtzorg, Netherlands — teams of twelve, no managers

Founded by a former district nurse, Jos de Blok, in 2006. Self-governing teams of 10 to 12 nurses take responsibility for 50 to 60 patients in a neighbourhood — assessment, rostering, care planning and hiring.

There are no managers: around 21 regional coaches support 10,000-plus nurses, roughly one coach per 45 teams, with overhead at about 8% against a 25% sector average. De Blok's formulation: "How do you manage professionals? You don't!"

By 2024 it served more than 70,000 clients with average client satisfaction of 9.2 out of 10, and it was named the best employer in the Netherlands in 2010, 2011 and 2012.

The KPMG evaluation of January 2015 found 108 hours of care per patient year against 168 elsewhere — a 36% reduction — while ranking among the best agencies on patient-reported experience.

But the famous "40% system saving" does not survive that evaluation. KPMG found Buurtzorg's case-mix adjusted total cost per client sat at the 49th percentile, essentially the national average, and subsequent curative care costs were at the 91st percentile.

The efficiency gain within district nursing is real and large; the system-wide saving is not demonstrated.

Replication outside the Netherlands has also been weak — a 2022 scoping review found implementations hindered by "the self-managed working culture, the organizational framework, or national healthcare policies", with staff needing to unlearn dependence on managers.

The transferable part: a coach ratio deliberately too thin to permit supervision. The overhead saving is arithmetic; the care saving comes from nurses being allowed to build patient self-reliance rather than bill authorised hours.

The Esther model, Jönköping, Sweden — one question, asked by everyone

From 1997, Jönköping County ran its integration work through a single question: "What is best for Esther?" Esther is a persona — an elderly woman living alone with several conditions whose care crosses hospital, primary care, home care and social care.

Region Jönköping describes her as "a symbolic person", and its own documents note that "whilst the first 'Esthers' were virtual", real older people were later recruited to be "the voice in the room".

Mats Bojestig, who initiated it: "I think it is very important that we call this work Esther. It helps us focus on the patient and her needs."

From 2006 it added Esther coaches — about 300 of them, deliberately frontline staff rather than managers, mostly nursing assistants and nurses, trained over eight days across eight months, with no extra pay and coaching folded into the day job.

Reported changes include hospital admissions falling from 9,300 in 1998 to about 6,500 by 2013 and heart failure bed days falling roughly 30%.

The caveat is important and comes from the Commonwealth Fund's own account: this "was not designed as a research project", similar declines were reported elsewhere in Sweden, and there is no control group. Attribution of the numbers to Esther specifically is not established.

The transferable part: a named persona converts "integration" from an abstraction into a question a nursing assistant in one organisation can use to challenge a process belonging to another. And the improvement capacity sits with 300 frontline people, not a management layer.

Virginia Mason, Seattle — anyone can stop the line

Virginia Mason adopted the Toyota Production System as the Virginia Mason Production System in June 2002 and built the Patient Safety Alert system from Toyota's andon cord. Its features are the ones England keeps recommending and not achieving: every team member, clinical or not, is required to report immediately any situation that has caused or could cause patient harm; they must stop the line, ceasing the activity; and alerts carry graded response deadlines, reported as 24 hours for the most serious, 72 hours, then one week.

The 50,000th alert was logged in September 2014, and the organisation set a target of 1,000 alerts a month — note the direction of that target. Staff safety-culture survey participation rose from 16% in 2004 to 88% in 2013, and 80% said they would speak up freely.

One correction worth carrying, since the story is usually told wrongly. Mary McClinton, 69, died in November 2004 after a topical disinfectant was injected instead of contrast dye — three identical clear liquids in unlabelled steel bowls.

Virginia Mason disclosed it publicly and named its safety award after her. But her death did not trigger the reform: the production system and the alert system were both in place from 2002, two years earlier. What her death changed was the decision to make safety the organisation's single overriding goal.

The transferable part: an obligation rather than a permission. Speaking up is not a protected right you may exercise; it is a duty with a logged channel and a clock on the response.

Martha's Rule, England — routing round the treating team

The best English bright spot came from a death. Martha Mills died on 31 August 2021, aged 13, at King's College Hospital, of sepsis following a pancreatic injury sustained in a cycling accident. The coroner found prompt referral to paediatric intensivists would likely have saved her.

Her parents, Merope Mills and Paul Laity, campaigned; Martha's Rule was announced by NHS England in February 2024, ran at 143 sites in its first phase from May 2024, and extended to all remaining acute inpatient services from April 2025.

It has three components: 24/7 staff access to a rapid review from a critical care outreach team; the same 24/7 access for patients, families, carers and advocates; and a structured approach to obtaining information about a patient's condition directly from patients and families at least daily.

Over the first 18 months to February 2026 there were 12,301 calls, 4,047 identifying acute deterioration, 2,310 resulting in changed treatment and 524 in transfer to enhanced care. Notably, 1,781 calls came from NHS staff — over a thousand individual staff members using a mechanism designed for families.

Be precise about its limits. It is an NHS England programme requirement, not legislation, and it is not a general right to a second opinion — the entitlement is to review by a critical care outreach team.

No outcome evaluation exists yet; the NIHR summative evaluation only began in June 2026, so current figures are activity counts, not measured mortality reduction.

And awareness is the binding constraint: NHS England's own 2025 survey found 32% public awareness, sharply graded by education and income, while a 2026 study of a mature equivalent service in North Wales found 93% of interviewed patients and relatives were unaware it existed, with some fearing that calling would count as complaining.

The transferable part: a named, advertised, 24/7 number that reaches a differently-accountable team, plus a mandated daily question whose answer is recorded. It converts family concern from a social act, negotiable by the person being challenged, into a logged escalation with its own responder.

Learning from Excellence — report what went right

Started by Dr Adrian Plunkett at Birmingham Children's Hospital in 2014, on the observation that "excellence in healthcare is highly prevalent, but there is no formal system to capture it". It uses the same reporting architecture as incident reporting, pointed at good practice, with feedback delivered to named individuals.

A thematic analysis of 400 reports from two English trusts found staff most value kindness, compassion and time-giving — qualities "underrepresented in existing non-technical skills frameworks".

The evidence is honest about limits: an independent Norwegian study found it "mostly contributed to a positive working climate" but flagged real harms — distress among staff who never received a report, and vagueness that produced no actual learning. There is no published evidence it improves patient safety outcomes.

The transferable part: excellence becomes a data stream that gets studied. With two cautions — appreciation must be widely distributed or it corrodes morale, and without a deliberate learning step it is decoration.

Schwartz Rounds — protected time to be human

Kenneth Schwartz, a Boston lawyer dying of lung cancer, wrote in 1995 that "for as skilled and knowledgeable as my caregivers are, what matters most is that they have empathized with me in a way that gives me hope and makes me feel like a human being, not just an illness", describing acts of kindness that made "the unbearable bearable".

Rounds came to the UK and Ireland in 2009 via The King's Fund's Point of Care programme, later the Point of Care Foundation, and now run in over 200 settings. Note for currency: the Point of Care Foundation closed at the end of 2025 and the programme transferred to the Schwartz Center in Boston.

The major NIHR-funded evaluation found psychological distress fell significantly more among regular attenders — caseness dropping from 25% to 12%, against 37% to 34% in non-attenders. But be straight about this: the study's primary hypothesis, on work engagement, was not supported; the wellbeing result was a secondary outcome in a self-selected observational cohort at p<0.05, with baseline differences between groups.

It is an encouraging association, not a demonstrated causal effect, and there is no randomised trial.

The transferable part: scheduled, confidential, multidisciplinary time that is explicitly flattened to equals and explicitly not for problem-solving. Remove either the flattening or the prohibition on fixing and it becomes a case conference.

ABLE / EPIC — the closest precedent for what healthcare has not yet done

Not a healthcare example, which is the point. Ervin Staub developed active bystandership training for policing after the Rodney King beating in 1991; it became EPIC (Ethical Policing Is Courageous) in New Orleans and then the ABLE Project at Georgetown Law. It is described as "a training that is also a vehicle for cultural change", is mandatory from leadership to rank-and-file, and is now used by more than 430 US law-enforcement agencies.

It is included here because it is an existence proof. Research on how ordinary people come to participate in atrocity has been translated into a scalable behavioural intervention for a hierarchical, high-stress, high-consequence profession with a defensive culture and a history of closing ranks — and adopted at national scale.

Healthcare shares every one of those characteristics. Active Bystander training exists in pockets of the NHS, including parts of maternity, but nothing with ABLE's scale, mandate or institutional backing.

The transferable part: the training is compulsory for everyone including the leadership, and it is framed as a cultural mechanism rather than a course. Both features are what NHS speaking-up initiatives have lacked — Freedom to Speak Up Guardians are a channel staff may use, not an obligation the organisation places on everyone, and least of all on its board.

Two more worth knowing

The Scottish Patient Safety Programme paired a small number of publicly stated numerical goals with a shared measurement infrastructure across every board from 2008. The Scottish Government reported a 10.6% reduction in hospital mortality over 2014 to 2017 — roughly 7,800 fewer deaths than expected — meeting its target 15 months early.

Attribution is genuinely unresolved: hospital mortality fell across comparable systems over the same period and there was no concurrent control, and the best-known paper was written by the programme's own clinical lead.

The transferable part is nonetheless real — improvement as a governed national obligation with a scoreboard rather than a local enthusiasm.

UCLH has been the top recommended place to work among English acute trusts for four consecutive staff surveys, 2022 to 2025, with 87% saying patient care is the organisation's top priority.

Northumbria is the trust usually cited for this, and it genuinely led the field around 2015 and 2020–21 with an 80% survey response rate and the highest staff morale in the NHS; its recent claims have shifted to patient-experience measures, and a top staff-survey position for 2022 onwards could not be verified.

Northumbria's own account of its method is the useful part: patient and staff experience data returned to the specific ward team within 48 hours, and cross-boundary "Big Room" sessions that always open with a patient's story.

Sir James Mackey: "how happy and satisfied staff are in their work directly links to the quality of patient care."

Closer to home: English councils that turned round

The most useful bright spots for England may not be in Minnesota or Alaska but in Leeds and Manchester — and they were built in social care, under the same funding pressure as everyone else. The care emergency annex sets out the full list of these lighthouses.

What belongs here is how the two biggest turnarounds were done, because they follow exactly the sequence this volume describes: an identity first, then practice built to hold it. The annex of examples reads these cases as designs — what was built in, and which part survives being copied elsewhere. Annex 15 · Humanity in healthcare: examples Annex 5 · The care emergency: social and children's care

Leeds: "with, not to". Ofsted judged the city's safeguarding services inadequate in January 2010 and the government issued an improvement notice. Under Nigel Richardson, Director of Children's Services for about six years from 2010, the city made two decisions.

The first was about identity: in 2012 it launched Child Friendly Leeds, an ambition to be "the best city for all our children and young people to grow up in", with cross-party support and more than 500 organisations signed up.

The second was about practice: it adopted restorative practice across the whole service, on the principle that outcomes are better when those in authority do things with children and families rather than to them or for them, and it ran family group conferences at scale, so that families were asked to find their own solutions before the state stepped in.

Ofsted rated Leeds good in 2015, with leadership outstanding, saying the investment in restorative practice was "having a transformational impact on culture and professional practice". It was rated outstanding in 2018, again in 2022 and a third time in 2025 — the first core city to do it three times in a row.

In 2018 inspectors wrote that "relationship-based practice is a clear feature of the work in Leeds" and that children stay with the same social worker wherever possible.

Child protection plans fell by nearly half between 2011 and 2017, and in 2018–19 the child protection rate was 31.8 per 10,000 children against 55.8 across England. The government's evaluation of the family-conference programme found significant falls in children in care and on protection plans, while warning that it was too early to be sure outcomes for families would improve consistently.

Manchester: "our child". In 2014 Ofsted judged Manchester's children's services inadequate on every leadership measure: 486 cases were waiting for assessment, some social workers held 40 to 50 cases each, and 1,406 children were in care.

Paul Marshall arrived as director in March 2016 with a vision called "Our Manchester, our children". The council recruited 40% more frontline social workers, cleared the unallocated cases and adopted a single practice model.

By 2017 leadership was rated good; by 2022 the whole service was good, with Ofsted praising an "'our child' approach" and continuing "financial commitment to the recruitment and retention of social workers"; and in November 2025 it was rated outstanding.

Inspectors credited leaders' "tenacious and determined efforts over many years", found that "relationship-based ways of working have been systematically embedded", and wrote a sentence that would not have appeared in 2014: "Social workers talk about children with warmth and kindness." Vacancy rates were below the national average and agency staff were converting to permanent posts.

What the two have in common is what Mayo and Buurtzorg have in common. A settled answer to what the service is for — a child-friendly city, our child — stated by leaders who stayed long enough to mean it.

Practice that keeps the same people around the same child. Money spent on keeping staff rather than on agency cover. And relationships put back at the centre, in the inspectors' own words. Both still have work to do — Ofsted asked each of them to improve its response to neglect — which is what honest bright spots look like.

And the models travel, when they are taken seriously. In 2019 the Department for Education committed £84m ($113m) over five years to spread three English council models to other councils: Hertfordshire's Family Safeguarding (multidisciplinary teams around the family), North Yorkshire's No Wrong Door (one team that stays with the young person) and Leeds's family conferences.

Nottinghamshire, where Anthony May led children's services from inadequate in 2010 to good in 2015 before taking on Nottingham University Hospitals, is the reminder that leaders who have turned one service round can be asked to turn round another.

The care emergency annex gives the figures and the caveats, including the IICSA findings that sit alongside Nottinghamshire's recovery. Annex 5 · The care emergency: social and children's care

Bright spots at the bedside

Some of the most robust bright spots are not institutions at all but ways of caring, and they are cheap. The annex on why being human is good for health gives the evidence. Annex 7 · Why being human is good for health

The Hospital Elder Life Program, which pairs staff and volunteers with frail older patients to keep them oriented, moving and in company, roughly halves the odds of delirium. Person-centred care training in English care homes reduced agitation and improved quality of life at lower cost than usual care.

Immediate skin-to-skin contact between mother and baby cut deaths among the smallest newborns by a quarter in a WHO trial. Bringing parents into the daily care of premature babies helped them grow.

None of these needs a new organisation. Each needs an organisation that decides this is how it treats people.

Values as infrastructure: the Aga Khan network and Elizabeth Fry

Two cases show the same principle from outside the Anglo-American mainstream and from a century and a half earlier.

Aga Khan University Hospital, Nairobi was the first hospital in East Africa to gain Joint Commission International accreditation, in 2013, and secured its fifth successive accreditation in December 2025 with a score of 98.4% across 268 standards.

The revealing part is chief executive Rashid Khalani's explanation: "although regulators do not require them, we continue to pursue these standards because it is the right thing to do for our patients", and "this accreditation is not simply a certificate on the wall.

It defines our way of life." That is the exact inverse of the English failure mode, in which compliance substituted for assurance and a good rating reduced scrutiny. Here the standard is sought where no regulator compels it.

A Patient Welfare Programme subsidises those who could not otherwise afford care, on the stated principle that high-quality healthcare "should be available to all, regardless of gender, faith or ethnic origin". The parent Aga Khan Development Network works "irrespective of their origin, faith, or gender", drawing on an explicitly Ismaili Muslim framework — "in Islam, service to God is not only worship, but also service to humanity" — while describing itself as not a religious organisation, its work underpinned by "the ethical principles of Islam — particularly consultation, solidarity with those less fortunate, self-reliance and human dignity".

Elizabeth Fry made the case for values as infrastructure before nursing was a profession. A Norwich Quaker, she began visiting Newgate Prison in 1813 and concluded that "much may be done amongst these unhappy outcasts merely by kindness accompanied by instruction and employment".

In 1840 she founded the Protestant Sisters of Charity, renamed the Institution of Nursing Sisters, whose probationers trained at Guy's Hospital from 1841 — a generation before the Nightingale school. In 1897 the British Medical Journal, reviewing the evolution of nursing, called her "the founder of nursing".

Her governing rule is the sentence that matters: the Institution's spirit was "that the poor should be attended in their houses with the same kindness and care bestowed upon the more wealthy part of the community".

She did not leave it as a sentiment — she pursued it through a standing institution with a training relationship, and through statute, giving evidence to a Commons select committee that led to the Gaols Act 1823, mandating sex-segregated prisons with female warders, chaplain visits, salaried gaolers and the prohibition of irons.

The American Jewish hospitals: a natural experiment in whether identity lasts

About 113 hospitals were founded on this pattern in the United States. Roughly 22 are left. Edward Halperin's study in Academic Medicine (May 2012) counts "approximately 113 Jewish hospitals" founded across American history, of which "there are now about 22 left, some of which are Jewish in name only".

They were founded, he writes, "to serve indigent Jews, to respond to anti-Semitism by creating opportunities for graduate medical education (GME) and medical practice, to provide culturally sensitive care to observant Jews, and to fulfill a religious commitment to healing" — which is to say, they were built by people the existing institutions would not admit, as patients or as doctors.

The same mechanism as Fry and the Quakers, and as the Aga Khan hospitals: the excluded build the thing, and then it serves everybody.

The institutions are the ones you would expect, and their own histories are specific about the purpose. Mount Sinai in New York was founded in 1852 as the Jews' Hospital in New York and opened in 1855, for "free medical care for indigent Jews in New York City"; sectarian by charter, it accepted emergency patients of any religious affiliation, and it abandoned the sectarian charter in 1866, taking its present name.

Montefiore in the Bronx was founded in 1884 by Jewish philanthropists as the Montefiore Home for Chronic Invalids and, in its own words, "offered care to people denied treatment by other medical institutions" — chronic patients nobody else would admit.

Beth Israel in Boston was established in 1916 by the city's Jewish community for immigrants in Roxbury, with Yiddish-speaking staff and kosher food, and was "accessible to everyone".

Michael Reese in Chicago opened in 1881 from a bequest whose heirs required that the hospital be "open to all people, regardless of creed, nationality, or race"; at its height it had 2,400 beds and was the largest hospital in the city.

Cedars-Sinai in Los Angeles traces to the Kaspare Cohn Hospital of 1902 and was constituted by merger in 1961, and it is the one of these that still says so aloud: its mission is "founded in the Judaic tradition, which inspires our devotion to the art and science of healing".

And here is the part that matters for this document, which is not the inspiring part. Halperin's explanation for the disappearance of the other ninety-odd is that the exclusion they existed to answer ended: "the economic pressures facing all community hospitals, a decline in anti-Semitism, open access to GME positions and hospital privileges, the general acceptance of Jews in the American mainstream, and a loss of Jewish community philanthropy".

That is this volume's pattern running backwards. These were identities expressed in structure, and they worked — they trained the physicians the establishment would not train and treated the patients it would not treat.

What they lacked was any mechanism for renewing the identity once the original grievance was settled: no Credo Challenge, nothing that made each generation argue the purpose out again.

Michael Reese closed on 31 August 2009 after years of operating losses. Mount Sinai Beth Israel in Manhattan — itself a Jewish-founded hospital, latterly part of the Mount Sinai system — began reducing services in 2023, was approved for closure in July 2024 and closed after an appeal court cleared the way on 8 April 2025, on grounds of financial losses and falling inpatient volumes, over the objections of elected representatives and of the low-income community of roughly 400,000 people it served.

An institution founded to treat the people others would not, closed because the numbers did not work.

Three things follow. First, a founding ethic is a real asset and it does real work; nothing here is sentimental about that.

Second, it is not self-sustaining — left unmaintained, an identity outlives its purpose and then dissolves into the ordinary economics of the sector, which is Volume 4's argument about the Credo seen from the other end.

Third, and most usefully for England: the NHS's position is the mirror image of this. Its founding purpose has not become redundant, because the exclusion it answered — that care depended on means — has not ended. What has never been built is the maintenance.

The faith-rooted pattern, tested — because the strong version of it is not true

It is tempting to conclude from Mayo, the Aga Khan network and Fry that faith-rooted systems simply run better. The evidence does not support that claim, and the version that survives testing is more useful than the version that does not.

What does not hold. The widely repeated figure that faith-based providers deliver 30–70% of health care in sub-Saharan Africa was examined in the Lancet's 2015 series on faith-based health care and found wanting: the basis for it "is vague and maybe overstated", and household surveys put the share below 10% against the 30–40% implied by facility counts.

An independent meta-analysis (PLOS ONE, 2012) puts it at 4.9% of outpatient and 6.1% of inpatient care across developing countries, and traces the much-quoted "60% of Nigeria" to an unverified personal communication.

There is no WHO or Joint Learning Initiative evidence that values-based systems outperform on safety, resilience or staff retention — that is an evidence gap, not a finding.

The only recent controlled comparison available, of 291 Californian hospitals between 2016 and 2020 (INQUIRY, 2025), found religiously affiliated hospitals performing significantly worse than secular ones on several outcome groups at higher revenue levels.

And on cost, the Lancet authors found faith providers dependent on user fees becoming more expensive for households as public coverage expands.

And the counter-case has to be stated, because it is the strongest test of this document's own argument. Ireland's Commission of Investigation reported on 12 January 2021 that around 9,000 children died in mother and baby homes between 1922 and the late 1990s, with vaccine trials conducted on children without parental consent; the Tuam home was run by the Bon Secours Sisters, whose successor body is today one of the largest Catholic health systems in the United States.

The Scottish Child Abuse Inquiry's first case study, on Smyllum Park, found places with "no love, no compassion, no dignity and no comfort". In 1994 the editor of the Lancet, Dr Robin Fox, visited Kalighat in Calcutta and found care "haphazard", no strong analgesics in the formulary and investigations "seldom permissible" — moral seriousness is not clinical competence.

Nor is any of this historical: a 2022 investigation reported Ascension executives boasting "about how they had slashed $500 million from the chain's labor costs" while clinicians filed complaints about possible preventable deaths; Methodist Le Bonheur filed more than 8,300 medical-debt lawsuits between 2014 and 2018 and garnished the wages of over seventy of its own employees; and roughly one in six US hospital beds now sits in a facility whose doctrinal directives restrict care.

So a covenant is not a control. Where a values statement substitutes for measurement, external scrutiny and the freedom to challenge, faith-rooted organisations fail as badly as anyone else — and sometimes worse, because the moral claim itself deters challenge.

That is the Chester mechanism wearing different clothes, and it is the reason this volume insists on identity carried into structure rather than identity alone.

What does hold is a mechanism, not a theology. One favourable finding is consistent across the literature: patients report higher satisfaction and more compassionate treatment at faith-based providers, though the Lancet authors note the attribution to values "required further substantiation".

And a small number of values-rooted institutions have changed practice in ways nothing else did — St Christopher's Hospice, founded by Dame Cicely Saunders in Sydenham in 1967 out of explicit Christian conviction, invented the modern hospice and palliative care movement and exported it worldwide; Aga Khan University Hospital sought international JCI accreditation in 2006, sixteen years before Pakistan had a healthcare commission requiring comparable scrutiny; the Sri Sathya Sai neurosurgery unit provides tertiary care at no cost to any patient; Intermountain absorbed a $9m annual revenue loss to improve obstetric care.

What those four share is not a creed. It is a constitutional feature: an owner whose stated purpose is something other than margin, which makes it institutionally possible to lose money doing the right thing and to treat data as inquiry rather than threat.

Which is why the case in this document does not rest on faith at all. Buurtzorg, Nuka and Virginia Mason arrive at the same arrangements from no religious tradition whatever. The faith examples matter because they are the oldest, the most documented and the most severely tested — including where they failed.

Britain's own instances run both ways: the hospice movement on one side, and on the other the institutions whose inquiries are still reporting.

Médecins Sans Frontières — the honest case

MSF belongs here, but not as a flattering example. Founded in December 1971, it is the clearest instance of values written as binding constraint: its charter commits it to "neutrality and impartiality in the name of universal medical ethics" and requires "complete independence from all political, economic or religious powers".

That independence is financially real — about 98% of income from private sources, some 7.1 million individual supporters giving roughly €2.3bn of a €2.36bn total — which is what lets it refuse money that would compromise it. Around 91% of its field workforce is locally hired.

And it failed anyway. After the killing of George Floyd in 2020, more than 1,000 current and former staff signed an open letter alleging MSF was built on "white supremacy" perpetuating racism "in our policies, in our hiring practices, in our workplace culture".

Leadership accepted the charge publicly, saying MSF had "failed people of colour, both staff and patients" and "failed to tackle institutional racism", and set seven reform priorities.

Its own published assessment since is unflattering: progress "has been uneven", and "the progress we have made does not meet the expectations of our colleagues who continue to face racism".

The transferable part is the lesson, not the model. Strong values do not immunise an organisation against the failures in this document. What MSF did differently was publish its own failure, in its own words, rather than wait for an inquiry to do it — which is precisely what the Countess of Chester, Nottingham and Durham did not do.

The common thread

Mayo, the Aga Khan network and Fry's Institution draw on entirely different confessional sources — Franciscan Catholic sisters with a frontier medical family, Ismaili Muslim, Quaker — and arrive at strikingly similar arrangements.

Each states its purpose in terms of the person served rather than the institution. Each extends that obligation explicitly to the poor while declining to restrict service by creed.

And, decisively, each embeds the claim in structures that constrain self-interest: Mayo in a deed of gift, a salary policy and bylaws reserving leadership to clinicians; AKDN in an endowed network with published welfare subsidy and voluntarily sought external audit; Fry in an institution and an Act of Parliament.

All three were founded by outsiders — religious minorities and women excluded from the professions they went on to reshape.

The secular cases point the same way. Nuka works because ownership is legally transferred, not because staff are told to be person-centred. Buurtzorg works because the coach ratio makes supervision impossible.

Virginia Mason works because reporting is an obligation with a deadline attached. Esther works because one question is short enough for anyone to ask.

The English councils point the same way again. Leeds works because restorative practice and family conferences are how the whole service operates, not a pilot. Manchester works because it paid to keep its social workers and kept the same people around the same children. Both began with a sentence about who they were for.

But it would be a serious mistake to read this as a lesson about mechanisms. The structures are not clever design devices that happen to produce good behaviour. They are what these people and institutions built because of who they had decided to be.

The Mayo brothers did not adopt a salary policy as a governance technique — they gave away a profitable practice and most of their savings because they understood themselves as custodians of "the sick man's dollar".

Fry did not reach for an Act of Parliament as a lever; she was a Quaker who held that every person carried equal worth, and legislation was what that conviction looked like when applied to prisons. The Aga Khan network does not pursue unrequired accreditation to signal quality — it does so because, in Khalani's words, "it is the right thing to do for our patients", which is a statement about identity before it is one about standards.

Identity first, then values, then structure. That is the actual sequence, and the order matters:

• Identity without structure depends on the founder and dies with them. Fry knew this, which is why she institutionalised.
• Structure without identity becomes compliance — and this is precisely England's failure mode. The NHS has a statutory duty of candour, Freedom to Speak Up Guardians, fit and proper person tests, a Leadership and Management Framework. It has mechanisms in abundance. What the inquiries find missing is any answer to who are we that staff actually inhabit, so the mechanisms get worked around, ticked off, or turned against the people they were built to protect.
• Values without either are the wall poster.

So the inference for English healthcare is not that it needs a religious foundation, nor that any of these models transplants wholesale. It is that an institution behaves humanely when it has decided that this is what it is — and then builds the arrangements that hold it to that when it is under pressure.

England has values statements in abundance, a great many mechanisms, and very little settled identity underneath either. That is diagnosed, not treated arrived at from the opposite direction.

What happens next

The 2026 reforms attach real consequences to leadership failure for the first time, and that matters. But they concentrate on detecting and punishing failure after it happens, not on the conditions that produce it — and Nottingham, Northern Ireland and the care system show how thin the margin now is.

The one thing to take away: accountability is necessary but not enough, so judge the reforms by whether behaviour changes on the ward, not by whether frameworks are published.

Research synthesis

What. For the first time, the reforms carry consequence rather than exhortation. An HCPC barring scheme for senior NHS managers, the NHS College of Leadership and Management (launched 9 July 2026), and Thirlwall's 17 dated recommendations — barring by September 2027, statutory regulation by September 2032, a deliberate-harm protocol by March 2027, and an NAO audit of inquiry-recommendation implementation from September 2027. Criminal investigations are live at both Chester and Durham.

What none of them does is protect the person who holds the line. Every case in this record was stopped by someone who accepted a personal cost for saying what they saw — and the reforms now being legislated act on the people who did the harm, after the event, not on the conditions faced by the person who reported it.

The National Guardian's Office closed on 30 June 2026; the speaking-up score in the 2025 staff survey is at a five-year low. Barring removes a person. It does not put anyone alongside the next consultant who has to say something unwelcome for eleven months. Annex 17 · Holding the line

And the profession being regulated says the diagnosis is wrong, which should be recorded rather than brushed aside. The editor of the Health Service Journal, Alastair McLellan, wrote that the Thirlwall inquiry "may be of some use to the families involved, but its summary findings and recommendations are patronising, confusing and in some cases plain ignorant".

The managers' union, Managers in Partnership, had already said of the government's scheme in July 2025 that statutory regulation is "a small piece of the jigsaw", and that the larger threat to good management is the scale of the job cuts now being made — management costs are being halved across integrated care boards and trusts.

The opposite case is made from inside the same profession. The BMA welcomed statutory regulation and pressed for it to be implemented sooner. Judith Smith, professor of health policy and management at the University of Birmingham, argued that what is needed is a code of conduct written into every manager's employment contract with patient safety at its heart, alongside proper training and development, because the previous code was voluntary.

This document's position is not between the two so much as underneath them. A barring scheme acts after the harm, on the individual. A code of conduct in a contract is worth having and is still a rule about conduct rather than a capability.

Neither answers the question of what a manager is actually rewarded for on an ordinary Tuesday, which is where the failures in this record were made.

It is also possible to say how isolated the trade press's position is, because it was measured. When the government consulted on regulating NHS managers it received 4,907 responses, and its published outcome records that "the vast majority of respondents (92%) agreed that NHS managers should be regulated".

Among the non-managerial workforce 98% agreed. Among members of the public 95% agreed. Among NHS managers and leaders themselves, 79% agreed.

A position held by roughly one manager in five is a real position and deserves argument rather than dismissal. It is not the profession's settled view, and it should not be reported as one.

The sequence of response is as familiar as the sequence of failure, and it is worth naming. Denial first. Then minimising — the incident was smaller, or different, or already in hand. Then reframing a specific failure as a systemic one, which has the advantage of being partly true.

Then an objection to the tone of the people who found it. And, eventually, acceptance. Chester's leadership spent years in the first four stages. Durham's board accepted reassurance rather than seeking assurance for thirteen.

One organisation in this record went straight to the last stage, and the Inquiry noticed: "Unlike other organisations, the starting point for the RCPCH was an uncomplicated recognition that their review team had made serious mistakes, for which the RCPCH acknowledged accountability and responsibility."

The apologies were "genuine and reflective" and the changes since "far reaching". Acceptance is not weakness. It is the precondition for changing anything.

And the public's test is simpler than the profession's. An institution that exists to serve the public, and is paid for by it, cannot answer a finding of neglect by saying that the job is hard. The job is hard.

The people who take the authority and the salary take the duty that comes with them, and the conditions cited as the excuse are set, in large part, by the same people citing them.

That is not an argument against the structural case, which this document accepts and makes at length. It is an argument about the order of things: acceptance first, then the argument about remedies. The strongest version of the managers' case is not self-protective at all, and this document has already conceded it — a regime that is only punitive drives concealment.

Thirlwall says the same thing from the other side: "whilst blame is always negative, responsibility is not".

The HSJ leader is behind a paywall; the words quoted are from the published standfirst. The Managers in Partnership statement is dated 21 July 2025 and responds to the government's announcement, not to the Thirlwall report.

So what. It is all detection and punishment, and none of it changes what a board is rewarded for on an ordinary Tuesday — which is finance, throughput and reputation. Nottingham University Hospitals shows the bind exactly: £106m ($143m) of savings against a £1.8bn ($2.4bn) budget, ~1,279 posts going, a £34.6m ($46.7m) deficit by July — while implementing Ockenden, rebuilding maternity safety and asking staff to speak up more freely.

Northern Ireland is the acute risk: 504,537 waiting for a first outpatient appointment, 53.7% over a year, ED performance around 31%, on a budget baseline falling from £8,898m ($12.0bn) to £8,481m ($11.4bn).

(Sterling figures carry a US dollar equivalent at $1.35 = £1.) Annex 9 · Joining up health and social care: what integration can and cannot do

Now what. Accountability is necessary, and it will not heal the system. Root and branch would mean four changes no current proposal touches: what leaders are measured and paid on; who leads and how close they are to care; where candour sits; and whether capacity is honest.

But even those are changes the system would have to make to itself, and the record of twenty-five years is that it does not — the wounded rarely recover unaided, and a system whose core mechanisms produced the harm is the least likely source of its own repair.

Two futures are equally live: 2026 as the year accountability acquired teeth, or a nearly empty barring register, a deferred NAO audit, a Northern Ireland catastrophe no staff survey was in place to anticipate, and another inquiry in the mid-2030s finding the same thing.

Which one arrives depends less on the reforms than on whether the service imports something it does not currently have — which is Volume 10.

What is changing now

The response to this cycle of inquiries is materially different from the response to Francis, because for the first time it carries statutory enforcement rather than exhortation.

Statutory barring of senior managers. Following its consultation response, the government is legislating for a statutory barring system operated by the Health and Care Professions Council.

Managers can be disbarred for serious misconduct — criminal convictions, deliberate bullying or harassment, dishonesty — and a barred individual cannot hold a senior NHS management post. This is the mechanism that closes the revolving door.

Three limits are worth noting. It is a barring system, not full registration, so there is no positive licence to practise. Its initial scope covers board-level directors, their direct reports, integrated care board members and DHSC arm's-length body directors — not primary care, social care, independent providers, or the mid-level operational managers who in practice handle ward-level grievances and whistleblower reports.

And the timetable is slow: parliamentary approval anticipated in the second half of 2026, with regulation commencing within twelve months of the legislation taking effect.

Thirlwall also noted the structural hole in it: if a hospital does not invoke disciplinary process and dismiss for serious misconduct, "there would be no basis for barring the manager".

The NHS College of Leadership and Management, launched 9 July 2026 under interim chief executive Steve McManus, addresses the "accidental manager" problem — the absence of any professional standard for healthcare administrators.

Co-designed with the Chartered Management Institute, the Faculty of Medical Leadership and Management and the Florence Nightingale Foundation, it brings a Leadership and Management Framework, a national code of practice, and duty of candour as an explicit leadership standard.

The rollout runs across 2026/27: organisations are asked to embed the framework in development, appraisal and recruitment during the year; every manager and leader is to self-assess against the framework stage relevant to their level and agree a development plan by the end of Q4 2026/27; board-level competency frameworks are aligned from 2027/28.

So 2026/27 is the year it should start showing up in behaviour, and 2027/28 is the first year in which anyone can test whether it did.

And the government changed while the reform programme was in flight, which matters more than it might sound. As at the Department of Health's release of 5 August 2026, Andy Burnham is Prime Minister and Yvette Cooper Secretary of State for Health and Social Care — the handover falls between late June and late July 2026, and the precise date is not established here, so this document does not state one.

The significance is not the personnel. It is that the Prime Minister is the author of Building the National Care Service, published when he was Health Secretary on 30 March 2010 and destroyed in the "death tax" row before that year's election.

On 29 July 2026 he committed to a National Care Service on four principles — seamless movement between hospital, home and community care; full and sustainable funding; prevention and independence; citizens' needs at the centre — brought Baroness Casey's commission on adult social care forward from 2028 to summer 2027, launched a public consultation and asked for cross-party talks, noting twenty-two failed attempts at reform in three decades.

No money was attached. In August the government said it would explore broadening the Fair Pay Agreement so that care pay and progression align with NHS standards. Whether a National Care Service helps will depend on what it does, not how it is organised. Annex 9 · Joining up health and social care · Annex 5 · The care emergency: social and children's care

Treat all of it as intent with a date on it, not delivery — but note that the accountability package described above was designed under one administration and now has to be carried by another.

Thirlwall's own 17 recommendations (Chapter 45) are notable for carrying dates rather than aspirations. The most consequential, grouped:

Also recommended: bereavement care pathway implementation (Aug 2027), interoperable NHS computer systems (Dec 2028), SUDIC procedures extended to sudden deaths of hospitalised babies, medical examiner safeguarding questions, and 37 paediatric and perinatal pathology training posts by 2033.

Chapter 45 opens on money, not process: "The long-standing underfunding of hospital services for babies and children is well recognised and requires action."

Two features of this list distinguish it from Francis. Recommendation 17 makes implementation itself auditable by the National Audit Office — the direct answer to the observation that Francis's recommendations were adopted in vocabulary and abandoned in substance.

And recommendation 14 goes at the root of regulatory blindness: inspectors are told to investigate whether care is actually safe, not to confirm that a process exists.

An honest assessment of the odds. Three reasons for cautious optimism: statutory consequence now exists where it did not; criminal investigation of former executives at both Chester and Durham has altered boardroom calculation in a way no framework can; and implementation is being tracked centrally.

Four reasons for scepticism, each of which has defeated reform before. There is already a gap between what Thirlwall asked for and what government is delivering: Thirlwall recommends barring for all managers by September 2027 with statutory regulation by 2032, while the HCPC scheme covers board-level directors, their direct reports and ICB members — not the operational layer where whistleblowers are actually silenced.

A college can educate and set standards but cannot enforce them. The financial and operational pressure that makes boards defensive has not eased, and a board under that pressure still has every incentive to manage the narrative.

And the sheer volume of overlapping recommendations across Thirlwall, Ockenden, Amos (the national maternity and neonatal investigation of 30 June 2026), Mann and the Covid inquiry creates a thicket in which prioritisation becomes genuinely difficult; Sir Robert Francis gave evidence to Thirlwall that the volume of structural change in the NHS makes following through on recommendations extremely hard.

The honest position: the architecture of accountability is being built correctly for the first time. Whether it changes behaviour depends on something the architecture cannot supply, which is whether consequence arrives quickly and consistently enough to be believed.

The implementation test, in one trust

Nottingham University Hospitals is the clearest live demonstration of why recommendations do not convert into changed behaviour. It is the trust the Ockenden review reported on in June 2026, and therefore the trust with the strongest possible mandate to change. As at September 2026 it is being asked to do the following simultaneously:

The operational detail matters because it is where culture actually lives. Corridor care is recorded as a patient-safety issue, not merely a flow problem — a category 4 pressure ulcer was linked to a frail patient's prolonged wait in an emergency department corridor.

Complaints performance stands at 12.6% against an 85% target, with investigations averaging 55 days. Maternity reports sustained consultant and resident-doctor gaps, with consultants stepping down to fill rota gaps.

The board has itself named duty of candour as something that has to become a culture rather than a process, and has warned explicitly about priority overload.

This is the mechanism of failure described in Volume 2 and worked through in the annex on organisational failure, observed in real time and in a trust that is trying. Annex 1 · Organisational failure: the worked cases

A second case is running now in Greater Manchester, at an earlier stage. The Northern Care Alliance, which runs a hospital twice rated Outstanding, was served a trust-wide warning notice by the CQC on 13 August 2026, two months after NHS England acted on a "fundamental failure in quality governance".

Consultants had warned a year earlier; staff allege that incidents were downgraded; patients were harmed by delays in a gynaecology service; and the trust was formed by absorbing a failing neighbour.

It is the sequence in this document, observed before any inquiry exists. Annex 12 · Humanity-First Healthcare: applying lessons from the past sets out what happened and what would have to be built for it not to happen again, and Annex 13 · Radical leadership sets out the mechanisms that do the building.

Psychological safety is easiest to build in calm conditions and is needed most in stressed ones. An organisation removing around 1,279 posts while asking staff to speak up more freely is attempting two things that work against each other, and no framework resolves that tension — it is resolved, if at all, in daily behaviour under pressure.

The generalisable point: every inquiry recommendation lands on an organisation that is already full. That is why diagnosed, not treated persists. Capacity to implement, not clarity about what to implement, is the binding constraint.

Threats and opportunities

Which way this goes is genuinely open, and the range of outcomes is wider than the current debate allows. The optimistic case is real: statutory consequence exists for the first time, and the architecture being built is broadly the right architecture.

The pessimistic case is that a system already failing patients is about to be asked to absorb financial retrenchment, and that the reform programme is a reorganisation of the visible parts of a structure whose foundations are the problem.

Northern Ireland: the nearest thing to a catastrophe in waiting

England gets the inquiries; Northern Ireland has the worse system. At 30 June 2026 the figures are not comparable to anywhere else in the UK:

Source: Department of Health NI quarterly waiting times; ED performance and excess-death estimate from Pivotal's 2026 tracker. No HSC trust met any of these targets. Emergency department performance has fallen from nearly 69% a decade ago to around 31%, and overcrowding has been linked to an estimated 1,032 excess deaths.

A waiting list of 504,537 in a population of roughly 1.9 million is on the order of a quarter of the population waiting for a first appointment, on one list alone. And this is the improving picture: outpatient and inpatient lists fell slightly in the quarter, while diagnostics rose.

Now add the money. The Department of Health's resource budget for 2026–27 is £8,481m ($11.4bn), down from £8,898m ($12.0bn) in the 2025–26 final plan — so it enters the period with a lower cash baseline.

It is simultaneously carrying forward pay awards, facing a further NHS Pay Review Body recommendation of 3.3% costing roughly £120m ($162m) in 2026–27 alone, and contributing to overspend recovery. The Fiscal Council's assessment is unusually direct: the Budget "faces a credibility challenge: it may satisfy the formal balancing requirement while still implying outcomes that departments cannot deliver without additional in-year funding", and sustainability is unlikely without "meaningful service transformation", greater local fiscal effort, or a changed funding settlement.

The structural trap is specific and worth understanding, because it is not a failure of local management. When the UK government funds an x% pay rise in England, the Barnett-derived block grant does not rise enough to fund the same percentage rise in Northern Ireland. So pay settlements set elsewhere land as unfunded pressure locally, every year, compounding.

Why this is the acute risk — and the lesson England has already taught. The single most consistent finding across the English inquiries is this: running a health service on financial and operational targets, without being rooted firmly in clinical and human priorities, has repeatedly preceded human harm.

Mid Staffordshire pursued foundation trust status and cost control while basic care collapsed. Nottingham's board knew about maternity safety for years while managing its reputation. Durham's own staff described a culture "predominantly performance-driven, with an emphasis on capacity, performance targets, and financial targets" in which "quality was deprioritised unless it directly impacted performance indicators". Three trusts, one recurring mechanism.

Northern Ireland is now being instructed to do precisely that, at national scale. Bringing an overspend of this kind into line, in a system already failing this comprehensively, means constraining capacity in a service that has none to spare — and it means boards spending the next three years judged primarily on whether they delivered the financial plan.

Every precondition the inquiries identify is therefore being assembled at once: acute pressure, no capacity to pause, boards judged on finance, and — uniquely — the thinnest culture data in the UK, with no equivalent of the NHS Staff Survey to detect the consequences while they accumulate.

On the evidence of this document, the next major inquiry is already in the pipeline. Not because anyone in Northern Ireland intends harm, but because the conditions that produced Mid Staffordshire are being recreated as the predictable consequence of the budget settlement, in the one part of these islands least equipped to notice.

One correction to a figure in circulation: the overspend is not 12–14% of the health budget. The Fiscal Council forecast a health overspend of roughly £212m ($286m) in 2025–26 against a budget near £8.9bn ($12.0bn), and the Department's own pre-mitigation gap has been reported at around £800m ($1.08bn), which is closer to 9%.

The sharper number is the falling baseline — roughly £417m ($563m) less in cash terms year on year while costs rise.

Threats, with the signal that would tell you they are materialising

Opportunities, which are equally real

Consequence now exists. For the first time a senior manager can be barred, and executives at two trusts are under criminal investigation. Nothing in the previous twenty-five years changed boardroom calculation the way that does.

Implementation is becoming auditable. Thirlwall's recommendation 17 puts the National Audit Office onto whether inquiry recommendations are actually implemented. If that lands, it breaks the specific failure mode that neutralised Francis.

The evidence base is strong enough to act on. Civility and psychological safety are no longer soft propositions: the effect of incivility on clinical performance is experimentally demonstrated and the intervention literature is clear that multi-component, leadership-led, sustained programmes work and one-off training does not. That is actionable knowledge, not an aspiration.

There is an exemplar to copy, and it is structural. Mayo, the Aga Khan network and Fry's Institution show that values bind when built into incentives, appointments and constitutions. England already knows how to legislate; the new barring scheme proves it.

The same logic could be applied to appointments, pay and board composition. And the closest exemplar is the NHS's own founding: Bevan built the financial half of the 1948 promise into an Act of Parliament and left the humane half to chance — what was built into structure survived seventy-eight years, and what was left to exhortation is what every inquiry here describes.

What transfers from Mayo, the Aga Khan network and Fry is not a creed but an owner whose purpose is something other than margin.

What transfers from them is the rule rather than the mechanism, and the copies that failed are as instructive as the ones that held: name the function you are trying to reproduce, redesign the form for your own place with the people who will run it, and change whatever sits above the team so that the new way of working is not quietly reabsorbed.

The work of building that into one organisation has a name — enablership — and it is the subject of its own annex. Annex 15 · Humanity in healthcare: examples · Annex 14 · Healthcare: why the Art of Enablership is critical

Families are now a permanent constituency. The inquiries of 2024–26 happened because families organised and would not stop. That capability does not dissolve, and it is the most reliable safety mechanism the system currently has.

Four nations, four natural experiments. Scotland's engagement-first approach, Wales's culture-focused restructuring and Northern Ireland's constrained position provide comparisons England has barely used. The variation is an asset if anyone studies it.

Root and branch, not the deckchairs

The honest reading of the inquiry record is that this needs root-and-branch reform, and that what is on offer is not yet that. The reason is not cynicism about the reforms; it is that almost every element of the current package addresses detection and punishment rather than the conditions that produce the failure.

Barring removes bad executives after harm is proven. A recommendation hub tracks compliance after recommendations are written. The College trains and sets standards. All are worth having, and none of them changes what a board is rewarded for on an ordinary Tuesday, which is finance, throughput and reputation.

Twenty-five years of inquiries say the failure happens at precisely that point — not when someone decides to cover up, but when the thing that gets measured crowds out the thing that matters, and the person who mentions it becomes inconvenient.

Root and branch would mean changing four things that no proposal in the accountability package touches:

1. What leaders are measured and paid on. Culture, candour and safety would have to carry the same weight in oversight frameworks and executive appraisal as finance and waiting times, with the consequence attached.
2. Who leads, and how close they are to care. Mayo reserves its presidency to a practising clinician by constitution. England has a mobile generalist executive class and no equivalent requirement. That is a choice, not a necessity.
3. Where candour sits. Duty of candour currently depends on the judgement of the people with the most to lose from disclosure — which is what produced Durham's nine-month delay. Thirlwall's barring proposals and an individual duty for managers help; independent triggering would help more.
4. Whether capacity is honest. A system running above 90% bed occupancy with the fewest scanners in the G7 cannot absorb the pause, the second opinion or the investigation. Reform that assumes otherwise is asking for behaviour the system cannot afford.

The deckchairs metaphor is apt in one specific respect. The NHS has been reorganised repeatedly — the internal market, foundation trusts, the 2012 Act, the 2022 Act, integrated care boards, the abolition of NHS England — and none of those reorganisations addressed any of the four items above. Structural churn has been the substitute for structural change.

Sir Robert Francis's own evidence to Thirlwall made the related point: the volume of structural change is itself part of why recommendations do not get implemented.

So the forward view is properly two-sided. There is a credible path on which 2026 is remembered as the moment accountability acquired teeth. There is an equally credible path on which the barring register sits nearly empty, the NAO audit is quietly deferred, Northern Ireland has a catastrophe that no staff survey was in place to anticipate, and a further inquiry in the mid-2030s finds a culture in which people who raised concerns were not heard.

Both are live. The determining variable is not whether the right things have been written down — they now largely have been — but whether anything changes about what the system rewards.

Humanity-first healthcare

Humanity-first healthcare means putting humanity back into health and social care: making how people are treated as much a part of the system's design as how it is paid for. It finishes what Bevan started, rests on one goal and three pillars, and is a performance case as well as a moral one. The one thing to take away: the NHS already has the right values — the task is to build them in.

In short: putting humanity back into health and social care

Humanity-first healthcare means putting humanity back into health and social care — making how people are treated as much a part of the system's design as how it is paid for.

The story in four sentences. In 1925, just over a century ago, a young miner held his dying father, whose lungs the pits had destroyed and whose life the system treated as disposable. Twenty-three years later that son, Aneurin Bevan, built a health service on the model his own town's workers had run since 1890, so that no family would face illness alone — and made its money permanent by writing it into law, while the humanity was left to chance.

Seventy-eight years on, the money has survived and the humanity has not: the inquiry record, from Mid Staffordshire to Chester, Durham and Nottingham, describes organisations that stopped seeing patients, families and their own staff as people.

Humanity-first healthcare finishes what Bevan started, by building the humane half of the founding promise into the structure so that it holds on a bad night.

The philosophy, the goal and three pillars. The foundation is the capacity to act for the good of others — the ability, under pressure, to see the person in front of you as a human being and do what is right for them.

At its best it becomes instinctive: the last of what Champions calls its Four Stages of Humanity, "unconscious humanity", where being human under pressure is simply another day at the office. The goal, taken from the fourth principle of the NHS Constitution and widened to all care, is that the patient will be at the heart of everything. Between them stand three pillars, set out here with what each means in practice:

1. Every person counts. Care is, at root, protection: the first job of care is to stand between vulnerable people and harm — the premature baby, the frail old person with dementia, the child in care, the patient who cannot speak up. Harm happens where someone falls outside the circle of people an organisation feels obliged to protect. And being human is good medicine: warmth, connection and being known by the same people over time are part of treatment, not extras laid on top of it. In practice: every board can say who is inside its circle of care and acts when someone is not; continuity, time to talk and family involvement are designed in, not squeezed out.
2. Caring for the carers. People treat others the way they are treated themselves. Fewer than half of NHS staff feel their organisation values their work, and staff who are exhausted, isolated and unheard draw their own circle smaller — and patients feel it. Every catastrophe in this document was seen coming by people who were not heard. In practice: teams, rest, belonging and a voice for staff are treated as patient-safety measures, and speaking up is an obligation on everyone, the board included.
3. It's who we are. Humanity is an identity, not a programme. The NHS already has the right values — compassion, respect and dignity, everyone counts — but nothing in how a trust is funded, measured, appraised or led makes them bind. The institutions that sustain humane care began with a settled answer to who they were, then protected it in how leaders are chosen and judged, how services are measured and how people are trained. And it spreads through people, the way one town's medical aid society became a national service. In practice: leaders are appointed, appraised and held to account on how people are treated; a critical mass of people carries the culture.

Humanity First. Performance Follows. In plain terms: look after people first — patients, and the staff who care for them — and good performance follows. People who are treated well think more clearly, share what they know and help each other; people who are belittled do none of these as well.

So this is not a choice between being kind and being good at medicine: the first is how you get the second.

The motto does not say that performance doesn't matter, or that kindness can replace skill, money or good systems. It says those things work best when they are built on a human foundation.

That is what the Champions pilot at Union Health Service in Chicago set out to test. Its independent evaluator, Lauren Easterday of EastLight, described the early results as a view of how rebuilding human capability — emotional regulation, trust and connection — "can translate into tangible performance lift in complex, high-pressure environments", and the review was published under the title Humanity as a Performance Multiplier.

It is one outpatient organisation at 90 days, and the evaluator describes her figures as "directional", not statistical — but it points the same way as the wider evidence.

The case is moral, and it is also practical: rudeness alone costs its target about a third of their performance, and people who are treated well — patients and staff — do better. The rest of this volume sets out the evidence.

Research synthesis

What. Every failure here ran through organisations that stopped treating people as people — patients, families and their own staff. Every institution that has sustained humane care at scale built humanity into its structures rather than urging it on its staff, because of a settled identity rather than as a technique.

And the evidence now runs both ways. Inhumanity is costly: incivility costs its target about a third of their performance and bystanders about a quarter; it makes colleagues 74–84% less likely to help; 91% of clinicians passed a simulated operative crisis in normal conditions against 64% when exposed to rudeness; and 27% of US practitioners knew of a medication error in the previous year in which disrespectful behaviour played a part.

Humanity is good for health: connection, warmth and continuity of care are consistently linked to better outcomes, and the strongest trials are at the two ends of life — skin-to-skin care for the smallest babies, attentive care for frail older people.

So what. Humanity is not a supplement to performance — it is a precondition of it, and part of the treatment. Civility restores diagnostic accuracy, information-sharing and help-seeking, which in a hospital are the safety system.

Care is also protection: a malevolent or careless act needs an unguarded gap, and the gap is what an organisation produces when patients and families fall outside its circle of obligation. And an organisation teaches its people how to treat others by how it treats them — fewer than half of NHS staff now feel their organisation values their work.

The capability is trainable (Staub) and the problem is a pressure symptom rather than a character defect (Porath: half of rudeness is overload). The Champions pilot at Union Health Service in Chicago, from June 2025, showed movement within 90 days in an organisation the pack described as "one setback away from catastrophic failure: not from clinical skill, but from emotional depletion".

Now what. The system cannot heal itself, so the fix has to be brought in — and this is the cheapest clinical performance the NHS is not yet buying. Accountability is necessary and it is not a cure.

Nor is restructuring: Northern Ireland has run health and social care as one system since 1973 and has the worst waiting times in the UK. A system missing every constitutional standard it has cannot afford what it currently loses to how people treat each other, and it has had twenty-five years to generate the remedy internally and has not.

Every mechanism that works came from outside: Toyota, policing, a district nurse who rejected management, and institutions built by outsiders. That is the case for putting humanity first — a performance case as well as a moral one. Humanity First. Performance Follows.

Humanity-first healthcare

The case in three sentences. Every failure in this document ran through people who stopped being able to act humanely under pressure, and every institution that has sustained humane care at scale did so by building humanity into its structures rather than urging it on its staff.

The reforms now underway address detection and punishment; they do not touch either the conditions that deplete people or the capacity of people to withstand them. A humanity-first approach addresses the second directly and makes the first visible — and it is not a moral supplement to performance, it is a precondition of it.

Why the bright spots are bright — and it is not their mechanisms. Every one of them begins with an identity: a settled, inhabited answer to what kind of institution this is, and what kind of person works in it.

The arrangements follow from that and exist to protect it. Four recur — purpose stated in terms of the person served rather than the institution; incentives removed that make the wrong behaviour rational; the obligation to speak placed on everyone including leadership; and the human capability built deliberately rather than assumed.

Mayo expressed its identity in a deed of gift and a salary policy. Nuka expressed it by transferring ownership. Buurtzorg expressed it by making supervision arithmetically impossible. None of it was exhortation — and none of it was technique either.

Identity is what makes the structure mean something; structure is what makes the identity survive a bad night.

This is the distinction England has missed, and it explains the pattern in this document better than anything else. The NHS is not short of mechanisms: statutory duty of candour, Freedom to Speak Up Guardians, fit and proper person tests, a new Leadership and Management Framework, and now a barring register.

It is short of a settled identity that staff and boards actually inhabit — so the mechanisms are experienced as compliance, and compliance is what gets worked around when the pressure arrives. A humanity-first approach is not another mechanism to add to that list. It is work on the layer underneath, without which the list does not function.

And it is not short of stated values either — which makes the point sharper, not softer. The NHS was founded on 5 July 1948 on an explicitly moral premise: that nobody should be deterred by cost from seeking help early, and that "preventable pain is a blot on any society".

The NHS Constitution for England still sets out six core values, developed with patients, the public and staff, and two of them are compassion — "respond with humanity and kindness to each person's pain, distress, anxiety or need" — and respect and dignity. Those are precisely the two that every inquiry in this document finds absent.

So the problem is not that the NHS lacks a statement of humanity. It has one, in a constitutional document, and it is a good one. The problem is that nothing in how a trust is funded, measured, appraised or led makes that statement bind on an ordinary Tuesday.

Humanity-first healthcare does not import new values into the NHS — it makes the ones it already has structural.

Why the fix has to come from outside

This is the part of the argument that decides everything else. The two responses now on offer — hold those responsible accountable, and implement what the inquiries established — are both right and both insufficient.

Barring a chief executive removes a person, not the conditions that made his conduct rational. An audit of recommendations counts compliance; it does not create capability. Neither touches what the failures actually consisted of: people, under pressure, who stopped being able to act humanely, inside organisations that rewarded something else.

And a system is rarely the source of its own repair. Any clinician knows the general version: the wounded recover more slowly unaided, and the self-harmed almost never recover without something brought in from outside — a different pair of eyes, a relationship, a structure they did not build for themselves. Institutions are no different, and the NHS has run the experiment at length.

The internal market, foundation trusts, the 2012 Act, the 2022 Act, integrated care boards, the abolition of NHS England, statutory candour, Freedom to Speak Up, fit and proper person tests, three separate leadership frameworks — all generated from inside, across twenty-five years, and the same failure is still being described in the same words.

That is not an argument that the people inside lack will or ability. It is an observation about what a system can see of itself while standing in it.

The principle has a name, and it is a great deal older than any of this. Rabbi Lord Jonathan Sacks reached for a Talmudic maxim to state it: "A prisoner cannot release himself from prison" (Brachot 5b).

His own gloss, written about despair rather than institutions, transfers exactly: "It needs someone else to lift you from depression."

The person in the cell may know precisely where the door is and what the lock needs — that is not the same as being able to open it. An organisation whose own logic produced the harm is in the same position: it can describe its failure in great detail, as every inquiry response demonstrates, and still be unable to release itself using only the logic that turned the key.

And Sacks's larger diagnosis is the one this document arrives at from the inquiry record. In Morality: Restoring the Common Good in Divided Times (2020) he argued that "we have outsourced morality to the market and the state, but neither is capable of showing us how to live", and that a moral order rests on we rather than I: "there is no liberty without morality and no freedom without responsibility… we must all play our part in rebuilding a common moral foundation."

Read that against the NHS record and it describes what happened precisely. The service outsourced humanity to mechanism and regulation — the institutional equivalents of the market and the state — and neither can tell a ward how to behave at three in the morning.

It also explains why the remedy has to be collective rather than individual: what the inquiries found missing was not good people but shared responsibility, in organisations where it was nobody's job to speak and everybody had a reason not to.

Sacks called that the ethics of responsibility in To Heal a Fractured World; Volume 5 calls it active bystandership. They are the same claim, arrived at from opposite ends.

The evidence for importing is the bright spots themselves — and almost none of them came from healthcare. Stop-the-line reporting came from Toyota's production system, by way of a hospital that chose to copy a car factory.

Active-bystander training came from policing, built on research into how ordinary people come to take part in atrocity, and is now mandatory across more than 430 US law-enforcement agencies while healthcare has it in pockets.

The examples annex reads those same cases as design problems rather than as inspiration: what exactly was built in, what happened when other people tried to copy it, and which part actually travels. Its short answer is that the mechanism rarely transfers, the principle usually does, and the copy only works when the people who have to run it own it. Annex 15 · Humanity in healthcare: examples

Buurtzorg came from a district nurse who rejected management altogether. The Esther model came from asking one question that a nursing assistant could ask of a hospital consultant.

And the three institutions with the deepest values architecture — Mayo, the Aga Khan network, Fry's Institution of Nursing Sisters — were founded by outsiders: religious minorities and women excluded from the professions they went on to reshape. Every one of those was an import, an outsider, or both.

Which is precisely what a humanity-first approach is, and why it is not another internal programme. It brings in a capability the system has demonstrated it does not generate for itself: the practised human capacity to hold the line under pressure, in the individual and in the team, built deliberately rather than hoped for.

It comes from outside healthcare — from work on violence prevention, on bystandership, on what makes ordinary people act well when the situation is pushing the other way — and that is a strength rather than a gap in credentials, because the situations are the same and the research base is deeper.

Two honest qualifications. Imported from outside is not the same as imposed by outside: every institution in Volume 8 took the outside thing and made it part of who they were, which is the difference between Virginia Mason's alert system and a hospital that installs a reporting form.

And an import without identity behind it becomes exactly what the NHS already has too much of — another mechanism, worked around under pressure. The outside brings the capability and the mirror. The organisation still has to decide what it is.

Humanity-first is the NHS's own founding idea

This is not an import. It is a restoration. The argument so far has said the fix must be inspired from outside a system that cannot heal itself, and that remains true of the inspiration. But the idea is not foreign to the NHS. It is the idea the NHS was founded on.

What Bevan argued was not administrative and it was not economic. It was moral, and it was about how a person in distress should be treated.

Moving the Bill in 1946, he said that "a person ought not to be financially deterred from seeking medical assistance at the earliest possible stage"; and in In Place of Fear (1952) he wrote that "no society can legitimately call itself civilised if a sick person is denied medical aid because of lack of means", and, more plainly, that "preventable pain is a blot on any society".

None of those sentences mentions capacity, throughput or efficiency. All of them are statements about human dignity, by the man who built the service. That is humanity-first, in the founder's own words.

The fuller founding record, and what the NHS Constitution does and does not bind, is in the Annex on the founding vision of 1948.

And it began with a human story, not a policy. Bevan held his father as he died of miners' lung disease in 1925, uncompensated because the law did not count it as an industrial disease. He grew up in a town whose miners paid into a medical aid society so that no neighbour faced illness alone, and he scaled that small community's answer up to a nation.

The NHS was one person's choice, carried by a few, rippling out to some fifty million people — the founding annex tells the story in full. Annex 4 · The founding vision of 1948

And it is still there in writing. The NHS Constitution's six core values — set out in Volume 1 — are that founding proposition translated into modern behavioural expectations, and three of them say it almost explicitly: compassion ("respond with humanity and kindness to each person's pain, distress, anxiety or need"), respect and dignity, and everyone counts.

So a humanity-first programme is not asking the NHS to adopt someone else's values. It is asking it to deliver its own.

Then why has it not held? Because Bevan built only half of it into structure. The financial half of the 1948 promise — care free at the point of need, funded collectively from taxation — was written into an Act of Parliament and a funding mechanism.

It has survived seventy-eight years, every government of both parties and every reorganisation described in this document, because it was constitutional rather than cultural.

The humane half — that people would be treated with kindness when they were frightened — was left to the professions, to training and to chance. That is the half that appears in every inquiry report in this document.

What was built into structure survived. What was left to exhortation did not. That asymmetry is the whole case for humanity-first healthcare, and it is not a criticism of Bevan: it is the unfinished part of his design.

The tradition behind it is also, awkwardly for a country that has mislaid it, British. Elizabeth Fry set the first humanity-first standard in healthcare in Norwich in 1840, a century before the NHS (the detail is below).

Mayo — the institution most often held up worldwide as the model of patient-first medicine — was founded by a man born in Salford, just outside Manchester, on 31 May 1819, who took his scientific training and, in Mayo Clinic's own account, "the Quaker values of tolerance and working together for the common good" from the Manchester chemist John Dalton before emigrating in 1846.

Mayo Clinic publishes that history under the title "English taproot of values". And Bevan was a miner's son from Tredegar who had watched his own community pay for illness in cash. One tradition, three outputs: an institution (Fry, 1840), a clinic (Mayo, from 1819 and 1846), and a national service (Bevan, 1948) — and the two that wrote their humanity into structure are the two still routinely cited as exemplars.

Which makes the proposition a conservative one, in the literal sense of the word. Humanity-first healthcare is not a new value system for the NHS. It is the completion of the founding settlement: treating how people are treated as a matter of design — built into appraisal, appointment, measurement, training and board accountability — in exactly the way that free-at-the-point-of-need was built into the 1946 Act rather than left to the good intentions of hospital almoners.

Read that way, the ask of any board is not "adopt our programme". It is: your constitution already commits you to this; what in your organisation actually makes it happen on a bad night?

Fry is worth taking properly, because the pattern is in how she built it. Elizabeth Fry, a Norwich Quaker, founded the Institution of Nursing Sisters in 1840 on a rule that reads as the first humanity-first standard in healthcare: that "the poor should be attended in their houses with the same kindness and care bestowed upon the more wealthy part of the community".

She did not leave it as sentiment — she built an institution, secured clinical training at Guy's, and drove the Gaols Act 1823 through Parliament. The British Medical Journal called her the founder of nursing in 1897.

That Quaker pattern — equality of persons, plain dealing, institutions designed so that values survive their founders — shaped a disproportionate share of British enterprise built on trust. It reaches Mayo indirectly and specifically: William Worrall Mayo, born in Salford on 31 May 1819, became a protégé in Manchester of John Dalton, a Quaker, and Mayo Clinic's own account credits Dalton with instilling "the Quaker values of tolerance and working together for the common good" (the Mayos themselves were not Quakers — the inheritance is intellectual, not confessional).

And the Aga Khan Development Network arrives at the same structures from an entirely different source, an Ismaili Muslim ethic in which "service to God is not only worship, but also service to humanity", expressed as care "irrespective of their origin, faith, or gender" and accreditation sought where no regulator requires it.

Three faiths, three centuries, one set of arrangements. That convergence is the strongest evidence that this is a design principle rather than a devotional preference.

Volume 8 tests it properly, including against the counter-case — faith-rooted organisations have produced some of the worst documented care failures on record, and no evidence shows them outperforming secular ones.

What survives the test is narrower and sturdier: what travels is not the theology but an owner whose stated purpose is something other than margin.

The Champions pilot: Union Health Service, Chicago. A 70-year-old non-profit outpatient organisation — six sites, around 250 staff, 40,000 members — commissioned the programme in June 2025. The starting point was not a skills gap: chronic emotional exhaustion, collapsed cross-team trust, a workforce in defensive mode, and a 1.6 Yelp rating.

One member of staff: "Coming to work felt like being alone on a raft in the middle of the ocean." The pack's own diagnosis is the relevant one here — the organisation was "one setback away from catastrophic failure: not from clinical skill, but from emotional depletion."

After 90 days, an independent evaluation by Lauren Easterday of EastLight found all interviewees reporting greater emotional awareness and self-regulation, around 90% improved cross-team collaboration, 85% reduced emotional reactivity and 70% positive changes in patient interaction.

The behavioural detail is more persuasive than the percentages: hallway interactions moving from avoidance to acknowledgement to warmth; "people stop now, they don't just walk past"; "before I'd ask, 'What's your problem?' Now I stop and wonder what they're going through"; leaders asking questions instead of reaching conclusions in conflict; calmer encounters with frustrated patients; and staff applying the same tools at home.

The chief executive, Marsha Liu: "I've done leadership training before, but this felt fundamentally different. Champions shifted how people think about one another — and you can see it."

Two honest bounds, and the evaluation states the first itself: the percentages are "a qualitative synthesis across structured interviews… directional prevalence, not intended as statistical measurement".

And this is one non-randomised outpatient case study at 90 days. It is indicative, not generalisable — which is exactly how the CREW programme and Schwartz Rounds evidence should also be read. The right claim is that it points the same way as the wider literature, in a real organisation, quickly.

Humanity as a performance multiplier. This is the load-bearing claim, and it is the best-evidenced one in the document.

Incivility costs its target about a third of their performance and bystanders about a quarter of theirs. It makes colleagues 74–84% less likely to help. In a simulated operative crisis, 91% of clinicians passed in normal conditions and 64% when exposed to rudeness.

In Australia, half of hospital staff said unprofessional behaviour had a moderate or major negative effect on patient care; in the United States, 27% of practitioners knew of a medication error in the previous year in which disrespectful behaviour played a part.

And the causal chain is experimentally established, not inferred: rudeness degrades cognitive processing, and the degraded processing degrades the work.

Read that in reverse and it is a performance case, not a wellbeing one. Civility does not make a team feel better while performing the same; it restores diagnostic accuracy, information-sharing and help-seeking — which in a hospital are the safety system.

Staub's finding that "learning by doing can change people in either positive or negative directions" says the capability is trainable. Porath's finding that half of rude behaviour is driven by overload says the problem is a pressure symptom, not a character defect.

Together they say something specific: under pressure, humanity is the first capability to go and the one whose loss costs most — so it is the one worth building deliberately.

Which is why this is not a soft extra. A system missing every constitutional standard, with no capacity margin, cannot afford the performance it currently loses to how people treat each other. On the evidence assembled here, humanity is among the cheapest available sources of clinical performance the NHS is not yet using.

What the annexes add

Humanity is good for health, not only for performance. The annex on why being human is good for health weighs the evidence and finds it points one way. Social connection is as consistently linked to survival as anything in public health.

Warmth changes outcomes even when the treatment is identical. Being known by the same GP for fifteen years or more goes with 25% lower mortality in Norway's national data. Annex 7 · Why being human is good for health

And the firmest evidence is where people are most dependent: attentive care roughly halves delirium in frail older patients, person-centred care reduces distress in care homes at lower cost, immediate skin-to-skin contact cuts deaths among the smallest babies by a quarter, and children moved from orphanages into families develop better.

The effects are often modest and some popular claims do not survive checking. But none of it has been found to do harm, and almost none of it costs money.

How an organisation treats its staff is how its staff will treat patients. The annex on doctors and nurses traces the chain: teams broken up by rotas, staff asked every shift to give care they know is not good enough, and a workforce that learns it is on its own. Annex 6 · Doctors and nurses: when the organisation stops caring

In a 2021 BMA survey, 51% of doctors reported moral injury; in the 2025 NHS Staff Survey, only 43% felt their organisation valued their work, and a record low of 63% would be happy with the standard of care if a friend or relative needed it.

"I'm alright, Jack" is what an organisation gets back when it has told its people "you're on your own".

Structure is not a silver bullet — in either direction. The care emergency annex shows what happens when care is bought by the minute and children are placed with providers earning margins of over a fifth.

It also shows that joining the boxes does not join up the care: Northern Ireland's integration has been, in the Nuffield Trust's words, "statutory rather than delivered in practice". Annex 5 · The care emergency: social and children's care

That is the argument of this volume from another angle. Structure is what makes an identity survive a bad night, but structure without the identity is just another organisation chart.

Loneliness is the silent side of the same failure. The loneliness annex shows that isolation is linked to a 32% higher risk of early death, and that young people — not the old — now report the most loneliness.

Health and care are among the few places left where people meet face to face every day; a service that rotates staff and shortens visits removes connection from the people who most need it. The annex on joining up health and social care adds the policy lesson: coordination around the person helps, slowly; reorganising the boxes, on its own, does not. Annex 8 · Loneliness, the silent killer · Annex 9 · Joining up health and social care

Humanity compounds, and so does its absence. The annex on Wales traces the proposition's deepest roots: from 1731 a parson's schools taught a nation to read, its people built chapels they governed themselves, then libraries and a medical aid society, and Bevan carried that spirit into the NHS — choice, connection and creation, handed down through the generations. Annex 10 · Why we all love the Welsh

The annex on why a top-down model fails tells the other story from the same century: benevolence that decided for people hardened into control and a ranking of human worth, and reached into health and care through the institutions the NHS inherited. Annex 11 · Why a top-down model fails

Excellence that rests on one leader does not last. The annex on applying lessons from the past follows what happened to Salford Royal, twice rated Outstanding, after the leader who built it left and the trust was enlarged to rescue a failing neighbour — and sets the three pillars against it: the pathway for the individual, circles that belong to a network outside the employer, and accountable leadership. Annex 12 · Humanity-First Healthcare: applying lessons from the past

And honesty has to be paid for in protection, or it does not happen. The annex on radical leadership takes the mechanisms from aviation, from DuPont and from the NHS's own patient-safety rules: protection that is conditional on reporting early, a line drawn by behaviour rather than outcome, a guaranteed call for help that anyone can make, and forgiveness that stops at silence and retaliation. Annex 13 · Radical leadership

And an organisation has to be built for the leader it might get, not the one it hopes for. The annex on enablership sets the NHS's three-to-four-year chief executive tenure against the years a culture takes to set, and takes the design principle from the people who faced the same problem hardest: "If men were angels, no government would be necessary", and "enlightened statesmen will not always be at the helm".

It ends with four questions any board can use to test how much of its safety depends on who currently holds the post. Annex 14 · Healthcare: why the Art of Enablership is critical

The lesson for humanity-first healthcare is the same from both sides: power exercised with people builds something that lasts; power exercised over them, however well meant, ends in harm. Annex 10 · Why we all love the Welsh · Annex 11 · Why a top-down model fails

And without change, it gets worse. Chester is the warning. A system that could not protect newborn babies then protected itself — and the failure of governance and leadership continued with different people: in 2022, four of its new senior leaders ran a covert campaign against the chief executive brought in after the scandal, once she had raised a whistleblowing complaint.

Research on how harm escalates is clear that unchecked harm makes more harm likely. This is not a problem that stabilises on its own.

State of the nation

What is established. English healthcare has world-class clinicians working inside a system that repeatedly fails to convert their skill into safe care. Every major constitutional performance standard is being missed.

Capacity — beds, scanners, staff per head — sits below most comparable countries, and outcomes on cancer survival and avoidable mortality sit below them too.

Across at least eight major inquiries the proximate cause of catastrophic harm has been the same: concerns raised and reframed, candour delayed, regulators misled, whistleblowers isolated. Culture failure precedes clinical failure with enough regularity that it should now be treated as the primary risk, not a secondary one.

What is contested. Whether the professionalisation of nursing genuinely traded humanity for technical skill, or whether that reading is nostalgia for a more paternalistic era that had failures of its own.

Whether England's culture is actually worse than its peers' or merely better documented — the international data suggest the latter. Whether the current reform package is a genuine break or a more sophisticated compliance layer.

And, at Chester specifically, the safety of the convictions on which the Thirlwall Inquiry's remit rests: the inquiry was expressly not asked to re-examine the medical and statistical evidence, and a Criminal Cases Review Commission application supported by an international expert panel remains live.

That is a live public controversy, not a settled matter, and any use of Chester as a case study should acknowledge it.

It bears repeating that the Inquiry's findings about management, governance and safeguarding do not depend on it: the executives' conduct towards the doctors, the families and the regulators is documented in their own emails and minutes.

Two further things are established and easily missed. England is not internationally exceptional — every developed system reports comparable levels of bullying and suppressed challenge, and no system has solved this; what distinguishes England is the quality of its measurement and the visibility of its inquiries.

And the institutions that do sustain humane care at scale, from Mayo to the Aga Khan network, differ from England not in what they profess but in what stands behind it: salary policies, deeds of gift, bylaws, voluntarily sought external audit.

Values bind when they are rooted in identity and then built into incentives, appointments and constitutions, and not otherwise.

Three further things are established and easy to lose in the detail. Incivility compounds — it is contagious to people who were not involved, it degrades bystanders' performance as well as targets', and each unchallenged act moves the threshold, which is why intervention at the small end is the only cheap intervention available.

Technical excellence is no protection — the most scientifically advanced medical profession in history was also the most complicit in genocide, and the danger the Lancet Commission identifies belongs neither to that regime nor to science as such, but to any human system not rooted in values of humanity.

And the capability required is trainable: Staub's evidence is that "learning by doing can change people in either positive or negative directions", and policing has already built mandatory bystander training from that research across more than 430 agencies. Healthcare, which shares every relevant characteristic, has not.

And the conclusion that follows. If culture fails before care, if the capability is trainable, and if incivility costs a third of a target's performance and a quarter of a bystander's, then humanity is not a soft extra to be funded when finances allow.

It is among the cheapest sources of clinical performance available to a system that is currently missing every constitutional standard it has. That is the case for putting it first, and it is a performance case before it is a moral one.

And the charge the document opened with still stands. The problems are known. The remedies are known. The failures are organisational, which means they were preventable by the people running the organisations.

To know all of that and act on none of it is negligence, and at this scale it is negligence against the health and wellbeing of a whole nation.

What to watch over the next eighteen months.

1. Whether the HCPC barring legislation clears Parliament on schedule, and whether its scope widens from board level towards the all-manager barring Thirlwall asked for by September 2027.
2. Whether the criminal investigations at Chester and Durham produce charges. Nothing would alter boardroom behaviour faster, and nothing would confirm the old pattern more completely than their quiet closure.
3. Whether County Durham and Darlington extends its breast surgery look-back back to 2015 at its public board meeting on 24 September 2026, bringing up to 10,000 records into scope. This is the nearest test of whether a trust under scrutiny chooses the wider, more expensive, more damaging truth over the narrower one.
4. Whether the National Maternity and Neonatal Taskforce action plan, due December 2026, sets measurable commitments or restates principles.
5. Whether NHS Staff Survey bullying and speaking-up measures move at all. Twenty years of flat lines is the single best test of whether anything has actually changed.
6. Whether the National Audit Office actually begins auditing inquiry-recommendation implementation from September 2027, and publishes it openly. This is the single most testable of Thirlwall's 17 recommendations, and the one that determines whether the others survive.
7. Northern Ireland's diagnostic waiting list and emergency department performance, as retrenchment bites. It is the least-observed system in the UK and currently the most exposed.
8. Whether the Casey Commission, now expected to conclude by summer 2027, and the government's National Care Service commitment treat care as a relationship to be protected or as a structure to be redrawn. Northern Ireland's fifty years of integration suggest which of the two decides outcomes.

The one-sentence version. Diagnosed, not treated: England has built an unmatched capacity to investigate its own healthcare failures and has not yet built the capacity to prevent them; the reforms of 2026 attach real consequence to that gap for the first time, and whether they work will be visible not in frameworks but in whether staff start believing it is safe to speak.

Organisational failure

What two organisations did once the alarm was raised — Chester from 2016 and again in 2022, and Durham — and what getting it wrong costs in money. The one thing to take away: these were not rogue individuals but normal behaviour inside those organisations, which is why removing people has not changed it, and why the bill for negligence alone runs to billions a year.

Research synthesis

What. At Chester, what the hospital did after the alarm was raised is worse than the failure to act on it. Nurses on the unit were misled about why a colleague was being supervised; the suspected nurse was supported by the most senior team in the hospital, given the external review before the doctors were, and told "we have got your back"; the doctors were told they had conducted a witch hunt, made to apologise, threatened with referral to the GMC, and made the subject of a written plan to "manage out"; the police were not called for a year.

At County Durham, warnings arrived for thirteen years, duty of candour was owed from May 2024 and delayed to February 2025, and the executive response to a governance review was to try to identify who had spoken to it.

So what. These were not rogue individuals — the behaviour was normal inside those organisations. Four executives in four different roles behaved the same way, none of them corrected by anyone above them, and the medical director's parting instruction to his successor was to refer the paediatricians to the GMC — said openly, as though it were obviously the right thing.

That is what a rotten system looks like from the inside: not deviant behaviour, but expected behaviour.

Thirlwall's own system finding is the same — "a culture of avoiding dealing with poor performance" and a "revolving door" that "allowed the movement of failed managers into new roles, irrespective of the reasons for their failure".

Now what. Barring individuals is designed for rogue actors, so it cannot be the whole answer. If the behaviour is what the system rewards, barring is necessary and nowhere near sufficient — and it is triggered by a disciplinary process the organisation itself has to start.

Read this annex as the evidence for the four root-and-branch changes in Volume 9, and for the cost section at the end: the money says the same thing the conduct does.

A note on names and live proceedings. Individuals are named here only where the Thirlwall report names them and makes findings about them; quotations are the report's. Cheshire Constabulary's Operation Duet continues into both corporate manslaughter and gross negligence manslaughter.

Three people described as part of the senior leadership team were arrested on suspicion of gross negligence manslaughter on 1 July 2025 and remain on bail, and a further person was arrested in April 2026 on suspicion of perverting the course of justice.

Police have not named them, and nothing here attributes any criminal conduct to any individual. Nothing has been determined against anyone.

Case one — Chester: what the hospital did next

The murders and attacks are Volume 3. This is the organisational conduct that followed, from the paragraph numbers of the Thirlwall report.

July 2016: the supervision that was not what staff were told. Letby was told on 14 July that she had been identified as "more regularly involved in the care" of the babies who collapsed. An email to all nurses said "all members of staff need to undertake a period of clinical supervision … Therefore Lucy has agreed to undergo this supervision first".

The Chair: "it was only ever Letby who was going to undergo supervision" (17.5), and "Ms Sian Williams — the Deputy Director of Nursing — was misled. All the nurses on the unit were misled as to the reason for supervision" (17.5).

On 18 July she was redeployed to the Risk and Patient Safety Department, described in writing as "temporary" and "a neutral act … taken in the best interest of all parties" (17.6–17.7). The Director of People's recollection of why she was not excluded: "there is not enough to exclude her, suspend her" (17.9).

Then the hospital's senior team supported her, personally. Between March 2017 and July 2018 one manager and Letby "exchanged over 400 WhatsApp messages" (17.34); a group of four ran to "over 750 messages" between April 2017 and July 2018 (17.36).

"Letby also had weekly or fortnightly meetings with Ms Hodkinson, Ms Kelly and Ms Rees. This was a very senior team, which supported Letby throughout (with Ms Rees as mentor). They made it plain to her that the expectation was that she would be returning to the neonatal unit … and they ensured she had a copy of the RCPCH report before the consultants did" (17.35).

The Chair: "Ms Kelly's participation in the meetings was ill judged. As a senior manager, she should not have been involved in supporting Letby" (17.39); and "every person who supported her stepped very quickly from professional support to personal friendship" (17.41).

The grievance, and the plan to put her back on the unit. The grievance outcome of December 2016 provided that the chief executive and a non-executive director would apologise to Letby in front of her parents (22.110); that she would be told she had "no case to answer" — a finding the Chair says "she was in no position to make" (22.111); that both consultants would apologise and mediate (22.113); and that "Failure to comply with these reasonable measures from the named parties, should result in disciplinary action taken by the Trust" (22.114).

At the meeting of 22 December 2016 the chief executive is minuted saying "We are within our rights to phone the police but we didn't believe it" and "Our commitment is now to meet with the consultants to get you back on the unit" (22.123).

On 10 January 2017 he told her "the Board were clear in their support for her returning" (23.32). On 6 February: "Lucy, don't worry, we have got your back" — the Chair: "To say 'we have got your back' is unambiguous … If it was to 'avoid any possible escalation', it was misconceived and dishonest" (23.61).

On 31 January 2017 Letby emailed all neonatal staff saying "I have therefore been fully exonerated. I have received a full apology from the Trust … I will begin making my return to the Unit". Nobody corrected it: "This was a further failure of management" (23.45–23.46).

By 28 March 2017 the plan was still to return her to the unit from the week commencing 3 April, for an hour a day (26.31). And the unit was told she was coming back: the chief executive called the staff together to say so, a manager proposed a tea party to welcome her, and one nursery nurse's written evidence is six words — "We did cakes and tea."

Letby came to the tea party (23.79; the nurse places it "around 12 months after" Letby was removed from the unit, and the Inquiry records it in the chapter covering January to March 2017). The staff were not being careless.

They had been told by their own chief executive that she was returning, and Letby's email of 31 January had told them she was "fully exonerated".

And the doctors who had raised the alarm were the ones handled. The phrase "witch hunt" first appears in an executive's note of late June 2016 — "Concerned witch hunt for Nurse" (14.46). The Chair: "There was no basis for the assertion that this was a 'witch hunt' … Declaring the expression of their honestly held beliefs a witch hunt was offensive but, far more important than that, it was a distraction from what the doctors were saying" (14.52).

One consultant was required to apologise for a remark he had not made: "the first time I knew that there had been any concern with what I had said was when Ian Harvey rang to ask me to apologise seven months later", and he was told he "hadn't [a] choice" (22.115–22.116). The Chair: "That he was required to apologise in writing to Letby was ridiculous" (22.118).

Of Dr Brearey being asked to apologise: "I can find no reason for his being chosen other than that he had raised concerns about Letby" (22.114). Of the whole episode: "the way the grievance and its consequences were handled was deplorable" (22.129).

At a meeting on 26 January 2017 the chief executive asserted the speaking-up process had been "professionally managed" and the medical director said the review "does not call out a criminal act" and there was "a need to draw a line under the 'Lucy issue'" (23.34).

Dr Jayaram's recollection: "I am drawing the line under it, you will draw a line under it and if you cross that line, there will be consequences" (23.36). The BMA representative present called the meeting "shocking" and the tone "dictatorial", and felt "deep concern" for the consultants "because there was potential that their jobs were at stake" (23.37).

Then the regulator was turned into a threat. On 1 March 2017 the medical director emailed both doctors that mediation "would also go a long way to protect you from a possible referral to the GMC from other parties" (26.6).

A note of external legal advice sought on 30 March records "? Consider refer cons [consultants] to GMC" — the Chair: "there was no question in the minds of the executives of protection for the doctors.

The contrary was the case. They were at risk of referral by their employer to the regulator" (26.35–26.36).

And then, on 12 May 2017, a note of the chief executive's plan: "RJ/SB plan re management. 1. GMC 2. Actions from grievance 3. Mitigation from SoS/whistleblowing 4. Action plan to manage out" (26.87).

The Chair: "This note is one of the most revealing documents in the Inquiry. It sets out an aggressive plan to get rid of Dr Brearey and Dr Jayaram" (26.88); "his obvious intention was to punish the consultants for persisting with their concerns" (26.91).

What stopped it: "What caused him to change course was the fact that, a few days later, contrary to his expectation, the police announced they were going to investigate" (26.91).

The protection that was supposed to exist did not operate. The concerns "were not … formally logged until 24 April 2017" (19.17), "at least seven months after every senior executive knew" (19.18), and "Despite raising concerns in good faith, the consultants were deprived of protection from recriminations under the Speak Out Safely Policy" (19.28).

On 16 March 2017 the executives were told the consultants "felt bullied, intimidated and victimised" — "Yet no action was taken under the policy" (19.19). Asked by police why they had not called them, Dr Jayaram's answer was: "Our career would be on the line if we contact [the] police, it would be whistleblowing" (26.26).

The year of delay, in the Chair's own words. "Almost everything the consultants raised with the police in May 2017 was known to the executives after the deaths of Baby O and Baby P … Responsibility for that delay lies squarely at the door of the executives" (26.97).

"The effect of the actions of the executives from June 2016 was that the police investigation was delayed by a year" (26.99). And: "Even at the Inquiry, Ms Kelly could not bring herself to say that the executives should have gone to the police then" (26.100).

When the chief executive did write to the Chief Constable on 2 May 2017, he asked for an investigation "with a view to excluding any unnatural causes" — which the senior investigating officer thought "had no place in the letter" and read as an attempt to "direct a mindset" (26.74).

Of the executives' subsequent briefings, the same officer: "it's like indeed doors are trying to be shut" (26.84). The police decided to investigate after meeting the doctors, on 15 May 2017 (26.93–26.95).

Everyone outside the hospital was given a partial picture. The NMC was told one registrant "has been present at nearly all these incidents" but not, the Chair finds, "what was known about the unexpected and unexplained nature of the deaths nor … the fact that, when Letby had been moved off the night shift, the unexpected deaths stopped happening at night" (18.26); a later update was "an economical approach to the facts" (18.27) and a 2017 call "misleading" (18.31).

The RCPCH produced two versions of its report, the second confirmed to have been made "purely to protect 'individuals for the HR issues'" (20.136) — "Providing a dissemination version to be shared with the paediatricians in circumstances where it was silent on the principal concerns raised was wrong" (20.138).

NHS England was told the review team "assured us that there were no immediate actions or concerns": "It was misleading. I am satisfied that its purpose was to mislead" (25.61). Its own people said so at the time — "I do believe that the Hospital was being evasive and uncooperative" (25.117), "I think the Trust executive team controlled the narrative so that it was hard to unravel the issues" (25.118), "I have not since experienced the same level of lack of cooperation from a hospital" (25.114).

The coroner was not told of the suspicions at all (24.100), and the note of one pre-inquest discussion recorded the desired outcome as "Narrative Verdict 'Unascertained' No negative comments No press" (24.99).

The Chair's one-paragraph summary of that year (14.84): "During that year, nurses on the neonatal unit were misled about what was going on. The RCPCH took on a task for which it was ill-equipped, the product of which was to be relied on much later as exonerating Letby when it did nothing of the sort.

Letby brought a grievance which was badly handled and ultimately led to the doctors being asked to apologise and/or mediate. The police were not called until Mr Chambers could see that he had no choice nearly a year later.

In the meantime, his focus was on bringing the doctors to heel, moving on, and drawing a line under the concerns about Letby."

The review the hospital commissioned that year has an annex of its own, because the outside check is supposed to be the backstop. The Royal College's Invited Review Service is peer review sold for a fee — in the College's own evidence, "not a regulatory process or a forensic investigation" — so the one question the Countess most needed answered was the one question the review could never answer.

What matters is what happened to that limitation once the review was commissioned: it disappeared, and the reassurance was used to close the question down. Annex 16 · When the sheriff gets it wrong

And the arrest did not end it. Dr Susan Gilby became medical director on 1 August 2018, three weeks after the arrest, and found "there was a denial that this meant that it needed to be taken really seriously … the focus continued to be on the people who had raised concerns" (27.18).

The chief executive "believed, in spite of the arrest, that no deliberate harm had been caused. He kept repeating that there was no single cause found" (27.19). The medical director briefed her that the reviews had found no deliberate harm and left without giving her the chance to read them: "He sought to control the narrative" (27.21).

When eight paediatricians wrote in April 2018 that "the chief executive and some members of the board treated the consultants' views with contempt", the draft reply claimed the neonatal team had "his full support and respect … and this has never changed" — which Gilby called "tone deaf" and the Chair calls, flatly, "untrue" (27.14).

Case one, continued — what happened to the people responsible, 2018 to 2026

This is the part that is hardest to read, and it is the part that decides whether the reform programme in Volume 9 is proportionate to the problem.

Ian Harvey, medical director. He announced his retirement and left on 21 August 2018 — a month after the arrest. His parting remark to his incoming successor, as she gave it in evidence and as the Chair accepts it: "as we walked out the door, he's packing his things away and he said to me 'You need to refer those paediatricians to the GMC'" (27.26). He denied it emphatically; the Chair: "I accept Dr Gilby's evidence about it."

He was referred to the GMC by the paediatricians at the end of August 2018. The GMC concluded that "whilst some of Mr Harvey's communication with the paediatricians fell below the standard expected, there was no realistic prospect of establishing that Mr Harvey's fitness to practise was impaired … The case was closed with no further action in May 2022.

Mr Harvey reapplied for voluntary erasure from the Medical Register and this was granted in June 2022" (27.28).

Tony Chambers, chief executive. He stepped down on 19 September 2018, ahead of a likely vote of no confidence. On 17 September NHS Improvement had been tasked with finding him a post (27.36–27.39); he secured a position at another trust with his notice paid by the Countess, and "for nine months, the Countess … funded Mr Chambers' salary package, which, including pension contributions and benefits in kind, amounted to £290,000 to £295,000 ($391,500 to $398,250), with no benefit to the Countess" (27.45).

The Chair also found it "troubling that public money could be spent in this way without any analysis of the reason for the move … and no consideration of whether Mr Chambers was a fit and proper person for the role he was taking on" (27.47).

The Trust's announcement said his departure "is not a judgement on his ability as a CEO but more a reflection of his integrity as a leader". The Chair: "That was pure spin" (27.48). His own evidence to the Inquiry was "I stand by the decisions that we made", of which the Chair says: "The lack of reflection and the failure to take any responsibility for the delays and the heartache caused to the families by his approach to the concerns about the deaths of their children is indefensible" (27.49).

Alison Kelly, director of nursing. "Ms Kelly remained in her role as Director of Nursing until she left the Trust in June 2021" (27.54) — nearly three years after the arrest. She subsequently held a nursing director post at another NHS trust, from which she was suspended in August 2023 following the conviction, and the NMC confirmed it would proceed with a fitness-to-practise investigation that had been paused at the request of the police.

No published outcome has been found. (Those last two facts are from contemporaneous news reporting, not from the Inquiry.)

Sue Hodkinson, director of people and organisational development. The report records her evidence but gives no departure date, no subsequent role and no regulatory action. On having continued to push for Letby's return even after being told about Baby K in March 2017, her reflection was "I apologise for that" — the Chair: "The apology did not begin to explain why she had continued with her efforts" (26.33).

Sir Duncan Nichol, chair. Stood down on 31 March 2020 (27.53). He accepted in evidence that the hospital did not exercise the "appropriate duty of candour" towards parents, which was a "serious failure" (26.67).

And the person who exposed it internally was the one who lost her job. Dr Susan Gilby was "unlawfully excluded from the hospital" on 2 December 2022 and "was later successful in claims against the hospital for constructive unfair dismissal, automatic unfair dismissal and detriment on the ground that she had made protected disclosures"; the Employment Tribunal judgment of 13 February 2025 "contains excoriating criticism of the leadership of the Countess at that time" (27.10). Four years after the arrest, the pattern was still running.

Project Countess: how it was done, and what it cost

The campaign had a name. Dr Gilby had been medical director from 1 August 2018, three weeks after the arrest, and became chief executive after her predecessor stepped down.

In July 2022 she made a whistleblowing complaint about the conduct of the Trust's new chair, Ian Haythornthwaite — a former BBC chief finance and operating officer who had taken up the post in 2021, after the chair who presided over the 2016–18 events had stood down — alleging bullying and harassment.

The Employment Tribunal found that the chair, the deputy chair, a further non-executive director and the chief people officer then ran a covert campaign to remove her, conducted through private emails and WhatsApp messages, which they called "Project Countess".

In November 2022 she was offered a secondment on terms that required her to withdraw the allegations; she declined. On 2 December 2022 she was suspended — Thirlwall calls it being "unlawfully excluded from the hospital" — and on 5 December she gave notice.

A director's words to her, as recorded: "time for you to go … if you don't agree to go, we will start a process against you."

The findings about the evidence are the part that should worry a board. Employment Judge Shotter found that hundreds of messages and documents had been deleted, that the evidence "points to a deliberate intention to hide documents", and that it was "more likely than not" that the chair was involved in the deletion.

She described him as an "inaccurate historian" who did not give credible evidence, and made comparable findings about the deputy chair, Ken Gill, whom she called a "key player" in the plot and "disingenuous".

Judgment was handed down on 13 February 2025. Thirlwall records that it "contains excoriating criticism of the leadership of the Countess at that time" (27.10). The BMA, which supported the case, called her treatment "distressing torment" and "nothing short of disgraceful".

What it cost, and what happened to the people involved. The chair resigned with immediate effect within days of the ruling. The deputy chair, the non-executive director Ros Fallon and the chief people officer Nicola Price all left the Trust.

The case settled in January 2026: £1.4m ($1.89m) in damages to Dr Gilby, and a total cost to the Trust of about £3m ($4.05m) once legal costs are counted.

Ms Price had moved in January 2024 to Mid Cheshire Hospitals NHS Foundation Trust as chief people officer; she resigned from that post following fit-and-proper-person concerns, and the CIPD — her professional body — imposed its most severe sanction, permanent lifetime exclusion from membership, for failing to investigate a grievance properly, participating in the effort to remove an employee and delete documentation, and dishonesty in the tribunal proceedings.

Why this is the most damning single item in the accountability record. Everything earlier in this annex happened before anyone had been convicted, when the organisation could at least tell itself it did not know.

This happened afterwards, in the same institution, to the person who had exposed it — with a public inquiry already in prospect.

The board's answer to a chief executive who had told the truth about the organisation's past was a covert campaign, the deletion of records, and an attempt to buy her silence with a secondment.

Nothing about the mechanism had changed in six years: the concern was still treated as the problem, and the person raising it as the thing to manage. What is different is that this time accountability arrived — and it arrived from an employment tribunal and a professional body, not from anywhere inside the NHS.

Regulators and inspectors over the same period. The CQC rated the Trust Requires Improvement after its November–December 2018 inspection (report published May 2019), a downgrade from Good; and Requires Improvement again after the February–March 2022 inspection, finding that the trust "did not have suitable governance systems and processes".

It issued two warning notices in 2022 requiring significant improvements to the safety and quality of maternity services, and carried out a focused follow-up inspection that July. (CQC inspection index; not an Inquiry finding.)

The criminal position, as at 15 September 2026. Cheshire Constabulary opened a corporate manslaughter investigation into the Trust in October 2023. Three people described as part of the senior leadership team were arrested on suspicion of gross negligence manslaughter on 1 July 2025 and remain on bail; a further person was arrested in April 2026 on suspicion of perverting the course of justice and also remains on bail. Operation Duet continues on both strands. Nobody has been charged.

In February 2025 counsel for the four former executives applied to pause the Inquiry pending the Criminal Cases Review Commission application; the Chair refused.

The score, then. Of four executives whose conduct the Inquiry criticises in detail: none was dismissed, none was subject to a disciplinary process recorded in the report, none has been barred, and one was funded for nine months by the hospital he left, into a job found for him by the national regulator.

The one clinician among them left the medical register voluntarily with his case closed. The doctors who raised the alarm were threatened with the GMC; the medical director who threatened them was referred to it and the case was closed with no further action.

The point this makes, and it is not about individuals

It would be comfortable to read Ian Harvey as a rogue actor. He was not. He was a product of the system he worked in, and his own words on the way out are the clearest evidence of it: a man who believed he was doing something indefensible does not stop, on his last day, to instruct his successor to report the whistleblowers to their regulator.

He said it because inside that organisation it was the obvious thing to say. The belief was not private, it was normal — shared, uncorrected, and consistent with everything the executive team had done for two years.

That is what "rotten to the core" means as a structural claim rather than an insult. Four people in four different roles — chief executive, medical director, director of nursing, director of people — behaved the same way at the same time, on the same instincts, with a board above them that did not intervene and regulators around them that accepted what they were told.

A system that produces identical behaviour in four different people is not only a system with four bad apples in it. It is a system whose ordinary incentives produce that behaviour, and which had no mechanism capable of interrupting it until the police arrived.

Thirlwall reaches the same conclusion as a national finding, not a local one:

• "it is inescapable that there is a culture of avoiding dealing with poor performance" (37.60).
• The "revolving door … allowed the movement of failed managers into new roles, irrespective of the reasons for their failure" (37.61).
• Quoting Mr Kark KC's evidence: "part of that settlement agreement is a vanilla reference. The director moves on to the next Trust down the road. This state of affairs is accepted with almost a collective shrug of senior shoulders. It should not be tolerated" (37.62).
• And the gap that remains in the fix: "if a hospital does not invoke the disciplinary process and dismiss for serious misconduct, then, on the current proposals, there would be no basis for barring the manager" (37.79).

Read those four findings together and the limit of the reform becomes obvious. Barring is triggered by a disciplinary process that the organisation itself has to start — and the organisations in this document did not start one, because the people who would have had to start it were the people concerned. Chester ran the opposite process: a grievance brought by the suspected nurse, against the doctors.

Case two — Durham: the same mechanism, thirteen years long (and here the harm was avoidable, not deliberate)

The Aubrey review's chronology is the clearest available picture of how an organisation can be told, repeatedly, and not act. Dates are the report's own; individuals are referred to by role, as the report does.

The verdicts that matter, verbatim. "The Board and its committees accepted reassurance rather than seeking genuine assurance, allowing unsafe practices and systemic weaknesses to persist unaddressed" (p.10).

"Executive leadership and senior management were not sufficiently focused on quality and outcomes, often prioritising activity and performance metrics over patient safety and learning" (p.11).

"Governance structures were bypassed, escalation pathways were ineffective, and clinical leadership was allowed to operate without oversight. This resulted in a culture of complacency, at least 27 missed opportunities, and outdated practice, which led to avoidable harm" (p.16).

"The evidence is unequivocal: The failures identified were systemic" (p.23). "Leadership and governance failed at every level: repeated warnings were ignored" (p.185). And on the board: "Despite repeated warnings and clear evidence of patient harm, the Trust Board and Non-Executive Directors did not provide effective challenge or scrutiny… contributing to a culture of tolerance for underperformance" (p.187).

On candour specifically, the review is unusually blunt for a document commissioned by the organisation it examines: "The review found clear evidence of obfuscation by the Executive Medical Director, who led on the Duty of Candour process… Despite the statutory threshold for Duty of Candour being met as early as May 2024, the Executive Medical Director failed to initiate the required processes, did not follow the clinical governance framework, and excluded Patient Safety Leads from involvement.

Instead, the Executive Medical Director relied on external advice to justify the inaction" (p.122).

Its recommendation 14 invites the Board to "formally consider whether continued appointment to this position remains appropriate in light of the evidence presented" (p.207).

Three features of this case are worth holding onto. Nobody audited the surgeons except the surgeons: "Breast surgeons audited their own practice without independent validation against national standards" (p.187), and internal reviews "were carried out by the same consultants whose practice was under scrutiny" (p.90).

And the reflex when the review began was control rather than candour: "Attempts by some Executive Directors to identify contributors not only breached this understanding but also reflected a leadership approach perceived to prioritise control over accountability" (p.13).

Staff had already said they could not "speak up to their manager", described a "chilling effect", and reported being warned that speaking up "could adversely affect their careers".

And the money flowed to the person whose practice nobody was auditing. About half of all patients referred into the Trust were seen outside standard Trust clinics — in insourced evening and weekend clinics and on Saturday theatre lists delivered by the same consultant clinical lead, who was a director of the two private companies providing them.

Payment was per patient, "creating a clear financial incentive" (p.91), and neither arrangement was subject to routine audit, formal quality monitoring or the clinical governance structures applied to NHS services.

Nearly £6m ($8.1m) was paid over six years to the diagnostic clinics and surgery companies the clinical lead ran privately.

The chief executive has since accepted that "there was a conflict of interest that wasn't managed properly".

Nor did the external check catch it: the CQC asked about breast services in April 2018, and the Trust's response omitted a known complaint and a known incident; the inspection that followed rated the Trust "Good" for well-led, and that rating stood until June 2026.

The structural conclusion Aubrey draws is the same one Chester's record supports: "A central finding of this review is the absence of a Board-level Director of Governance with sole, independent responsibility for governance, compliance, and assurance… governance was subordinated to operational and clinical portfolios, undermining its effectiveness as a protective mechanism for patients and staff" (p.197).

The precedent she cites for the fix is Morecambe Bay — which is to say, a previous inquiry, whose lesson had been available for a decade.

What these cases establish together

1. The cover-up is not a separate event from the harm; it is the same failure continuing. At Chester the conduct after the concerns were raised cost a year, and the last baby was harmed inside that year. At Durham the candour delay cost nine months while patients went un-recalled.
2. The organisation's defensive instinct survives exposure. Chester's executives were still managing the narrative after the arrest; Durham's were still trying to identify who had spoken to the reviewer during the review.
3. Individual accountability arrives late, partially, or not at all — and every route to it in both cases was external: police, an employment tribunal, a coroner, a commissioned review, an inquiry. Not one was internal.
4. The people who raised concerns paid, and the people who did not act did not. That asymmetry is the single most corrosive fact in this document, because it is the one every member of staff can see, and it is the one that teaches the next person to say nothing.
5. Therefore barring individuals cannot be the whole answer. If the behaviour is what the system rewards, the fix has to change the reward. That is the argument for the four changes in Volume 9 — what leaders are measured and paid on, who leads and how close they are to care, where candour sits, and whether capacity is honest.

The financial cost of getting it wrong

The human cost is Volume 3. This section is the money, because the third option in the three-way choice — keep commissioning inquiries — is the most expensive of the three and the only one that buys nothing.

Every figure below is dated and sourced, and where a commonly quoted number does not survive checking it is corrected rather than repeated.

Clinical negligence: the national bill

The second largest provision on the entire government balance sheet, after nuclear decommissioning, is the NHS's liability for harming its own patients.

Three corrections to figures in circulation. First, the widely reported "£60bn" ($81bn) is a provision — the actuarial present value of all incurred claims, reported and unreported — not a bill and not an annual cost. Dividing it by an annual budget is a category error that appears constantly in coverage.

Second, 13,329 is claims resolved in 2024/25, not claims received; received was 14,428 that year and 15,236 in 2025/26.

Third, there is no published average cost per claim. The £230,000 ($310,500) figure is £3,088.9m ($4.17bn) of cash paid divided by 13,329 claims resolved — two different flows, since much of the cash relates to claims settled years earlier.

The same method on 2025/26 data gives £273,000 ($368,550). Use instead the figures NHS Resolution and the NAO do publish: 48% of claims settled with no damages at all, and claims of £1m ($1.35m) or more were 2% of volume but 68% of cost.

The Public Accounts Committee's verdict, 30 January 2026: the Department "has failed to tackle the rising costs of clinical negligence despite repeated warnings", and legal costs remain "unacceptable".

The cost of finding out

Police cost is largely invisible, which is itself a finding. Cheshire Constabulary answered an FOI request about expenditure on Operation Duet — the corporate manslaughter and gross negligence manslaughter investigation — with "no information held".

A figure of around £9.7m ($13.1m) to September 2023 for Operation Hummingbird, the original criminal investigation, appears in a Cheshire Police and Crime Panel paper but could not be verified directly.

For the Shrewsbury and Telford maternity investigation, only the funding is public: £1.6m ($2.16m) from government plus over £720,000 ($972,000) from the Police and Crime Commissioner for a single year, for an inquiry into roughly 600 cases.

Gosport's investigation reviewed the records of more than 750 patients and took 1,200 witness statements; no cost was published and nobody was charged.

And there is the opportunity cost nobody invoices for. Detective time spent for years reconstructing what a hospital already knew is detective time not spent on other crime. The same applies to the coroners, the regulators, the lawyers on both sides, and the clinicians who spend years giving evidence instead of treating patients.

Thirlwall's own record shows one consultant spending "most of the last ten years or so reliving many of the events about which he gave evidence".

The employment bill: what burnout and attrition cost

The agency reduction is real and worth noting honestly — spend has fallen sharply. But the per-nurse replacement figures often quoted ($61,110 per bedside nurse, $289,000 per percentage point of turnover) are American, and there is no published national count of NHS employment tribunal claims and no published national figure for NHS spend on settlements or non-disclosure agreements. Anyone quoting one is aggregating FOI responses.

What it costs elsewhere when an organisation fails at scale

The pattern across all of them is the same and it is worth naming: the longer an organisation denies, the more of the eventual bill goes to lawyers rather than to the people it harmed. Coal spent more than £1 on process for every £2 that reached a miner.

Horizon's claimants received about a sixth of their own settlement. In the NHS's low-value negligence claims, legal costs run at 3.7 times the damages. None of that is the cost of the harm. It is the cost of contesting it.

The Durham illustration — what one department could cost

This is the scenario that should concentrate any board's attention, and it has to be handled carefully because it is arithmetic on an unfinished review, not a forecast.

What is known: of 514 cases reviewed in the County Durham breast surgery look-back, 315 patients were harmed (61.3%), of whom 77 were significantly harmed (15.0% of those reviewed), including 20 unnecessary mastectomies and one known death.

The review currently covers January 2023 to February 2025. The Trust has been considering extending it back to 2015, which would bring up to 10,000 records into scope, with a decision due at the public board meeting on 24 September 2026.

If — and it is a very large if — the same rates held across 10,000 cases, that would be on the order of 6,100 patients harmed and 1,500 seriously harmed from one department in one trust. Applying even the crude £230,000–£273,000 ($310,500–$368,550) per-claim ratio criticised above gives a claims bill of £1.4bn–£1.7bn ($1.9bn–$2.3bn).

Three caveats belong with that number every time it is used: the harm rate in the earliest cases reviewed is likely to be higher than in a full population, because the review began with the cases flagged as highest risk; harm is not the same as a successful negligence claim, and nationally 48% of claims settle with no damages; and the per-claim ratio is not a published average.

But damages are only one line of the bill, and on this kind of failure they are not the whole of it. Four other lines follow automatically, and none of them is optional once the look-back is opened:

And the coal precedent says the administration can rival the compensation. The British Coal respiratory disease and vibration white finger schemes paid £4.1bn ($5.54bn) in compensation against £2.3bn ($3.11bn) in administration, £1.3bn ($1.76bn) of it in solicitors' fees — the Public Accounts Committee's summary in March 2008 being that "for every £2 paid out in compensation more than £1 has been spent on administration".

That is the closest structural analogue available: a mass look-back, hundreds of thousands of individually assessed claims, most of modest value, resolved over years. Apply that ratio to a £1.4bn ($1.9bn) claims bill and the total approaches £2bn ($2.7bn).

So the honest statement of the scenario is not "about £1bn". It is that one service, in one trust, over one decade, is capable of generating a total cost well in excess of £1bn ($1.35bn) — and on the coal precedent, plausibly approaching twice that — with a substantial share of it going to lawyers rather than to the women who were harmed.

The defensible version of the point does not need the arithmetic to be precise. One service, in one trust, over one decade, is capable of generating a claims volume approaching half the NHS's entire annual intake of clinical negligence claims — 15,236 in 2025/26 — on the implausible assumption that every harmed patient claims.

And not one pound of that bill buys a single patient any care. Almost all of it is the cost of not having acted in 2012, when the GMC withdrew the trainees.

The summary a finance director would recognise

Put the strands together and the economics of not acting are indefensible on their own terms.

• The NHS carries a £60.3bn ($81.4bn) provision for harm it has already caused, and pays out £3.2bn ($4.3bn) a year, rising to a projected £4.1bn ($5.5bn) by 2029-30. A quarter of that annual spend is legal costs.
• Finding out what happened cost £18.6m ($25.1m) and counting for Chester alone, on top of an uncosted multi-year police investigation and whatever the trust spent on its own lawyers. The median statutory inquiry runs 3.5 years at £6,600 ($8,910) a day.
• Depleted, burnt-out staff cost an estimated £12.1bn ($16.3bn) a year in absence, presenteeism and temporary cover — four times the negligence payout bill.
• And the comparators in other sectors say that once a failure of this kind is finally admitted, the bill is measured in billions, with a large fraction of it going to lawyers rather than to the harmed.

Against those numbers, the cost of building the capability described in Volume 10 — the cheapest of the three options, and the only one available this financial year — does not register as an expense. It is the only line in this document that would reduce the others.

UK healthcare in numbers

Orientation for a reader who is not from the UK — and a reference for one who is. Everything is dated. England and the UK are never used interchangeably: health is devolved, so "the NHS" in most of this document means NHS England, which covers 85% of the population.

Sterling figures carry a US dollar equivalent in brackets at $1.35 = £1 — a convenience conversion at a single rate, not a purchasing-power comparison.

The one thing to take away: the UK's system is universal and mostly paid for from taxation, but it runs with fewer beds, scanners and staff than comparable countries, it is missing every major performance standard, and its social care is separate from the NHS, means-tested and chronically short of money — which is where many of the failures in this document begin.

The ten numbers that describe the system

If you take one thing from this annex, take this. The UK does not have one care system. It has five — tax-funded healthcare, means-tested social care, a private market, a charitable hospice sector, and several million unpaid relatives — and they are funded on entirely different principles, regulated to entirely different standards, and scrutinised to entirely different degrees.

The failures in the rest of this document are concentrated in the one that is best measured, best funded and most inspected. That should be read as a warning about the other four, not as reassurance.

How the parts fit together

Seventy-eight years of reorganisation, and of the same findings

The timeline's horizontal scale expands after 1990 so that recent events are legible; decade ticks show where it changes.

Where the NHS came from, and what it already says it stands for

The NHS was founded on 5 July 1948, under the National Health Service Act 1946, and launched by the Health Minister Aneurin Bevan at Park Hospital in Davyhulme, Manchester — now Trafford General. Seventy-eight years old, and for most of that time the largest employer in Europe.

The founding premise was moral before it was administrative: that illness is a misfortune rather than a fault or an indulgence, and should not carry a financial penalty. Care available to all, free at the point of need, funded from general taxation rather than insurance contributions — Bevan rejected Beveridge's insurance model precisely because people who could not work would never accumulate enough contributions to qualify. The 1948 household leaflet put it in one line: "It is not a charity. You are all paying for it."

And the values are not missing — they are written down. The NHS Constitution for England sets out seven principles and six core values, developed with patients, the public and staff:

• Working together for patients — "Patients come first in everything we do."
• Respect and dignity — "We value every person – whether patient, their families or carers, or staff – as an individual."
• Commitment to quality of care — "insisting on quality and striving to get the basics of quality of care – safety, effectiveness and patient experience – right every time."
• Compassion — "respond with humanity and kindness to each person's pain, distress, anxiety or need."
• Improving lives — "We strive to improve health and wellbeing and people's experiences of the NHS."
• Everyone counts — "we maximise our resources for the benefit of the whole community, and make sure nobody is excluded, discriminated against or left behind."

Hold that list against the rest of this document. Every inquiry in it describes the failure of at least two of those six, in an organisation that had them on the wall.

The gap is not aspiration — it is that nothing structural stands behind the aspiration. Volume 10 makes the argument in full: Bevan built the financial half of the 1948 promise into an Act of Parliament, and left the humane half to chance.

Scale and money

Free at the point of use, funded from general taxation, and covering essentially the whole population — which is the feature that makes the failures in this document a matter of national rather than commercial accountability.

The workforce

Registration and employment are different counts: the registers include people working privately, through agencies, part-time, abroad or not at all.

How it is organised

England is the odd one out among the four nations. It alone keeps a purchaser/provider split: integrated care boards commission care from trusts and from non-NHS providers. Scotland (14 territorial health boards), Wales (7 local health boards) and Northern Ireland (5 HSC trusts, with health and social care in one system) plan and provide directly, with no internal market.

One consequence runs through this whole document: England generates far more data about itself, and therefore far more visible failure.

One change to watch, because it is not yet done. The abolition of NHS England as a separate body, and its merger into the Department of Health and Social Care, was announced in March 2025. The Health Bill cleared the Commons on 8 September 2026 and is now in the Lords, with a target of March 2027 and a 50% headcount cut across DHSC, NHS England and ICBs by March 2028. Treat it as pending, not as fact.

Public and private

The private sector is real but small, and mostly complementary rather than competing.

• UK independent healthcare market: £13.8bn ($18.6bn) in 2024 — £7.2bn ($9.7bn) acute hospitals, £5.7bn ($7.7bn) clinics and private practice.
• 12.2% of the UK population (8.43 million people) has private medical cover; self-pay is nearly a quarter of independent hospital revenue.
• Independent providers delivered 7.5% of NHS elective care in England in February 2026, down from a peak of about 10% in late 2024.
• Private hospitals are built around planned, single-specialty work — which is exactly why the Branson case in Volume 7 matters: money removes the excuse of scarcity, and the failure still happens.

Dentistry: the part of the NHS that has already gone private

Dentistry is the control experiment for this whole document — the one service where the public offer has visibly thinned and the private market has simply taken its place. Nobody legislated for that. It happened because a payment contract unchanged since 2006 made NHS work unattractive, and patients went where the appointments were.

The contract is being modified, not replaced. From 1 April 2026 urgent care is paid at £75 ($101) a course instead of 1.2 units of dental activity, with a minimum urgent-care requirement, a quality programme at £3,400 ($4,590) per practice a year, and complex-care pathways from 23 June 2026.

The UDA remains the currency. The 2024 recovery plan's new-patient premium spent £88m ($119m) and was followed by 3% fewer new patients getting NHS care.

Meanwhile the Competition and Markets Authority has turned its attention to the private market, opening a study on 5 March 2026. Its own numbers: 29% of adults in England used private dental care exclusively in 2023/24; an initial private consultation rose more than 23% to £80 ($108) and a routine check-up more than 14% to £55 ($74) between 2022 and 2024.

Its chief executive: "For some, turning to private dentistry is a choice – but for many, it's a necessity."

And the consequence lands on children. In 2024/25, 21,162 children aged 5–9 in England were admitted to hospital with tooth decay — more than for acute tonsillitis (13,667), making decay the leading cause of hospital admission in that age group (Royal College of Surgeons, September 2025).

Around 70 children a day have teeth extracted because of decay. 22.4% of five-year-olds have obvious decay. This is an entirely preventable disease, and access to prevention now depends on where you live and what you can pay.

Hospice and end-of-life care: mostly charity, not NHS

Dying, in the UK, is largely funded by charity shops, raffles and legacies. People not from Britain routinely assume the NHS runs hospices. It does not. Hospices are overwhelmingly independent local charities, and they raise most of their own money every year.

Government money has come as capital, not as a funding formula. £100m ($135m) of capital was announced on 19 December 2024 alongside £26m ($35m) of revenue for children's hospices, paid as £25m (February 2025) and £75m (July 2025); a further £25m in January 2026 took the capital programme to £125m ($169m); £80m ($108m) over three years was committed to children's hospices from 2026/27. A Modern Service Framework for Palliative Care is planned for autumn 2026.

The Health and Care Act 2022 made commissioning palliative care a duty on integrated care boards. The Public Accounts Committee's finding in March 2026 is that the duty "has yet to result in universal access to quality services."

That is the pattern of this entire document in miniature: a duty written down, nothing structural behind it, and 0%–80% variation is what local discretion turns out to mean.

And the unmet need is measurable. Marie Curie's February 2026 analysis puts it at 203,442 people a year in the UK, 31% of those who need palliative care, dying without it — "nearly one person every three minutes".

More than two in five people die in hospital, and 41% of that group die alone. Of £11.7bn ($15.8bn) spent annually on people in their last year of life, £6.3bn ($8.5bn) is hospital care — the most expensive setting, and usually not the one people choose.

And here is the part that should stop anyone in their tracks: the least funded part of the system is the best performing part of it. When the CQC last published a cross-sector comparison — The state of hospice services in England, October 2017 — 25% of hospices were rated Outstanding and 70% Good, against roughly 6% of NHS acute hospitals and about 2% of care and domiciliary services.

The figures are nine years old and CQC has published no equivalent update (its inspection backlog and the 2023 change of framework make one impossible for now), so quote them with the date attached. But nothing since contradicts the direction, and the structural explanation is the one this whole document is about: hospices are independent charities governed by local trustees, staffed alongside 89,000 volunteers who make a closed culture almost impossible, and measured on how a person's last weeks go rather than on bed turnover.

The sector with the clearest identity, the most community visibility and the least money produces the best-rated care in England.

Health and social care are two different systems

This is the distinction most often missed from outside the UK, and it matters.

Northern Ireland is the exception: health and social care are integrated in a single system there.

Social care in its own right

First, a scope point that matters for anyone reading the headline numbers. The £345bn ($466bn) total health spend uses the international System of Health Accounts, which is broader than the NHS and, in the ONS's own words, "include[s] some services considered social care in the UK".

The test it applies is whether care supports basic activities of daily living — bathing, dressing, walking — in which case it counts as health; help with shopping, cooking or managing money sits outside the headline and is reported separately as "social long-term care".

So the honest statement is this: personal and nursing care is partly inside the headline health figure; the wider social care system is not, and there is no national social care budget at all.

Total long-term care, health and social combined, was £80.3bn ($108.4bn) in 2024, growing faster in real terms than health spending as a whole. The NHS's own contribution to social care runs to low single-digit billions — the Better Care Fund and NHS-commissioned care — not tens of billions.

Second, the sector is enormous, and structurally the opposite of the NHS. It is means-tested rather than free, commissioned by 153 local authorities rather than one national body, and delivered overwhelmingly by private and voluntary providers.

Third, the market is under visible strain, and the harm data is substantial. Two-thirds of councils (66%) saw provider closures, ceased trading or contract handbacks in the six months to May 2026; councils overspent their social care budgets by £715m ($965m) in 2025/26, a fourth consecutive year; over 400,000 people are waiting for an assessment, care, a direct payment or a review; and over half of directors of adult social services have only partial or no confidence in meeting their statutory safeguarding duties.

The fee gap is explicit: the average council home care price is £25.05/hr ($33.82) against a sector-calculated minimum sustainable price of £34.42/hr ($46.47).

On harm, England recorded 640,240 safeguarding concerns in 2024-25 and 185,270 formal section 42 enquiries involving 148,830 people; the most common risk was neglect and acts of omission, at 40.8%, and the most common location was the person's own home.

Fourth, and directly relevant to this document's argument: social care has had the exposés without the inquiries. Winterbourne View, a private assessment and treatment hospital, was exposed by undercover filming in 2011; its serious case review called it "a case study in institutional abuse".

Whorlton Hall was exposed the same way in 2019, and the regulator's own independent review found it had been wrong not to publish a critical 2015 inspection. Neither produced a statutory public inquiry.

So the sector with the larger workforce, the more fragmented provision, the lower pay, the higher turnover and 640,000 safeguarding concerns a year has had less formal scrutiny than the NHS, not more — which should temper any assumption that the inquiry record in this document represents the worst of it.

Social care is not the better-run half of the system. On the evidence it is the worse one.

The natural assumption is that social care, being smaller and closer to home, is less prone to the failures in this document. The opposite is true — and the reason it looks calmer is that far less of it is being watched.

The regulator has largely stopped looking. Dr Penny Dash's independent review, published 15 October 2024, found "significant failings in the internal workings of CQC, which have led to a substantial loss of credibility within the health and social care sectors".

CQC's own accounts show published assessments falling from 10,356 in 2022/23 to 3,825 in 2024/25 — a 63% drop.

The consequence, measured by the Homecare Association in June 2026: 83.5% of community social care services in England have no current CQC rating — 36.9% have never been assessed at all, and 46.6% are carrying a rating between four and ten years old.

Billions of pounds of public money buy care from providers the state has not looked at this decade.

And in England there is still no professional register for the people delivering that care. A care worker dismissed for neglect in an English care home can be hired by an agency down the road the following week, because there is no register to be removed from and no fitness-to-practise process to answer to. England is the only UK nation in that position.

Scotland's SSSC had 173,117 registered individuals in July 2025; Social Care Wales registers adult care home, domiciliary and residential child care workers by law (27,498, 24,124 and 5,663 respectively in August 2026); Northern Ireland made registration compulsory in March 2017 and NISCC had 49,253 registrants at 31 March 2025, with 510 fitness-to-practise referrals in 2024-25.

In England, qualified social workers are regulated (Social Work England, 108,345 registered at 31 August 2026) and registered managers register with the CQC rather than a professional body; everyone else is unregulated.

Set that against Volume 9: the gap this document identifies in NHS management regulation exists across an entire 1.6 million-person workforce next door, and no scheme to close it has been announced.

Children's residential care is the clearest case of the state losing control of a market it depends on. Ofsted's Annual Report 2024-25 records that nearly nine in ten local authorities placed children in unregistered homes because they could not find registered places, and that Ofsted opened almost 900 investigations into potential unregistered homes.

In July 2026 Ofsted put it at over 900 unregistered homes identified in 2024, around one in ten children in residential care living in one, at a cost to the taxpayer of £353m ($477m), with some providers charging up to £30,000 ($40,500) a week.

The Children's Commissioner counted 669 children in illegal children's homes at 1 September 2025, at an average of £10,500 ($14,175) a week.

For context on the economics, the Competition and Markets Authority found in 2022 that the fifteen largest children's home providers were running average operating margins of 22.6%.

The money is failing in the ordinary way too. ADASS's autumn survey projected a £623m ($841m) overspend mid-year; the outturn for 2025/26 was £715m ($965m), 3.2% of net budgets, with 80% of councils overspending in both 2024/25 and 2025/26. Adult social care now takes 37.9% of councils' entire net budgets in 2026/27, up from 35.5%.

Councils plan £909m ($1.23bn) of savings this year and only 15% of directors are fully confident of delivering them. Demand is not obliging: the number of 18–24 year olds needing packages costing over £7,000 ($9,450) a week rose 30% in a single year, from 547 to 712.

All of which lands back on the NHS. CQC's State of Care, October 2025: on any given day in March 2025, nearly six in ten patients who were ready to be discharged experienced a delay. The Nuffield Trust counted 13,750 patients a day delayed in acute hospitals in January 2026.

This is the boundary problem in its most expensive form — a hospital bed used because a care package does not exist.

And almost nobody outside the sector knows any of this. Ipsos polling for the Health Foundation in May 2025 found that 31% of adults think the NHS provides most social care for older people and 33% think it is free at the point of need — rising to 52% and 44% among 16–24 year olds.

Both beliefs are wrong, and the consequence is political: there is no public pressure to fix a system most people assume is already the NHS. That is the single best explanation for why a sector with a larger workforce, worse pay, higher turnover, thinner regulation and 640,000 safeguarding concerns a year has had a fraction of the scrutiny.

There are bright spots, and they point the same way as the rest of this document. Shared Lives — an adult who needs support moves in with an approved carer and their family — supported 9,802 people across the UK in 2023-24 with 10,081 carers, and 96% of its English services are rated Good or Outstanding by the CQC, the best quality profile in social care; the Local Government Association puts the saving at £12,000–£26,000 ($16,200–$35,100) per person a year against residential or supported living for people with a learning disability.

Somerset has taken the opposite approach to scale, backing very small local providers: 1,250 new micro-providers developed since 2015, now delivering over 30,000 hours of support a week to almost 6,000 people (the impact evaluation is the delivery partner's own, so treat the savings claims with care).

And Wales has legislated to take profit out of children's care — the Health and Social Care (Wales) Act 2025, with no new for-profit registrations from 1 April 2026 and restrictions on expansion from 1 April 2027.

What these have in common is not a technique. It is small scale, visible relationships and an owner who is accountable to the person rather than to a distant balance sheet — the same structural point Volume 8 makes about Mayo, Nuka and Buurtzorg.

One honest caution on importing models, which belongs here rather than in a footnote. The Buurtzorg self-managing team model, rightly admired, has been tested three times in England — Guy's and St Thomas' (2016-17), East London in Tower Hamlets (2017-18) and West Suffolk (2018) — and the independent evaluations are sobering.

Patient experience and holistic care improved; but the King's Fund's review of West Suffolk concluded it "cannot say from this review whether the Buurtzorg model 'works' in the English context", with the team shrinking from six staff to under two full-time equivalents through resignations, and no analysis of admissions or cost completed at all.

The lesson is the one this document keeps arriving at: a structure imported without the identity and the support behind it becomes another initiative, and fails.

The United States, for comparison

For a US reader the shape is familiar but the scale and the payer mix are different, and the workforce numbers are the sharpest contrast.

Three things a US reader should take from that. The US sector is roughly four to five times the UK's by spend and three times by workforce, but its turnover is four times worse, and its staffing floor has just been removed rather than raised.

The UK's problem is not a different problem: it is the same problem — low pay, high churn, fragmented private provision, means-tested access, and neglect as the single largest category of recorded harm.

And in both countries the failures that get investigated are overwhelmingly the healthcare ones, not the social care ones, even though social care employs more people and touches more lives at their most vulnerable.

Whatever works for humanity under pressure in a hospital has a larger and less scrutinised market waiting for it in care.

Care in the community — and paying the family to do it

Most care in England does not happen in a hospital or a care home. It happens in people's own homes, delivered by private companies and by relatives — and the relatives are the larger workforce by far. This is the single biggest thing a non-UK reader misses about the system.

The policy origin is the NHS and Community Care Act 1990 (community provisions live from 1993), which moved responsibility to local authorities and separated commissioning from provision. Long-stay institutions closed and care moved into homes, supported living and reablement.

Four decades on the effect is in the numbers: of the 889,000 people who received long-term support in England in 2024-25 (672,000 at 31 March 2025), 67% of those with a known setting were living at home or with family, rising to 86% of those aged 18–64.

Of £23.5bn ($31.7bn) of long-term care spend, community care took £11.6bn ($15.7bn), 49%, against £7.9bn ($10.7bn) residential and £3.0bn ($4.1bn) nursing.

It is a private market. England had 15,232 CQC-regulated non-residential locations at March 2025, up 72% since 2018, with 740,000 filled posts, a 9.9% vacancy rate and 23.7% turnover. In the week to 16 February 2026, 505,886 people were receiving regulated home care — counted across the 84.9% of providers who responded, so the true figure is higher.

The state does pay family members, in two limited ways. The first is a direct payment: cash instead of commissioned services, taken by 24.5% of people on long-term support in 2024-25, with around 70,000 recipients employing their own staff and ~131,000 posts working directly for them.

The rule on relatives is narrower than usually assumed: a direct payment may not normally be used to pay a spouse or partner, or a close relative living in the same household, unless the local authority agrees it is necessary (Care and Support (Direct Payments) Regulations 2014, regulation 3). A relative living elsewhere is not caught by the restriction.

The second is Carer's Allowance, and its arithmetic is the point. It pays £86.45 ($116.71) a week in 2026/27 for providing at least 35 hours of care a week — £2.47 ($3.33) an hour, against a National Living Wage of £12.75 ($17.21). Earnings above £204 ($275) a week disqualify the claim entirely: a cliff edge, not a taper.

At March 2026, 1.4 million people were claiming, but 26.7% were entitled and not paid because of overlapping benefits; around 902,000 were actually paid in 2023-24, at a cost of £3.7bn ($5.0bn), forecast to reach £4.5bn ($6.1bn) in 2025-26.

That cliff edge produced a scandal of its own, and it is instructive. Outstanding Carer's Allowance overpayment debt reached £250m ($338m) in 2023/24, up from £150m in 2018/19, held by tens of thousands of carers who had earned slightly too much in a week.

The independent review by Liz Sayce (report dated 21 July 2025, published 25 November 2025) found the cause was "conflicting policy drivers" and an absence of "integrated, concerted leadership" rather than dishonesty — and, critically, that the department's own guidance on fluctuating earnings did not accurately reflect the statutory position between 2015 and summer 2025, so "some carers could not have known that they were building up overpayments". Carer's Allowance claimants received 46% of all DWP civil penalties in the year to March 2025.

Government accepted or partly accepted 38 of the 40 recommendations; from 13 April 2026 it began reviewing over 200,000 cases, with around 25,000 carers expected to have debt reduced, cancelled or refunded, and £75m ($101m) allocated across three years.

Same mechanism as everything else in this document: people raised it for years, the system recorded it as their error, and it took an independent review to reframe it as the institution's.

And the scale of what unpaid carers do dwarfs everything on this page. At Census 2021 there were 5.0 million unpaid carers in England and Wales (around 4.7 million in England), of whom 1.7 million across the UK provide 50 or more hours a week.

The most recent valuation of that work is £184.3bn ($248.8bn) a year (2021/22) — which the Commons Health and Social Care Committee noted is "equivalent to a second NHS".

Set against it, local authorities spent £203m ($274m) supporting carers in 2024-25. That is the ratio to remember: the state spends roughly one pound supporting unpaid carers for every nine hundred pounds of care they provide.

Carers UK's State of Caring 2025 (10,539 respondents, self-selecting, so treat percentages as direction not prevalence): 35% report bad or very bad mental health, 42% say their physical health has suffered because of caring, 49% have cut back on essentials such as food and heating, 61% provide 50+ hours a week, and 73% had not had a carer's assessment in the previous twelve months.

One claim to make carefully. It is often said that supporting family carers is cheaper than the alternative. No authoritative body states that in those terms, so the honest version uses what is costed: delayed discharges cost £1.89bn ($2.55bn) in 2022/23, and the Institute for Fiscal Studies found that reductions in social care spending for over-65s were associated with a 3.8% rise in the probability of an A&E attendance.

The Health and Social Care Committee adds its own caution, which is worth keeping: social care "should not be valued only for how it supports the NHS".

Mental health, and beds

• Mental health, learning disability and dementia: £20.6bn ($27.8bn) planned for 2025/26 — 14.2% of total integrated care board funding, up 31% in real terms since 2016/17. 4.13 million people were in contact with secondary mental health, learning disability and autism services in 2024/25, just over 7% of the population and the highest since the series began.
• 129,905 overnight hospital beds in England (quarter to December 2025): 103,824 general and acute, ~18,000 mental health, ~8,000 maternity. General and acute occupancy 91.5%, against the 85% usually treated as the safe ceiling.
• About 2.4 beds per 1,000 people, against an OECD-EU average of 4.6 and Germany's 7.8 (BMA analysis). On scanners, the most recent traceable comparison is now dated: 8.8 CT and 7.4 MRI per million people against OECD averages of 25.9 and 16.9 (2021 data, submitted to Parliament in 2022) — the direction is not in doubt, the precision is.

What is changing — and the argument now running at the top of government

The direction of travel is out of hospitals and into communities, and for the first time the person setting that direction has been arguing for it for sixteen years. Three things a reader should know.

One: mental health support is being moved into neighbourhoods, at scale. On 5 August 2026 the government announced £343m ($463m) for 100 new community mental health centres (£187m / $252m), offering walk-in support without a referral, and 59 new mental health emergency departments (£156m / $211m).

The base is small — 6 community centres and 23 emergency departments exist today — so this is a roughly six-fold expansion, with the first sites opening in autumn 2026 and more from March 2027.

Read it precisely: this is capital delivering a commitment already in the 10 Year Health Plan, and it is larger than that plan's "up to £120 million" for mental health emergency departments, but it is not a new funding settlement for mental health.

The Commons Health and Social Care Committee reported in December 2025 that mental health took 8.71% of NHS spending (£15.6bn / $21.1bn in 2025/26) against more than 20% of demand — the first proportional decline since 2016/17.

The Royal College of Psychiatrists' president welcomed the announcement as a "recommitment", which is the accurate word.

Two: the 10 Year Health Plan (3 July 2025) commits to three shifts — hospital to community, analogue to digital, sickness to prevention — with a neighbourhood health centre in every community, starting where healthy life expectancy is lowest.

It sets no target for reducing hospital spending, which is the standard criticism: the shift is stated as an intention and left to be delivered by organisations that are already full.

Three: social care reform is live again, and the politics of it are unusually personal. As Secretary of State for Health, Andy Burnham published Building the National Care Service on 30 March 2010 — free personal care at home for those with the highest needs, national standards, and a comprehensive National Care Service — and its funding proposal was destroyed in the "death tax" row before the election. He is now Prime Minister.

On 29 July 2026 he committed to a National Care Service on four principles (seamless movement between hospital, home and community care; full and sustainable funding; prevention and independent living; citizens' needs at the centre), brought Baroness Casey's commission on adult social care forward from 2028 to summer 2027, launched a public consultation — "The Big Conversation on Care" — and asked for cross-party talks, noting 22 failed attempts at reform in three decades.

No money was attached. In August the government said it would explore broadening the Fair Pay Agreement so that care pay and progression align with NHS standards.

And what has not happened is as important. The Dilnot cap — an £86,000 ($116,100) lifetime limit on personal care costs, with the capital threshold rising to £100,000 ($135,000) — was cancelled on 29 July 2024 and has not returned.

The Fair Pay Agreement's own timetable is regulations in October 2026, negotiations in 2027 and implementation in 2028. The Casey commission's first-phase report had not been published as at September 2026.

The structural fault all of this circles is simple, and it is worth stating plainly for a non-UK reader. The NHS is free at the point of use; social care is means-tested above £23,250 ($31,388). So the boundary between them decides who pays — and when social care thins, the cost does not disappear, it arrives in a hospital. Delayed discharges alone cost £1.89bn ($2.55bn) in 2022/23.

That is the case for shifting care into the community, and it is a strong one. But note what it does not touch: moving a service from a ward to a high street changes where care happens, not how people are treated when they get there. Every failure in this document happened inside a building someone had already reorganised.

Who regulates what

• Care Quality Commission (CQC) — independent regulator of quality and safety for all health and adult social care providers in England: registers, inspects and rates.
• NHS England — national policy, funding allocation, and performance and financial accountability for ICBs and trusts (merger into DHSC pending).
• GMC, NMC, HCPC — UK-wide professional regulators for doctors; nurses, midwives and nursing associates; and 15 allied health and psychological professions. They set standards, hold the registers and run fitness-to-practise proceedings, independently of employers. Until 2026 there was no equivalent for non-clinical managers — the gap Volume 9 is about.
• Coroners and inquests (England and Wales) — a coroner must investigate any death that is violent, unnatural, of unknown cause or in state detention. An inquest establishes who, when, where and how, not liability, and a coroner may issue a Prevention of Future Deaths report that the recipient must answer.

Sources for this annex: ONS population and UK Health Accounts; House of Commons Library briefings on NHS expenditure, NHS key statistics, mental health statistics and adult social care; NHS England Digital workforce, GP and activity statistics; NMC and GMC register data; King's Fund; Skills for Care; LaingBuisson; IHPN; BMA; Nuffield Trust.

On dentistry: NHS Business Services Authority, the Public Accounts Committee and the British Dental Journal. On hospices: the National Audit Office, the Public Accounts Committee, Hospice UK, Together for Short Lives and Marie Curie.

On community care and carers: DHSC adult social care activity and finance reports, the Care and Support (Direct Payments) Regulations 2014, DWP benefit rates and statistics, the Sayce review, ONS Census 2021, the Centre for Care and Carers UK.

On what is changing: the DHSC announcement of 5 August 2026, the 10 Year Health Plan, the Casey Commission and the Prime Minister's statement of 29 July 2026. Figures are the most recent published as at 20 September 2026, and the accuracy notes for all of them are in the Sources volume.

The impact: when people stop caring

What it actually costs when humanity fails under pressure — in people, in organisations and in money. Three columns of the same ledger, and they are the same event counted three ways. Sterling figures carry a US dollar equivalent at $1.35 = £1.

The one thing to take away: inhumanity is not a soft cost. The people harmed, the staff who leave and the money paid out in negligence are one failure counted three ways, and preventing it is far cheaper than paying for it.

The short version

The point of putting them side by side. Each column is usually argued separately — the human one by families, the organisational one by safety researchers, the financial one by auditors. They are the same failure.

A ward where people have stopped being able to treat each other decently produces the harm, loses the information that would have caught it, and generates the bill. Fix the first and the other two move with it.

1 · The human impact

What happens to patients

The harm is not abstract and it is not rare, and in the worst cases it is deliberate. At the Countess of Chester, fourteen babies were murdered or attacked between June 2015 and June 2016 — the unit recorded eight deaths from 468 admissions in 2015 against one to three a year before.

Survivors carry it for life: Thirlwall records one baby left with "profound brain injury… with lifelong consequences", requiring 24-hour care, eleven years on.

Most harm, though, is the ordinary kind, and there is far more of it. At County Durham, 315 of 514 patients reviewed were harmed (61.3%), 77 of them significantly, including 20 unnecessary mastectomies — breasts removed that did not need to be removed — in a service that lost its GMC trainees in 2012 and was still being warned in 2025.

Ockenden's Nottingham review covers more than 2,500 families and care from 2012 to 2025. At Gosport, 456 patients' lives were shortened. Infected blood ran from the late 1970s, with four decades of institutional denial after it.

And it does not stop at the ward door. 203,442 people a year in the UK die with palliative care needs that were never met — nearly one person every three minutes. More than two in five people die in hospital, and 41% of those die alone.

In adult social care, England recorded 640,240 safeguarding concerns in 2024-25, with neglect and acts of omission the largest category at 40.8%, most often in the person's own home. At Nottingham, a category 4 pressure ulcer — the most severe grade there is — was linked to a frail patient's prolonged wait on a corridor trolley.

What happens to families

They find out last, and often from someone else. At Chester, families learned the truth from the police in 2018 or at the trial in 2023. One mother: "to not inform us at all until somebody is arrested is unforgivable."

At Durham, the duty of candour was owed from May 2024 and began in February 2025 — nine months in which people went on not knowing that their own treatment had been reviewed and found wanting.

And then they have to do the institution's job for it. In every case in this document the truth emerged because someone bereaved would not stop asking. Families were the warning bell, the investigation and the pressure for an inquiry — which is precisely what the designed mechanisms were supposed to be.

The cost of that is measured in years: at Chester, families were still giving evidence about the deaths of their children ten and eleven years afterwards.

What happens to staff

The people who raise concerns pay, and the people who ignore them do not. At Chester the consultants who raised the alarm faced a documented plan — "1. GMC 2. Actions from grievance 3. Mitigation from SoS/whistleblowing 4. Action plan to manage out". One consultant told the inquiry he had spent most of the last ten years reliving the events.

Where whistleblowers win, the bill lands on the public: £1.4m ($1.89m) awarded against the Countess of Chester in the Gilby case, £700,000 ($945,000) spent by one trust on the Dr Chris Day litigation.

Where they lose, they leave — and, as Kark KC told Thirlwall, the executive "moves on to the next Trust down the road" with a "vanilla reference".

The wider workforce shows the damage in its own numbers. NHS Staff Survey 2025 (766,285 respondents): 31.5% burnt out, 56.0% worked while unwell, 29.6% often think about leaving and 16.8% would leave as soon as they could — an all-time high.

Sickness absence stood at 5.0% in May 2026, with mental illness the single largest cause at 29.9% and over 680,000 full-time-equivalent days lost in that month alone. 30,323 nurses left the NMC register in the year to March 2026.

In adult social care, turnover runs at 23.1% — about 335,000 people leaving a year. Among unpaid carers, 35% report bad or very bad mental health.

And the harm travels home. The Champions pilot in Chicago recorded staff describing work before the programme as "being alone on a raft in the middle of the ocean", and afterwards using the same tools with their own families. People do not leave this at the door in either direction.

2 · The organisational impact

Start with what a hospital's safety system actually is. It is not the policies. It is people noticing things and telling each other — a nurse saying the baby looks wrong, a junior asking the consultant to look again, a colleague offering help before being asked.

Every one of those is a social act, and every one of them degrades when people stop treating each other decently.

That degradation is measured, experimentally, not asserted. Incivility costs its target about a third of their performance (33% worse on a task, 39% fewer creative ideas) and bystanders who merely witnessed it about a quarter (22% and 28%).

It makes colleagues 74–84% less likely to help someone. In a simulated operative crisis, 91% of clinicians passed in normal conditions and 64% when exposed to rudeness.

And the mechanism is established: rudeness impairs working memory, and the impaired working memory degrades the clinical work. 27% of US practitioners knew of a medication error in the previous year in which disrespectful behaviour played a part; 49.8% of Australian hospital staff said unprofessional behaviour had a moderate or major negative effect on patient care.

Then the failure follows a script, and it is the same script every time. A concern is raised. It is reframed as a conduct problem about the person raising it. Candour is delayed while the organisation establishes its position.

The regulator is shown a curated picture. The person who spoke is managed out. Chester ran all five steps and wrote the fifth one down. Durham delayed candour for nine months and tried to identify which staff had spoken to the review.

The threshold moves while this happens, which is why nobody inside notices. Each unchallenged act licenses the next, until behaviour that would once have been unthinkable is unremarkable and the person who finally objects is the one who looks unreasonable. Silence is not neutral: it is information, and what it tells the ward is this is tolerated here.

The Edenfield Centre is the clearest recent anatomy of it. After the 2022 undercover investigation into that NHS medium-secure mental health unit, Professor Oliver Shanley's independent review (January 2024) found unsafe staffing, a closed culture with no psychological safety, incivility between staff and the suppression of bad news, and fear-based senior leadership prioritising targets and expansion over care quality — with board members invisible in the services and governance that failed to escalate warning signs already visible in staff survey results and vacancy rates. The warning signs were in the data. Nobody acted on them.

The institution then pays in ordinary operational currency. Nottingham University Hospitals answers 12.6% of complaints on time against an 85% target, with investigations averaging 55 days. Vacancies run at 97,475 (6.5%) across NHS hospital and community services and 96,000 (6.2%) in adult social care.

Agency and bank cover, presenteeism and absence together cost the equivalent of a fifth of NHS staff spending.

And the organisation loses the people most likely to have spoken up next time — which is the compounding cost, because a culture that removes its challengers becomes progressively less able to detect its own failures.

And there is a slower loss that no ledger records: the withdrawal of trust. When families conclude that the only way to the truth is a campaign, a lawyer or a journalist, the system has lost its cheapest and fastest source of safety information. Every inquiry in this document was triggered by someone who had already given up on being told.

3 · The financial impact

The full detail, with sources and corrections, is in the organisational failure annex. This is the summary a board needs.

Three things about that table are worth saying out loud.

First, none of it buys any patient any care. Every figure in it is the cost of harm, of contesting harm, or of finding out about harm. It is the single largest avoidable line in British healthcare and it produces nothing.

Second, the cheapest-sounding failures are the most expensive to resolve. A large volume of individually modest claims — an unnecessary excision, a delayed diagnosis — is the worst possible mix, because each one carries a full legal process regardless of its value, at 3.7 times the damages.

That is exactly the profile a service-wide look-back produces. Scaled across the 10,000 records that County Durham may bring into scope, the same rates imply roughly 6,100 patients harmed, a claims bill of £1.4bn–£1.7bn ($1.9bn–$2.3bn), and — on the coal precedent for administration — a total approaching £2bn ($2.7bn), from one service in one trust over one decade.

Third, the longer an organisation denies, the more of the eventual bill goes to lawyers rather than to the people it harmed. Coal spent more than £1 on process for every £2 that reached a miner. Horizon's claimants received about a sixth of their own 2019 settlement.

In NHS low-value claims, legal costs run at nearly four times the damages. Candour is not only the decent option; it is the cheap one.

And what it costs to do the opposite

Set the ledger above against what the alternative actually costs, because the asymmetry is the argument. The interventions in this document are training, facilitated conversation, and changes to how boards measure and appoint.

Schwartz Rounds run in hundreds of organisations for the price of a room and a facilitator. The Champions pilot at Union Health Service covered around 250 staff across six sites and showed change within 90 days.

Intermountain's obstetric work cost it about $9m a year in forgone revenue — against a US health system's balance sheet, and it did it deliberately.

Be honest about the evidence. The intervention literature is indicative rather than conclusive: multi-component, leadership-led, sustained programmes show effect; one-off training does not. What is not in doubt is the size of the thing being avoided.

Against £12.1bn a year of staff ill-health and £3.2bn a year of negligence payments, an intervention does not have to work spectacularly to be worth funding — it has to work a little, and keep working.

Which is the whole case, stated as arithmetic. Humanity under pressure is not a soft extra to be funded when finances allow. It is among the cheapest sources of clinical performance available to a system that is currently missing every constitutional standard it has — and every year it is not funded, the three columns of this annex are paid in full.

The founding vision of 1948

Part one tells why the NHS was founded — a son holding his dying father, a town that looked after its own, a nation that chose to lift the fear of illness from everyone at once — and measures the chasm between that promise and today.

Part two is the record. The one thing to take away: the NHS was founded on a moral promise about how people should be treated; the money was written into law, the humanity never was, and the Constitution that restates it still binds almost nothing.

Part one · What it was for

It began with a son holding his dying father. In February 1925 — just over a century ago — in Tredegar, David Bevan died of pneumoconiosis — the miners' lung disease, the slow suffocation that comes from a working life breathing coal dust.

He died, as his son later recalled, in the arms of Aneurin, then twenty-seven; and even at the end, Aneurin said, his father was asking for the Daily Herald and for news of John Wheatley, the Labour minister who was trying to build decent homes for working families.

Then came the insult on top of the grief. Pneumoconiosis was not classed as an industrial disease under the Workmen's Compensation Act, so the job that had killed him owed his family nothing, and paid nothing.

A man had given his lungs to the pits, and the system he had given them to treated his life as disposable.

That is where the NHS really starts: moral first, organisational second. Bevan grew up in a mining town that had already worked out how to look after its own. In Tredegar the miners paid a few pence a week into a medical aid society so that when someone fell ill a doctor would come, whoever they were.

So he had seen both things at close quarters — a society that would let a man suffocate uncompensated, and a community that refused to let its neighbours face illness alone. The first gave him his anger; the second gave him his answer; together they gave him the steel to carry it through a hostile profession and a reluctant Treasury twenty years later.

When he rose in the Commons in April 1946 to move the Bill, he did not talk about administration. He talked about what it is like to be ill and afraid of the cost — "the financial anxiety caused by having to pay doctors' bills" — and about the people who therefore "tend to postpone consultation as long as possible".

Before the fifth of July 1948, illness came with a bill, and the bill came first: a family with a sick child did the sum before they did anything else.

And then, on a Monday morning, that fear was lifted from everyone at once. Every household had been sent a leaflet: "It will provide you with all medical, dental and nursing care. Everyone — rich or poor, man, woman, or child — can use it or any part of it … it will relieve your money worries in times of illness."

A radiographer who was there, Ethel Armstrong, remembered it simply: "It was just unbelievable. You couldn't believe that whatever it was that you needed, that required medical attention, was going to be given to you totally free of charge."

At Park Hospital in Davyhulme the first patient, thirteen-year-old Sylvia Diggory, remembered Bevan asking whether she understood the significance of the day, and telling her it was "the most civilised step any country had ever taken".

In its first year the new service supplied some four and a half million pairs of spectacles, with three to four million more on order. People who had lived for years unable to see properly simply came.

What it promised was not a funding arrangement. It was a way of treating people. Bevan wrote that "no society can legitimately call itself civilized if a sick person is denied medical aid because of lack of means" and, more plainly, that "preventable pain is a blot on any society".

Filmed later, he put it in a sentence anyone could carry: "It's a piece of real socialism. It's a piece of real Christianity too, you know. We had to wait a long time for it." In the same film he reminded his audience that it had been built immediately after the war, when Britain was, in Churchill's words, "a bankrupt nation" — a broken country choosing to do it anyway.

Take away the politics and what is left is the proposition this whole document stands on: that the person in front of you is owed care because they are a person, and for no other reason.

Bevan the man: a champion's story

He was a champion of humanity before he was anything else, and millions of lives are better because of who he was. Aneurin Bevan was born in Tredegar on 15 November 1897, the sixth of ten children of a miner, David, and his wife Phoebe; not all of them survived childhood.

He left school at thirteen and went down the pit the following year. None of that was unusual in a Welsh mining town. What he did with it was.

The power of one: a boy with a stammer who became the great parliamentary speaker of his age. As a child he had a severe stammer. He did not wait for it to pass. He built a huge vocabulary so that he could swap a hard word for an easier one, recited poetry, forced himself to speak in public, and later took elocution lessons.

He read his way through the Tredegar Workmen's Library and ended up on the committee that chose its books.

His father was a Baptist who sang in the chapel choir and his mother a Methodist, and he grew up in a chapel culture that taught that everyone was equal before God. He did not keep the faith, but he kept the equality.

(Labour's general secretary Morgan Phillips said in 1950 that British socialism owed little to Marx and much to religious thought — usually remembered, more neatly than he put it, as "more to Methodism than to Marx".)

In 1919 the miners' union sent him to the Central Labour College in London. He came home and was elected to the local council in 1922, the county council in 1928 and, in 1929, to Parliament for Ebbw Vale, which he represented until he died.

In between, in 1925, he held his father as he died. One person, making one choice after another about who he would be: that is where it starts.

The power of the few: Tredegar showed him it could be done. Bevan did not invent the NHS out of nothing; he scaled up what his neighbours had already built. The Tredegar Medical Aid Society, formed in 1890, took about threepence in every pound of the local ironworks and colliery workers' wages and in return provided doctors, a hospital and medicines.

By the mid-1920s it covered around 20,000 people. Bevan sat on its hospital committee from 1924 and chaired it in 1929–30. He was also a leading member of the Query Club, a small group of young Tredegar men who met to argue, help one another and plan local politics.

When the time came he is reported to have said that he was only extending to the whole of Britain what Tredegar had had for a generation: "We are going to 'Tredegarise' you." (The line is widely printed, including in the British Journal of General Practice, but no record of when he said it has been found.) A town of working families proved the idea; a handful of determined people carried it.

It is a remarkable fact, and easily lost: the idea at the heart of the NHS is not seventy-eight years old but more than 130, and its roots go back nearly 300 years, to the Wales of the 1730s (Annex 10 · Why we all love the Welsh). It was one of many workmen's medical aid societies in the South Wales coalfield, but it was the one Bevan knew from the inside — and a community that started out looking after its own went on to shape how a whole nation would be cared for.

The power of two: Nye and Jennie. In 1934 he married Jennie Lee, a Fife miner's daughter who had been elected to Parliament for North Lanark at twenty-four. They were partners in the fight as much as in marriage.

She went back to the Commons in 1945, became the country's first Minister for the Arts, and drove the creation of the Open University — another institution built on the belief that opportunity should not depend on means.

At the end she carried the heaviest load. When surgeons at the Royal Free found cancer, the doctor told Jennie, and she chose not to tell Nye or their friends how serious it was, so that he could spend his last months at their farm making plans.

She later wrote about it in My Life with Nye (1980). The decision is still debated. It was also an act of love.

He was a fighter, and he made powerful enemies. Churchill told the Commons in December 1945 that Bevan would be "as great a curse to this country in time of peace, as he was a squalid nuisance in time of war."

Two and a half years later Bevan had the NHS open anyway, over the opposition of the doctors' own association. When he died at Asheridge Farm on 6 July 1960, aged sixty-two, a memorial service was held for him at Westminster Abbey, and in 1972 Michael Foot unveiled memorial stones to him on the open hillside at Waun-y-Pound, between Tredegar and Ebbw Vale.

A champion's story — and the NHS was founded on the Champions Pathway. This is the author's lens, not a claim anyone made in 1948, but the pieces fit. The Champions Pathway runs from Choice to Connection to Creation.

Choice is the power of one: a person deciding who they will be under pressure — a stammering boy who chose to speak, a grieving son who chose to act. Connection is the power of two and of small groups: Nye and Jennie; the Query Club; the few committed people who ran a town's medical aid society. Creation is the ripple: what a small, committed group builds becomes the way everyone lives.

Champions draws here on the political scientist Erica Chenoweth's finding that no government in her study withstood a peak mobilisation of 3.5 per cent of its population — a finding she herself describes as a pattern in the evidence, not a law, in which organisation and momentum matter as much as numbers.

Tredegar is an older, smaller version of the same truth: a few thousand families showed a nation what was possible, and one man carried it to some fifty million people. The ripple went further than he lived to see — into the people who came from across the world to staff it.

Ripples of humanity: the people who carried it. The service opened some 54,000 nurses short. A fortnight earlier the Empire Windrush had arrived at Tilbury, and from 1949 the Ministries of Health and Labour were recruiting hospital staff directly from the Caribbean; by 1955 recruitment programmes were running in sixteen colonies and former colonies, and by the end of 1965 between three and five thousand Jamaican nurses were working in British hospitals.

By 1960 somewhere between 30 and 40 per cent of the NHS's junior doctors had come from India, Pakistan, Bangladesh and Sri Lanka.

They were not always welcomed. Many overseas nurses were steered — some, historians say, duped — into the lower State Enrolled qualification rather than State Registered training; patients refused their care; and decades later some of that generation were treated by the Home Office as though they had no right to be here at all.

And still they carried the ethos, often more faithfully than the society that had asked them to come. Daphne Steele arrived from British Guiana in 1951 and in 1964 became matron of St Winifred's in Ilkley — the first Black matron in the NHS — and received more than 350 letters from around the world: "It did something for race relations and it also did something for other blacks because if one can do it we all can do it."

Allyson Williams came from Trinidad in 1969, trained at the Whittington, and retired as a deputy head of midwifery. Yvonne English, recruited from Jamaica in 1965 to train at Harrogate, said only that "we need to leave something behind to show that we made our contribution". They did.

Today around one in five NHS staff in England report a nationality other than British. And on 8 December 2020 the first person in the world to be given a Covid-19 vaccine outside a clinical trial, ninety-year-old Margaret Keenan, received it in Coventry from May Parsons, a matron who had come to the NHS from the Philippines.

"The nurse in me kicked in," she said, "where the patient is first." That is the ripple, seventy-two years on from the Windrush: humanity carried into the service by people who were so often offered less of it than they gave.

Seventy-eight years on: the chasm

Set the promises beside the present, and the distance is hard to look at. The leaflet promised relief from worry; today 7.33 million pathways are waiting for planned treatment, three in four A&E patients are dealt with inside the four hours in which the standard says nineteen in twenty should be, and two-thirds of the nurses who answered their college's survey say they have given care in corridors and other unsuitable places — nine in ten of them that patients' safety was compromised.

Bevan said nobody should be deterred from seeking help early; today four adults in ten have seen an NHS dentist in the last two years, and tooth decay is the leading reason children aged five to nine are admitted to hospital.

The Constitution says everyone counts; Lord Mann found racism "normalised in the NHS". The founders took away the fear of a bill; the families in this document have had to fight for years simply to be told the truth, at Chester, at Durham and in Nottingham.

And the people who work in the service are saying something too: fewer than two-thirds would now be happy with the standard of care if a friend or relative needed it — the lowest figure ever recorded — and fewer than half feel their organisation values their work.

At Chester the chasm reached its lowest point: a system that could not protect newborn babies, and then protected itself. Between June 2015 and June 2016 a nurse on the neonatal unit at the Countess of Chester Hospital murdered seven babies and tried to murder seven more — fourteen babies, fifteen offences.

(The Thirlwall Inquiry proceeded on those convictions; an application to the Criminal Cases Review Commission is pending.) Some of the babies who survived live with lifelong injuries.

The Inquiry, reporting on 15 September 2026, ten years on, found that the hospital's executives focused "on the effect of a situation upon reputation rather than the effect of the situation on patient safety", that they "ignored safeguarding, patient safety and the need to protect whistleblowers", and it recorded the concession, made by almost everyone who gave evidence, that "there was a total failure of safeguarding at every level".

One mother put it more simply: "I feel that the Countess of Chester Hospital have covered up what happened to our daughter for years."

A health service founded so that no family would face illness alone had allowed the most defenceless patients it has to be harmed, and then treated the families' questions as the threat. You would hope that this is the bottom.

The research on atrocity says it is not the bottom — unless something changes at the root. Ervin Staub, who has spent fifty years studying how ordinary people and institutions come to do terrible things, found that harm develops step by step: "Once perpetrators begin to harm people, the resulting psychological changes make greater harm-doing probable."

What stops the slide is an early response from the people around it. The same logic applies to institutions: an organisation that meets harm with silence teaches itself that silence works. (This is an analogy about direction of travel, not a comparison between a hospital and a genocide.)

Chester shows it happening. Six years after the babies died, with the hospital under national scrutiny, four of its senior leaders ran a secret campaign to remove the chief executive who had raised a whistleblowing complaint — Project Countess, upheld in full by an employment tribunal in 2025. The organisation had not learned; it had repeated.

The national record set out in Volume 2 — eight major inquiries in twenty-nine months, the same failures found each time — is what a whole system on that path looks like. Thirlwall's own warning is plain: "It is inexcusable to ignore such failings.

Even where recommendations are accepted, there is no mechanism to track progress or enforce implementation." Without root-and-branch change, the evidence points one way, and it is not up.

And the tragedy that started it all is being repeated, on a far larger scale. Not pneumoconiosis, and not usually a bill — but the same experience David Bevan's family had in 1925: of a system that treats people as though their lives were disposable, and of families left alone to count the cost.

The parents at Chester who were not told. The women at Durham who were not recalled. The children placed in illegal homes, the adults with learning disabilities abused behind locked doors, the patients in corridors.

The NHS was created so that no family would again have to watch someone they loved treated as though they did not matter. Seventy-eight years later, too many are watching exactly that.

And the most telling fact of all is how late the values were written down. The NHS ran for sixty years before anyone set its values on paper. The NHS Constitution was published on 21 January 2009, sixty-one years after the launch, with six values: working together for patients; respect and dignity; commitment to quality of care; compassion; improving lives; everyone counts. They are the right values.

But the Constitution says of itself that it "does not create legal rights"; its pledges are "not legally binding"; the values carry no mechanism of any kind; and the review the law requires every ten years was opened in 2024 and closed in March 2025 without a new text. The values were written down and never wired in. That is the chasm, in one sentence.

Yet hope is not lost, because a great deal has been achieved. This is a service that turned a nation's hidden, untreated suffering into ordinary, routine care. Between 1980 and 2021 infant mortality in England and Wales fell from 12 to 3.7 deaths per 1,000 live births.

British clinicians delivered the world's first IVF baby in 1978, and the NHS gave the world's first Covid-19 vaccination outside a trial. It is still kept going, every day, by some 1.5 million people in England alone who go to work to look after strangers.

What has been lost is not the capability. It is the thing the capability was for — the ethos, the reason the service was created — and without it the NHS is on a steep and dangerous slide.

Bevan did not live to see much of it. He was taken into the Royal Free Hospital on 27 December 1959 for what was expected to be ulcer surgery; the surgeons found cancer. He went home to his farm in February and died there on 6 July 1960 — the day after the service he built turned twelve.

The line most often put in his mouth, that the NHS "will last as long as there are folk left with the faith to fight for it", appears in nothing he is recorded as saying or writing; it belongs to the people who loved the service, not to him, and this document does not attribute it to him.

But whoever first said it was right. The founding ethos was never self-sustaining. It has always needed people prepared to fight for it — and, this document argues, it now needs something Bevan never gave it: a structure that does the holding on the nights when nobody has the strength left to fight.

Lighthouses, and a north star

The way back is not unknown. It has been charted, in places, already. This document is full of them: children's services in Manchester and Leeds taken from Inadequate to Outstanding; Camden's adult social care, the first in the country rated Outstanding under the new inspection regime; Hertfordshire's family safeguarding model, with 30% fewer children in care; hospices rated Outstanding four times as often as acute hospitals; the Esther network, Martha's Rule, and Intermountain choosing to lose money rather than give worse care.

They are lighthouses in the dark — not proof that the night is over, but proof that there is a safe passage through it, and that someone has already found it. The two annexes that follow set out the nearest of them: in social and children's care, and in what it takes to keep doctors and nurses believing in the place they work.

And the Constitution, for all its failings, is the north star. It came late, it was never embedded, and in law it binds almost nothing. But its words are right, they are already the NHS's own, and they point exactly where this document argues the service must go.

Humanity-first healthcare does not ask the NHS to find new values. It asks it to steer by the ones it wrote down in 2009 — and to build, at last, what Bevan built for the money and never built for the people: the structure that holds those values in place on a bad night.

Part two · The record

What follows is the evidence behind Part one: what the founders said and did, in their own words, and what the NHS Constitution does and does not bind. It is set out with sources because the argument deserves to survive a hostile reader.

Why the record matters

The whole argument of this document is that the NHS was founded on a moral proposition and never built the machinery to hold itself to it. That claim only carries weight if the founding proposition is set out accurately, in the founders' own words, and if the present constitutional position is described exactly.

This annex does both: the legislation, the launch, what Bevan said he was doing, what was conceded to get the service started, and what the NHS Constitution says, binds and does not bind. Every quotation has been traced to a primary or authoritative source.

The final section lists the famous lines that could not be, and which this document therefore does not use.

One note on money: historic sterling amounts here are as published in their own year. The document's $1.35 convention is for present-day figures and is not applied to 1940s sums.

What the NHS was built against

Before 1948 the question was not how good your care was, but whether you could pay for it. The Beveridge report, Social Insurance and Allied Services (Cmd. 6404), was presented to Parliament in November 1942 and published that December.

It made "the establishment of comprehensive health and rehabilitation services" one of the three assumptions on which the whole plan for social insurance depended — health treated as a precondition of the welfare state, not one of its programmes.

The wartime coalition's own answer stopped short of the one that was built. The White Paper A National Health Service (Cmd. 6502, 17 February 1944) proposed a comprehensive service to "provide for everyone all the medical advice, treatment and care they may require", at an estimated £148m a year from central funds, rates and contributions.

But it did not propose taking the hospitals into public ownership: "the problem of creating a National Health Service is not that of destroying services that are obsolete and bad and starting afresh, but of building on foundations laid by much hard work."

The difference between that paper and the 1946 Act is the difference between improving a patchwork and replacing it.

The Act, and the appointed day

The National Health Service Act 1946 (9 & 10 Geo. 6 c. 81) received Royal Assent on 6 November 1946. Its long title is as plain as legislation gets: "An Act to provide for the establishment of a comprehensive health service for England and Wales, and for purposes connected therewith."

Scotland was covered by the National Health Service (Scotland) Act 1947 (10 & 11 Geo. 6 c. 27), and Northern Ireland by the Health Services Act (Northern Ireland) 1948, which came into force on the same day.

The appointed day was 5 July 1948, fixed by Bevan and announced to the Commons on 9 February 1948. He marked it at Park Hospital, Davyhulme, which he opened that day as the first NHS hospital, receiving the keys from Lancashire County Council; the first patient treated there was Sylvia Diggory, aged thirteen. The hospital was renamed Trafford General in 1988.

Every household was told what it was getting, and told plainly that it was not a favour. The Ministry of Health's leaflet The New National Health Service went out ahead of the appointed day: "Your new National Health Service begins on 5th July.

What is it? How do you get it? It will provide you with all medical, dental and nursing care. Everyone — rich or poor, man, woman, or child — can use it or any part of it. There are no charges, except for a few special items.

There are no insurance qualifications. But it is not a 'charity'. You are all paying for it, mainly as taxpayers, and it will relieve your money worries in times of illness." The line usually quoted — "it is not a charity, you are all paying for it" — is genuine, but it is the Ministry's wording rather than Bevan's, and the clause that completes it, "mainly as taxpayers", is usually dropped.

What Bevan said he was doing

His case was moral before it was administrative, and he said so in terms. From In Place of Fear (1952), chapter 5, "A Free Health Service":

"The collective principle asserts that the resources of medical skill and the apparatus of healing shall be placed at the disposal of the patient, without charge, when he or she needs them; that medical treatment and care should be a communal responsibility that they should be made available to rich and poor alike in accordance with medical need and by no other criteria. It claims that financial anxiety in time of sickness is a serious hindrance to recovery, apart from its unnecessary cruelty. It insists that no society can legitimately call itself civilized if a sick person is denied medical aid because of lack of means."

Two things about that passage are usually got wrong. The famous last line is a separate sentence beginning "It insists that…", not a continuation of the first; and no page number is verifiable from the available transcriptions, so it should be cited to the chapter, not a page.

From the same chapter, and directly to the point of this document: "Preventable pain is a blot on any society." That line is genuine and it is Bevan's own written words — earlier drafts of this document had it flagged as unverified, and that flag has been withdrawn.

Also from chapter 5: "A free Health Service is a triumphant example of the superiority of collective action and public initiative applied to a segment of society where commercial principles are seen at their worst."

In the Commons, moving the second reading of the Bill on 30 April 1946, he put the purpose three ways. On deterrence: "It is cardinal to a proper health organisation that a person ought not to be financially deterred from seeking medical assistance at the earliest possible stage."

On the particular cruelty of paying for diagnosis: "It is one of the evils of having to buy medical advice that, in addition to the natural anxiety that may arise because people do not like to hear unpleasant things about themselves, and therefore tend to postpone consultation as long as possible, there is the financial anxiety caused by having to pay doctors' bills."

And on charity: "I believe it is repugnant to a civilised community for hospitals to have to rely upon private charity." (HC Deb 30 April 1946, vol 422, cols 43–63.)

And the ambition in one phrase, in his words rather than the usual paraphrase. The familiar version is "universalising the best". What Bevan actually said, in the same speech, was "universalise the best": "if it be our intention that we should universalise the best, that we shall promise every citizen in this country the same standard of service…" (vol 422, col. 49.) The document uses his wording.

Which is the point. Read together, these are not statements about funding mechanisms. They are statements about what it is like to be ill and afraid, and about an obligation owed to a person regardless of their means.

That is a humanity-first proposition, stated by the founder, in 1946 and again in 1952, before the phrase existed.

What was conceded on the first day

The service was not founded on unanimity; it was founded on a deal, and the deal is visible in the statute. In January 1948 the BMA declared the scheme "so grossly at variance with the essential principles of our profession that it should be rejected absolutely by all practitioners", and a BMA plebiscite early that year went against it by roughly nine to one; a former BMA secretary, Alfred Cox, had already likened Bevan's powers to those of a "Medical Fuehrer". Five months before the appointed day, Bevan told the Commons what he had given in exchange:

"First of all, in the hospital services we have accorded paid bed blocks to specialists, where they are able to charge private fees. We have accorded, in addition to those fees for those beds which will have a ceiling, a limited number of beds in the hospitals where there is no ceiling at all." And: "We have also conceded that general practitioners and specialists can have private patients."

(HC Deb 9 February 1948, vol 447, cols 35–50, at col. 47.) This was not an afterthought. Pay beds are in the founding Act itself — section 2, "Accommodation available on part payment", and section 3, "Accommodation for private patients".

Consultant salaries from 1948 ran £1,836–£2,970 with distinction awards of £500–£2,500. Bevan's assurance to the profession on clinical autonomy was equally explicit: "My job is to give you all the facilities, resources and help I can, and then to leave you alone as professional men and women to use your skill and judgement without hindrance."

The structure that resulted was three services, not one. Hospitals were taken into public ownership under 14 regional hospital boards, 36 boards of governors for teaching hospitals and 388 hospital management committees.

GPs, dentists, opticians and pharmacists remained independent contractors, administered by 138 executive councils. Public health, community nursing, health visiting, ambulances and maternity services stayed with 147 local health authorities.

The fault lines that every subsequent reorganisation has tried to close — between hospital and community, between primary and secondary care, between health and local government — were designed in on day one, as the price of getting anything at all.

Why this belongs in a document about culture. The 1948 settlement is a precise illustration of the argument in Volume 4. What Bevan needed to make permanent, he put in the Act: ownership of the hospitals, the funding route through general taxation, the abolition of charges at the point of use, and — as the price — the private practice concessions. All of it structural, all of it still there in some form.

What he stated as purpose, he stated in a speech and a book: the obligation to the frightened person, the repugnance of relying on charity, the refusal to let means decide. None of that went into anything with a mechanism behind it.

Seventy-eight years later the first list is still operating and the second list is a set of quotations.

The NHS Constitution: what it is, and what it actually binds

The Constitution was published on 21 January 2009, and that date is fixed in statute. Section 1(1) of the Health Act 2009 defines the NHS Constitution as "the document entitled 'The NHS Constitution' published by the Secretary of State on 21 January 2009", or any revised version; section 1(2) does the same for the Handbook.

The duty it creates is a duty to think about it. Health Act 2009, section 2, "Duty to have regard to NHS Constitution": each listed body "must, in performing its health service functions, have regard to the NHS Constitution".

The listed bodies now include NHS England, integrated care boards, local authorities, NHS trusts and foundation trusts, NICE, the Care Quality Commission and the Health Services Safety Investigations Body, and section 2(4) extends the duty to everyone providing health services under NHS contracts, including sub-contractors.

The Constitution's own summary uses the same register: those bodies "are required by law to take account of this Constitution in their decisions and actions". Take account of — not comply with.

And the Constitution says of itself that it creates no rights. From the Handbook to the NHS Constitution for England (DHSC, last updated 24 January 2025), under "Rights and pledges": "A right is a legal entitlement protected by law.

The constitution does not create legal rights but it sets out a number of rights, which include rights conferred explicitly by law and rights derived from legal obligations imposed on NHS bodies and other healthcare providers."

The pledges are weaker still, on the Constitution's own statement: "Pledges go above and beyond legal rights. This means that pledges are not legally binding but represent a commitment by the NHS to provide comprehensive high quality services."

The values sit outside even that. The Constitution distinguishes rights (restatements of entitlements that exist in law elsewhere) from pledges (expressly not binding) — and makes no comparable statement of any kind about the six values.

They are listed, prefaced by the note that they were "derived from extensive discussions with staff, patients and the public", and the Handbook adds only that "these values are not intended to be limiting".

There is no enforcement provision attached to them anywhere. The only legal hook is the generic section 2 duty, which attaches to the document rather than to the values, and which requires consideration rather than compliance. In law, compassion, respect and dignity are matters to be taken into account.

The review machinery has not been working either. Section 3 of the Health Act 2009 requires a full review of the Constitution at least once every ten years, with consultation of patients, staff, carers and the public; the first was due by 5 July 2018. A ten-year review consultation finally opened on 30 April 2024 and closed on 25 June 2024.

On 3 March 2025 the government closed it without a revised Constitution, acknowledging that the general election "disrupted the consultation and people's ability to respond" and that "more could have been done to ensure the accessibility and inclusivity of the consultation", and undertook to develop "the next consultation on the NHS Constitution to align with the 10 Year Health Plan's ambitions", with timings "set out in due course".

The document in force is still substantively the 2021 text; the last substantive change was made on 1 January 2021, following EU exit.

The seven principles and the six values, as published

The Constitution's own framing sentence: "Seven key principles guide the NHS in all it does. They are underpinned by core NHS values which have been derived from extensive discussions with staff, patients and the public."

The seven principles

1. The NHS provides a comprehensive service, available to all
2. Access to NHS services is based on clinical need, not an individual's ability to pay
3. The NHS aspires to the highest standards of excellence and professionalism
4. The patient will be at the heart of everything the NHS does
5. The NHS works across organisational boundaries
6. The NHS is committed to providing best value for taxpayers' money
7. The NHS is accountable to the public, communities and patients that it serves

Worth noticing what is not in that list: "free at the point of delivery" is not one of the seven principles. Nor is the familiar triad — that the service meet the needs of everyone, be free at the point of delivery, and be based on clinical need rather than ability to pay.

That formulation was official: it appeared on nhs.uk as "NHS core principles", dated 23 March 2009. The page has since been withdrawn and its URL now redirects to the Constitution, and no current official source states it. Where this document uses it, it is attributed to that withdrawn page.

The six values, verbatim

1. Working together for patients — "Patients come first in everything we do. We fully involve patients, staff, families, carers…"
2. Respect and dignity — "We value every person — whether patient, their families or carers, or staff — as an individual, respect their aspirations and commitments in life, and seek to understand their priorities, needs, abilities and limits."
3. Commitment to quality of care — "We earn the trust placed in us by insisting on quality and striving to get the basics of quality of care — safety, effectiveness and patient experience — right every time."
4. Compassion — "We ensure that compassion is central to the care we provide and respond with humanity and kindness to each person's pain, distress, anxiety or need."
5. Improving lives — "We strive to improve health and wellbeing and people's experiences of the NHS."
6. Everyone counts — "We maximise our resources for the benefit of the whole community, and make sure nobody is excluded, discriminated against or left behind."

Law, or chance: the gap this document is about

Set the two lists side by side and the argument makes itself. Written into the Act and still running: public ownership of the hospitals, funding from general taxation, no charge at the point of use, and the private practice concessions that bought the profession's cooperation.

Written into a speech, a book and, sixty years later, a document that creates no rights: that nobody should be deterred by cost from seeking help early; that financial anxiety in illness is a cruelty as well as a clinical problem; that relying on charity is repugnant in a civilised community; that compassion should be central and that respect and dignity are owed to every person.

The two values every inquiry in this document finds missing — compassion, and respect and dignity — are values four and two of the NHS's own constitution. They were not omitted. They were written down and left without a mechanism.

No executive is appointed, appraised or paid against them; no regulator rates a trust on them as such; no board has failed a test because of them; and the statutory duty that attaches to the document containing them is a duty to have regard to it.

Which is why humanity-first healthcare is not an import. It does not propose new values for the NHS. It proposes doing for the humane half of the 1948 settlement what Bevan did for the financial half: putting it into structure, so that it survives a bad night and a hard winter.

Volume 10 makes that case; Volume 4 explains, through Johnson & Johnson's Credo, what the maintenance of such a commitment actually looks like when an organisation takes it seriously; this annex is the evidence that the commitment already exists and has simply never been wired to anything.

The lines that do not survive checking

• "The NHS will last as long as there are folk left with the faith to fight for it." Universally attributed to Bevan and traceable to no primary source — no speech, no date, no Hansard record, and it is not in In Place of Fear chapter 5. The wording drifts between versions, which is the usual sign. Not used in this document.
• "I stuffed their mouths with gold." A reported private remark, attributed to Bevan by Brian Abel-Smith in The Hospitals 1800–1948 (1964); no contemporaneous record. It is also unnecessary: Bevan's own words in the Commons on 9 February 1948 say the same thing and are directly citable. Used only with that attribution, if at all.
• "Preventable pain is a blot on any society." Previously flagged here as unverified. It is genuine — In Place of Fear, chapter 5. The flag has been withdrawn.
• The "three core principles" triad. Genuine official wording from nhs.uk, 23 March 2009, but that page has been withdrawn and no current official source states it. Attribute to the withdrawn page, or quote principles 1 and 2 of the Constitution instead.
• A page number for In Place of Fear. Earlier drafts cited p.100. No pagination is verifiable from the available transcriptions. Cite the chapter, not a page.
• The Northern Ireland legislation. Named as the Health Services Act (Northern Ireland) 1948, in force 5 July 1948. Its chapter number and date of Royal Assent could not be verified — legislation.gov.uk does not hold it. Do not assert either.
• Distinction awards as a negotiated concession. Their existence from 1948 at £500–£2,500 is sourced. That they were granted as part of the bargain with the consultants is widely repeated and not established from an authoritative source. Assert the existence, not the motive.

Fuller reference set, including the social care and dentistry figures and the founding material behind Volumes 1 and 10, is in the project doc 15 — UK healthcare and social care in numbers. Accuracy flags for everything in this annex are consolidated in the Sources and accuracy volume.

The care emergency: social and children's care

Social and children's care is where the most vulnerable people meet the least oversight: forty years of inquiries, a market in placements for children, councils squeezed by overspends, and a join with the NHS where people get stuck.

Leeds, Manchester and Camden show it can be turned round. The one thing to take away: this is an emergency, and restructuring alone will not fix it — stable staff, lasting relationships and leaders who stay do.

Why the word is emergency

Social care is not the better-run half of the system; it is the worse-run half, and the least watched. The annex on UK healthcare in numbers sets out the structure. This annex is about what that structure does to people — because in social and children's care the failure is not a backlog or a missed target.

It is children placed in illegal homes, adults with learning disabilities abused behind locked doors, and families told nothing until a journalist or an inquiry does the telling.

The word emergency is used deliberately. An emergency is a situation in which people are being harmed now, the harm is known, and the response is not yet equal to it. Every part of that definition is met.

Five facts that set the scale.

• 669 children were living in illegal, unregistered children's homes on 1 September 2025, at an average cost of £10,500 ($14,175) a week — about £353m ($477m) a year (Children's Commissioner, 11 January 2026). In September 2024 it was nearly 800, about one in ten children in residential care (Public Accounts Committee, January 2026).
• 83.5% of community social care services have no current CQC rating; the number of social care assessments the regulator published fell by 63% in two years.
• England is the only nation in the UK with no professional register for frontline care workers. A care worker dismissed for neglect can be hired by another agency the following week.
• Private equity owns or part-owns 11 of the 20 largest children's home and fostering providers (Common Wealth, 3 September 2026); the fifteen largest providers made operating margins averaging 22.6% between 2016 and 2020 (Competition and Markets Authority, March 2022).
• 80% of councils overspent on adult social care in both 2024/25 and 2025/26, and on any given day in March 2025 nearly six in ten hospital patients ready for discharge were stuck because the care to go home to was not there.

Children's care: forty years of the same finding

These are not recounted out of morbid interest. They are recounted because each one names a failure point, and because the failure points have barely changed.

The pattern is the one this whole document describes. Information held in one place and not acted on in another. Concerns treated as someone else's to own. Children who became invisible because nobody's job required them to be seen.

The Independent Inquiry into Child Sexual Abuse reported in full in October 2022, with 20 recommendations; the same failure points appear in the reviews that came after it.

The same failure to listen, outside the NHS. Oxford shows what happens when the children most at risk are not believed. The 2015 serious case review into child sexual exploitation in Oxfordshire, which followed the Operation Bullfinch prosecutions, found that more than 370 children had been identified as at risk, and that "at times, the girls' accounts were disbelieved or thought to be exaggerated". They were "treated without common courtesies"; one victim described "snide remarks".

In September 2026, survivors told BBC Panorama that "hundreds" of abusers are still free, one said police had not believed her for years, and the former detective who led the first investigation said: "We should've done more."

Thames Valley Police says it works "tirelessly" to bring offenders to justice and that it is "never too late to come forward"; Oxfordshire County Council, which has apologised for its failings, says today's safeguarding arrangements are very different.

A statutory national inquiry, chaired by Baroness Anne Longfield, began work on 31 March 2026 into grooming gangs across England and Wales and "the failures of institutions responsible for protecting" children.

The lesson for healthcare is the same as Chester's: the warning was there, in the words of the people with least power, and the institutions did not listen.

Children's care now: the market in vulnerable children

The current scandal is not a single home; it is the system of placing children at all. Councils short of foster carers and registered beds are placing children in homes that are illegal because they are unregistered — holiday lets, caravans, rented houses with agency staff.

They are also extraordinarily expensive: Ofsted has described costs of up to £30,000 ($40,500) a week for a single child, and an April 2026 study put the range at £20,000–£40,000 ($27,000–$54,000).

An investigation by the Bureau of Investigative Journalism published on 8 September 2026 found £255m ($344m) paid to more than 480 companies for around 1,800 placements over two years.

What that looks like for a child. A fifteen-year-old girl, moved by her council to escape sexual exploitation, was placed in an unregistered home staffed by two former soldiers; they plied her with alcohol and abused her, and were jailed in April 2025 for eleven and nine years.

An agency supplying staff to such placements, Prospero Health and Social Care, had 44 safeguarding allegations made against its workers between January 2024 and April 2026, sixteen of them upheld, while receiving £7.6m ($10.3m) from councils.

Another operator, SafeSpace4U, is under criminal investigation; nothing has been determined against it.

The law is only now catching up. In March 2026 Ofsted secured its first ever successful prosecution of an illegal children's home provider: Catalyst Care Limited, in Kent, which had received more than £1.7m ($2.3m) from councils, was fined £92,400 ($124,740) in August 2026.

The Children's Wellbeing and Schools Act 2026 (Royal Assent 30 April 2026) gives Ofsted power to fine providers, including of unregistered homes, introduces financial oversight of the largest providers, and creates a power to intervene on excessive profits — a backstop, not a cap.

Lincolnshire: a local story that is the national picture

Pain in the body, left untreated, does not resolve itself; it is telling you something. Lincolnshire is what that looks like in a council's accounts. In September 2026 Lincolnshire County Council was forecasting, from its first three months, an overspend of £12.6m ($17.0m) across children's services for 2026/27, around £9m ($12.2m) of it on children's social care.

Some children in its care cost "up to £18,000 ($24,300) a week", the deputy chief executive, Heather Sandy, told councillors: "I'm not sure that is well known."

The council had already taken £5.4m ($7.3m) of its £7.6m ($10.3m) contingency fund to cover it, only three months into the year. "The even more worrying thing is that the contingency has already been raided," said Councillor Martin Hill.

An independent member of the scrutiny committee, Andrew Middleton, named the problem: "None of us have control over demand. None of us has control over complexity of need. The difference between a moderate need and a higher-level need can be three, four, five thousand a week."

More children are coming into care, and they are more troubled when they arrive. Lincolnshire's rate of children in care rose from 55 to 61 per 10,000 children between June 2025 and June 2026, above the council's own target of 59, with the sharpest rise among 15–17-year-olds. "We're seeing families in crisis, we're seeing deprivation," Ms Sandy said.

The councillor responsible, Natalie Oliver, was clear about the priority: "Some children have a very high level of trauma and in order to give them the therapeutic care that they need, sometimes they are really expensive … At the end of the day, we're here to protect children and their welfare is our overwhelming priority."

The council has joined a regional care co-operative to find more stable placements and set up a reunification team to help children return home safely. And the pressure is spreading: its schools budget, driven by special educational needs, is forecast to overspend by £51m ($68.9m) this year.

Lincolnshire is not unusual. It is typical. Older and more rural than most of England, it is caught in exactly the same squeeze as everywhere else.

• More children in the most expensive care. England had 81,770 children in care at 31 March 2025, 2% fewer than a year before — but 9,480 of them were in children's homes or secure units, up 9% in a single year (Department for Education).
• The cost of a place has almost doubled. Councils spent £3.1bn ($4.2bn) on residential care in 2023-24. The average children's home place now costs £318,400 ($429,840) a year, almost double five years earlier and eight times the cost of foster care; 84% of the market is run by private providers (National Audit Office, September 2025).
• The extreme cases are no longer rare. Placements costing £10,000 ($13,500) a week or more rose from about 120 in 2018/19 to more than 1,500 in 2022/23, and the share of councils with at least one rose from 23% to 91%. The highest reported was £63,000 ($85,050) a week (Local Government Association survey, November 2023).
• Children are sent away. At 31 March 2025, 18,120 children in care — 22% of them — were living more than 20 miles from home, and 47% of those in children's homes or secure units (Become, from Department for Education data).
• And it is taking over council budgets. Children's social care rose from 18% of council budgets in 2011/12 to 27% by 2024/25; in 2023/24, £2.3bn ($3.1bn) of the £13.3bn ($18.0bn) spent had not been budgeted for (Institute for Government).

The pain has a cause, and for fifteen years it was left untreated. Between 2010/11 and 2022/23, councils' spending on early intervention — the help that keeps families together before a crisis — fell by £1.8bn ($2.4bn), or 44%, while spending on late intervention rose by £3.6bn ($4.9bn), or 57%.

In 2022/23, for the first time, councils spent more on residential care (£2.4bn) than on all early intervention put together (£2.2bn) (Children's Charities Coalition and Pro Bono Economics, September 2024).

Spending on youth services fell by 76% in real terms over the same period, to £419m ($566m) in 2024–25 (YMCA, February 2026). The problem did not go away. It moved down the line, to the point where it is most expensive and does most harm.

And the ripples have barely begun. A child moved far from home and between placements loses school, friends and the adults who knew them; the annex on why being human is good for health records that unstable placements raised the probability of behavioural problems from 22% to 36%.

At 31 March 2025, 40% of care leavers aged 19 to 21 — more than 15,000 young people — were not in education, employment or training, up 15% in five years, and 2,540 were living in accommodation their own council judged unsuitable (Become, from Department for Education data).

Those costs arrive later, in other budgets — benefits, housing, mental health and the justice system — and in lives. The same pattern is running through special educational needs: councils' accumulated deficits were projected at £6.6bn ($8.9bn) by February 2026, and the government has committed to write off 90% of them, conditional on local reform plans (County Councils Network).

The treatment is known, and it is the same everywhere. Keep families together where it is safe to do so; find, support and keep foster carers; build homes close by so that children are not sent away; and let the same people stay with the same child.

The lighthouses later in this annex — Leeds, Hertfordshire, North Yorkshire — did exactly that. The question Lincolnshire's accounts put to every council, and to the government, is whether it can find the money to do the cheap thing early while it is still paying for the expensive thing now.

Teesside: the society that prevention was meant to protect

If Lincolnshire shows the cost of late help in a council's accounts, Teesside shows it in a community. Middlesbrough has one of the highest rates of children in care in England: 150 per 10,000 children in 2023, or 513 children, more than double the England average of 71 (Blue Cabin, from Department for Education data).

The North East as a whole has the highest rate of any region, 114 per 10,000 in 2024 against 70 nationally. And in the Middlesbrough and Thornaby East constituency, 52% of children live in relative poverty after housing costs, one of the highest rates in the country (End Child Poverty, 2025).

The warning was given, in plain words, six years ago. When Ofsted judged Middlesbrough's children's services inadequate in January 2020, it found "a culture in which harm is tolerated, rather than one in which there is a focus on how to protect children".

Young people's exploitation was being treated as "lifestyle choices", and "too often care leavers are treated as independent adults who will seek support if needed".

Services have improved since, under a government-appointed commissioner and years of monitoring. But after its full inspection in June 2026 Ofsted still judged every area to require improvement, with a "churn of leadership" blamed (Teesside Live).

A few years on, the same pattern shows up on the streets. Cleveland has the second highest knife crime rate in England and Wales, 154 offences per 100,000 people, and ranks seventh for hospital admissions for assault with a sharp object.

The violence is concentrated: 47% of violence with injury takes place in just ten wards, and Middlesbrough's Central ward alone accounts for more than 11%. The police and crime commissioner's own assessment names what lies behind it: "Deprivation, experience of services, exclusion from school, experience of domestic abuse and prior exposure to the criminal justice system are all identified as strongly linked with serious violence offending" (Cleveland Strategic Needs Assessment, data 2017–2022).

The North East also has the highest rate of drug misuse deaths in England and Wales: 107.1 per million people in 2024 registrations (ONS). In August 2026 more than 200 extra police officers were sent to Teesside.

A Volkswagen Passat on cloned plates, fleeing police, had driven the wrong way down the A66 and hit an armed response car head-on, killing two officers, PC Tom Clough and PC Matthew Blades, and all five young men in the Passat; days later a house fire killed a woman and a seven-year-old girl.

Residents of one estate told a reporter they were afraid to be seen talking to him (UnHerd, 31 August 2026).

The young men in the Passat had been seen by the system again and again. Four of the five had criminal records which, on court records and Cleveland Police information, came to more than thirty offences between them: six assaults on police officers, carrying a knife in public, possessing drugs, racially aggravated harassment, vehicle theft, aggravated vehicle taking, theft from cars, a dozen shoplifting offences and a string of driving offences.

One was already under a youth rehabilitation order. Two of them were seventeen. The fifth had no previous convictions (ITV News Tyne Tees, 24 August 2026). Nothing in the reporting says whether any of them had been in care, and nothing here is a judgment on them as people.

The point is about the system: boys who came to the attention of the police and the courts dozens of times before they were adults, and whose direction of travel nobody changed.

And at eighteen, the support stops. More than a quarter of people aged 20 to 34 in the UK still live with their parents (ONS, 2024). Young people in care are expected to become independent at eighteen, and sometimes at sixteen or seventeen when they are moved into supported accommodation — what the charity Become calls the "care cliff".

Care leavers are nine times more likely than other young people to become homeless; the number of young care leavers facing homelessness in England has risen by 54% in five years; and only 38% of councils offer a rent deposit guarantee (Become). Nationally, 40% of care leavers aged 19 to 21 are not in education, employment or training.

Teesside is a mirror, not a proof. Its troubles have many causes that no children's service could have prevented: the closure of the Redcar steelworks in 2015, decades of lost industry, and a police force that, on one account, shrank from 1,756 to 1,198 officers between 2009 and 2019.

The point is narrower, and it is the one this document keeps arriving at. Children who were not protected early, and were left on their own at eighteen, do not disappear from the system. They reappear, as the victims and the perpetrators of violence, in addiction services, in hospital, in court and in grief.

When a society stops investing in prevention, it does not save the money. It pays it later, with interest, and in lives.

Adults: behind the locked door, and against the clock

The adult scandals are quieter, and they recur because the conditions that produce them are permanent.

• Winterbourne View (2011) — adults with learning disabilities abused at a private hospital near Bristol, exposed by an undercover reporter, not a regulator. Margaret Flynn's review called it "a case study in institutional abuse" and concluded that "the apparatus of oversight across sectors was unequal to the task". Eleven staff were convicted.
• Orchid View (2013) — a Southern Cross care home in West Sussex. The coroner found that neglect contributed to five of nineteen deaths and called the home "an accident waiting to happen".
• Whorlton Hall (2019; trial 2023) — again exposed by undercover filming, not inspection. Nine staff were tried; four were convicted of ill-treatment.
• Blue Ribbon Healthcare (2026) — after an ITV News investigation in July and August 2026 into supported living and care services for adults with learning disabilities, autism and mental health conditions, the CQC suspended ratings at two sites and opened an investigation. Nothing has yet been determined.
• Care against the clock. The Equality and Human Rights Commission's Close to home inquiry (November 2011) heard from 1,254 people and found home care commissioned by "time and task" breaching older people's human rights. Two years later, 73% of councils were still commissioning fifteen-minute visits (UNISON, 2013) — fifteen minutes to wash, dress, feed and medicate a frail person, with no time to talk to them.
• Silenced by eviction. An ITV investigation in 2023 found more than 11,000 notices to quit issued to care home residents in eighteen months to April 2022; families described being too frightened to complain in case their relative was moved.

Mental health: the same pattern, and three inquiries open now

• Edenfield Centre (Greater Manchester Mental Health NHS FT) — filmed by Panorama in 2022. Oliver Shanley's review (January 2024) found "all the hallmarks of a closed culture, including an absence of psychological safety, incivility between staff, poor leadership". Two staff have been charged with ill-treatment; nothing has yet been determined.
• Tees, Esk and Wear Valleys NHS FT — a statutory public inquiry into the deaths of young women in its care, including Christie Harnett, Nadia Sharif and Emily Moore. His Honour Judge Roderick Potter was appointed chair on 17 September 2026; the inquiry began on 21 September 2026.
• Nottinghamshire Healthcare NHS FT — the Nottingham Inquiry into the care of Valdo Calocane, chaired by Her Honour Deborah Taylor, heard evidence from February to June 2026 and received closing submissions in September 2026. The CQC had already found premature discharge and poor follow-up.
• St Andrew's Healthcare, Northampton — in 2026 NHS England instructed commissioners to find alternative placements, affecting 287 inpatients.

The money, and the join with the NHS

The overspend looks small. The squeeze it causes is not. English councils overspent their adult social care budgets by £715m ($965m) in 2025/26 — about 3.2% of a net budget of £22.6bn ($30.5bn) — after £774m ($1.04bn) the year before, the highest in a decade; four in five councils overspent in both years (ADASS Spring Survey, July 2026).

Children's social care is worse: of £13.3bn ($18.0bn) spent in 2023/24, £2.3bn ($3.1bn) — 21% — had not been budgeted for (Institute for Government).

In most organisations a 3% overspend is absorbed. A council cannot run one. It is legally required to set a balanced budget every year, and if spending is heading beyond its resources its chief finance officer must issue a section 114 notice — in effect, a declaration that it cannot pay its way.

So the gap has to be found within the year, from the rest of the council: two-thirds of councils covered their 2025/26 adult care overspend partly from underspends in other departments, and six in ten from reserves (ADASS).

For 2026/27, 35 councils were approved for about £1.5bn ($2.0bn) of exceptional financial support from government simply to balance their books, and the government's list now runs to 36.

Because care is now most of what a council does, there is almost no slack. Adult social care alone takes 35–38% of the entire net budget of a council with care responsibilities (ADASS: 35.5% in 2025/26, 37.9% in 2026/27).

Add children's social care and the two consume about 65%, up from 57% in 2014 (County Councils Network); the Institute for Fiscal Studies puts them at 65% of councils' non-education spending, up from 50% in 2010–11.

Everything else a council does has to fit into what is left. Between 2010–11 and 2019–20 spending per person fell by 45% on transport, 44% on housing and 58% on planning (IFS), and the County Councils Network names libraries, road repairs, street lighting and parks as the services that lose out.

A social care problem is therefore never contained in social care. It is paid for by every other local service people see.

The join between the NHS and social care is where people get stuck. On a typical day in 2025/26 nearly 13,000 patients in English hospitals were well enough to leave but still there, at an estimated cost of £2.7bn ($3.6bn) a year, or £562 ($759) a bed day (King's Fund, July 2026); the Health Foundation puts delays at 11% of all bed days.

For a frail older person every extra day on a ward is a day away from home, from routine and from the people they know.

Who is responsible is less clear than the old term "bed blocking" implies: the last national data that assigned responsibility, in February 2020, put 60% of delays down to the NHS, 30% to social care and 10% to both (Nuffield Trust), and today's data record reasons rather than responsibility.

What is clear is that two systems, funded, run and held to account separately, meet at the ward door, and the patient waits in the gap between them.

But joining them up in structure is not the answer on its own — and Northern Ireland proves it. Northern Ireland has run health and social care as one integrated system since 1973; today five Health and Social Care Trusts deliver both.

On paper it is what England keeps trying to build. In practice it has the worst waiting times in the UK: in December 2022, 50.9% of patients there had waited more than a year, against 5.4% in England (Northern Ireland Audit Office, 2024), and in March 2026 more than half of the 510,680 people waiting for a first outpatient appointment had waited over a year (Department of Health NI).

Delayed discharges persist, with about 40% of lost bed days linked to a shortage of care home places (departmental figures, reported July 2026).

The Nuffield Trust found integration there "statutory rather than delivered in practice"; that pooled budgets "have not succeeded in shifting more resources towards community and social care services"; and that inside a single structure social care risked having its "pocket picked", with cuts to home care "one of the primary means to deliver savings".

Care packages for older people fell from 26.2 to 20.6 per 1,000 people aged sixty and over between 2014 and 2024 (Commissioner for Older People for Northern Ireland). The expert review led by Rafael Bengoa in 2016 was titled, pointedly, Systems, Not Structures.

Scotland tells the same story. It legislated for integration joint boards in 2014. In January 2026 Audit Scotland reported "a long-term picture of generally declining performance and satisfaction", with days spent in hospital by people over seventy-five who were ready to leave up 20% since 2018/19.

In England, the National Audit Office found in 2017 "no compelling evidence" that integration leads to sustainable savings or reduced hospital activity.

The lesson: structure is not a silver bullet. Every reorganisation promises that joining the boxes will join up the care. Northern Ireland has had the joined-up boxes for more than fifty years. What decides whether a frail person gets home is whether there is a care worker to go to them, a budget that protects that care when money is tight, and people on both sides who treat the patient's life as their shared responsibility.

That is a question of priorities and culture, not of organisation charts — which is the argument of this whole document.

What the scandals share

Every one of them fails at the same few points, and none of them is a medical point.

1. Closed cultures. Winterbourne View, Whorlton Hall and Edenfield were all exposed by cameras, not by the organisations or their regulators. Where the people being cared for cannot speak for themselves, the only protection is the conduct of the staff — and nobody was watching it.
2. Oversight by paperwork. Inspectors were shown records; the records were not the care.
3. Care bought by the minute. Fifteen-minute visits and placements priced per week turn a relationship into a transaction, and the transaction is what gets managed.
4. Profit taken out of the most vulnerable. Margins of over a fifth, earned from councils that cannot find anywhere else to put a child.
5. Risk that nobody owns. From Laming to Southport the same sentence recurs: every agency held a piece, and none of them held the child.
6. A workforce nobody registers. In the part of the system with the most vulnerable people and the least supervision, England alone does not know who its care workers are.

That is the universe-of-obligation argument of Volume 3, in its purest form. The people in these places are, very often, the people with the least voice anywhere — children in care, adults with learning disabilities, the very old, the detained. They fall outside the circle first.

The lighthouses: what good leadership has already done

None of this is inevitable, and the proof is in English councils, not foreign case studies. Each of the places below started from failure or from the same constraints as everyone else, and each changed direction because someone in charge decided what the service was for and stayed long enough to make it stick.

Manchester children's services: Inadequate to Outstanding. Rated Inadequate by Ofsted in 2014, Good in 2022, and Outstanding in November 2025 — the first in the North West.

Across Greater Manchester more widely, under devolution and Andy Burnham's mayoralty, the measured gains are real but modest, and should be stated as such: life expectancy rose an extra 0.2 years against comparable areas (Lancet Public Health, 2022); alcohol-related admissions fell 11.1% and satisfaction with adult social care rose 17.6%, while A&E waits lengthened 12.2% and adult obesity rose (Social Science & Medicine, 2024).

His "A Bed Every Night" programme saw rough sleeping fall 57% between 2017 and 2020 against 43% nationally, though its evaluators also flagged weak move-on into permanent homes. Several Greater Manchester councils committed to paying care workers the real living wage — a commitment rather than a measured outcome, but the right signal about whom the system values.

Nottinghamshire, and Anthony May: running towards the fire. When he was the county's Director of Children's Services, serious case reviews in 2009 found child protection procedures that "should have been initiated but were not", and Ofsted judged safeguarding inadequate in 2010.

The improvement notice was lifted in 2011 and children's services were rated Good in 2015, the year he became the council's Chief Executive.

In 2022 he took over Nottingham University Hospitals in the middle of its maternity crisis; by 2023 the CQC had moved the trust's leadership from Inadequate to Requires Improvement and recorded "significant progress in improving the culture". He was appointed OBE in 2023.

The honest context belongs alongside it: the IICSA report of 2019, published while he was Chief Executive, found that "neither of the Councils learned from their mistakes" over five decades, while crediting the County with a public apology and good later work on harmful sexual behaviour.

A lighthouse does not make the night disappear. It shows that one leader facing the problem squarely changes what is possible.

What the lighthouses have in common is not money. It is leaders who stayed; practice models that keep the same people around the same child or adult; relationships instead of transactions; and an organisation that treated the person being cared for as its reason for existing rather than as a case to be closed or a placement to be priced. That is the founding ethos of 1948, applied to the part of the system Bevan's settlement left out.

Where it goes next

Baroness Casey's commission on adult social care has not yet reported. It was set up to produce a first report in 2026 and a final one in 2028; in July 2026 the government said it wanted the commission's work completed sooner, by summer 2027.

In March 2026 Casey asked for early action — a dementia tsar, a national safeguarding board, a fast-track care passport for people with motor neurone disease — and described a system held together with "sticking plasters and glue".

The public consultation, the Big Conversation on Care, opened on 29 July 2026. Andy Burnham, now Prime Minister, is the author of the 2010 white paper Building the National Care Service. The opportunity is real. So is the record of commissions whose findings were admired and not implemented.

Doctors and nurses: when the organisation stops caring

Doctors and nurses are accused of an "I'm alright, Jack" attitude. This annex tests the charge fairly and finds it is half the story: teams broken up, a broken career ladder, moral injury, and organisations that stopped treating their staff as people.

The one thing to take away: an organisation teaches its people how to treat others by how it treats them.

The charge, and why it is only half the story

Many people now see a generation of doctors that looks after itself first. The perception is not imagined. Public support for resident doctors' strikes fell from 52% in June 2024 to 28% by the autumn of 2025 (Ipsos); in April 2026, during what the Health Secretary called the fifteenth strike since 2023, YouGov found 37% in support and 55% against.

When the dispute finally settled, on 29 June 2026, the offer was accepted by 52.9% to 47.1% — nearly half still voted to reject — and the committee chair said the profession was "still nearly a fifth behind 2008 levels of pay".

To a patient whose operation was cancelled for the third time, or to a senior clinician who remembers unpaid nights as the price of the job, it can look like "I'm alright, Jack".

But "I'm alright, Jack" is what an organisation gets back when it has spent fifteen years telling its people "you're on your own". This annex is not a defence of every strike or every demand, and it is not a charge sheet against a generation.

It is an account of what happened to the job — because the argument of this whole document is that behaviour follows conditions, and the conditions around doctors and nurses in training have been dismantled piece by piece.

An organisation that stops caring for its people should not be surprised when they stop caring for it. The damage does not stop there, because people who feel unheld do not have much left to hold others with.

What happened to the job

The name changed because the old one had stopped being true. From 18 September 2024 "junior doctors" became "resident doctors", with 91% of those surveyed in favour. Many of them are in their thirties, several years qualified, carrying the hospital at night.

The team they belonged to disappeared. The old "firm" — a consultant with the same registrars and juniors, working, learning and eating together — has largely gone, replaced by rotas that move doctors between wards and hospitals to fill gaps.

The GMC's own review, Caring for doctors, caring for patients (West and Coia, 2019), heard that "changes in many organisations, such as the loss of team structure, have left them feeling exposed, with increases in bullying, blaming and undermining", and that multiple rotations undermined doctors' "ability to be a member of any team".

Health Education England's report Enhancing junior doctors' working lives (2017) recorded the plainest version: "I do not feel valued"; "most doctors in training felt disconnected from their employers"; rotations leave them "feeling isolated and disconnected".

Much of the working day is not doctoring. One observational study timed 137 resident doctors with stopwatches: 43.4% of their time went on non-patient-facing administration and only 11.7% on patient-facing tasks that needed a doctor (QJM, 2025). It is a small study; it matches what they say.

The ladder broke. In 2024, 25,496 doctors applied for 12,743 specialty training posts, and in 2025 the number of applicants rose to 33,108.

Only 31.4% of doctors finishing the Foundation Programme in 2025 went straight into UK specialty training, and 12.8% went abroad; in 2011, on an earlier survey method, 71.6% went straight into training.

In 2024, 16,124 doctors stopped holding a licence to practise, up from 11,309 the year before. Parliament passed the Medical Training (Prioritisation) Act in March 2026 to put UK graduates first for training posts.

The pay fell — by how much depends on the measure, and honest accounts say so. The BMA's figure of a 26% real-terms cut (2008/09 to 2021/22) uses the retail price index, which the UK Statistics Authority considers unfit for purpose.

The Nuffield Trust puts the fall since 2008 at 4.7% on consumer prices and 17.9% on RPI; Full Fact, using the Household Costs Index, at six to seven per cent.

Awards since then have been substantial — around 22% over 2023–25, 5.4% for 2025/26, and a further average 6.6% under the June 2026 settlement by April 2027. Meanwhile doctors can graduate with student debt of over £100,000 ($135,000).

Whichever measure is used, the old bargain — years of hardship now in return for security and standing later — has visibly broken.

And the strain shows. The GMC's 2025 training survey found one in five doctors in training at high risk of burnout, 21% apprehensive about raising concerns with their supervisors, and 28% experiencing micro-aggressions; in 2026 19% were at high risk and a further 42% at moderate risk.

Nurses: the same story, told more quietly

Nursing has lived through the same loss, with less public attention and less leverage. On 15 December 2022 the Royal College of Nursing held the first national strike in its history.

In January 2025 more than 5,000 nurses answered its survey on corridor care: two-thirds said they had delivered care in corridors or other unsuitable places, and nine in ten said patient safety was being compromised.

Of those who left the register in 2024/25, 47% left earlier than they had planned, with workload and staffing levels the commonest contributing reasons.

And the pipeline is jamming at the far end: in August 2025 the government announced a graduate guarantee after finding up to three times as many newly qualified nurses as vacancies in some areas, and a year later the RCN found a third of entry-level posts advertised in July 2026 were closed to new graduates.

People who trained for three years to care for others are being told there is no room for them, while patients wait in corridors.

How an organisational failure becomes a failure of humanity

The chain is short, and each link is measured.

1. The organisation stops seeing its people as people. Only 42.95% of NHS staff feel their organisation values their work, and only 32.83% say there are enough staff to do the job properly (NHS Staff Survey 2025, 766,285 respondents).
2. Staff are asked, every shift, to do what they know is not good enough. In the BMA's 2021 survey, 78.4% of doctors recognised moral distress in their own work and 51.1% moral injury; the leading cause was insufficient staffing.
3. People protect themselves. Burnout is the body's version of "I'm alright, Jack": 31.47% of NHS staff report feeling burnt out. Withdrawal, detachment and a narrower sense of duty are what the literature on moral injury predicts.
4. The relationship becomes a contract. Rotas without a team, training without a ladder and pay without a bargain turn a vocation into a job, and a job can be bounded in a way a vocation cannot.
5. And the patient feels the difference. Fewer than two-thirds of NHS staff — 62.84%, the lowest figure ever recorded — would be happy with the standard of care if a friend or relative needed treatment. Engagement is not a soft variable: a one-standard-deviation rise in staff engagement goes with mortality around 2.4% lower (West and Dawson, King's Fund, 2012).

Chris Turner's story is the chain in one life. He joined the emergency department at Mid Staffordshire as a junior doctor in 2007, found a culture of bullying and staff "utterly demoralised", stopped sleeping, then eating, and "after two years I broke under the pressure and resigned". He went on to become a consultant in Coventry and co-founded Civility Saves Lives.

Not everyone who breaks finds a way back into the profession. Staub's line from Volume 5 applies here in both directions: "learning by doing can change people in either positive or negative directions." An organisation teaches its people how to treat others by how it treats them.

And staff can fall outside the universe of obligation too. Volume 3 argues that harm follows when a group falls outside the circle of people an institution feels obliged to. The circle can shut out those who work in the institution as well as those it treats — the doctor moved around by a rota that does not know their name, the nurse who is told there is no post, the overseas recruit whose qualification is quietly downgraded. When staff are outside the circle, they draw a smaller one of their own.

Fairness to the strikers

The evidence does not support the claim that the strikes killed patients — nor that they were harmless. A study of the 2016 junior doctors' strike by the Institute for Fiscal Studies found "no statistically significant impacts" on mortality or readmission overall, but that Black patients experienced worse outcomes in the hospitals most exposed to the strike.

Coroners have linked strike disruption to five deaths, where causation is hard to establish. By July 2024 about 1.49 million appointments had been rescheduled, and the King's Fund put the total across all NHS industrial action since late 2022 at at least 1.7 million.

The cost is contested: the government gave NHS England £1.7bn ($2.3bn) to cover 2023/24 strikes, and independent estimates run to £2bn–£3bn ($2.7bn–$4.05bn), with around £300m ($405m) a round more recently. During the April 2026 strike, 94.1% of planned activity went ahead.

The BMA: both things are true

The doctors' own organisation has always had two faces, and both are on the record. It fought the creation of the NHS: a plebiscite early in 1948 went against the scheme by roughly nine to one, and a former BMA secretary, Alfred Cox, had likened Bevan's powers to those of a "Medical Fuehrer".

It registered as a trade union in 1971, and in 1975 doctors took industrial action for the first time.

It has also used its authority for the public good — pressing for compulsory seatbelts (it told Parliament in 1980 that there was "no alternative to … legislation"), campaigning for smoke-free public places from 2002, and calling in 2015 for a tax on sugary drinks.

(The link between smoking and lung cancer, often credited to it, was established by Doll and Hill's research, published in the BMJ, and pressed by the Royal College of Physicians.)

In 2026 both faces were visible at once. Responding to Lord Mann's review on 4 June, the BMA supported its training and monitoring measures but opposed new GMC appeal powers as "unnecessary" and said a "blanket ban" on political identifiers "would go too far".

At its annual meeting on 23 June it passed a motion, as reported, asking the government and NHS England to revoke the mandatory adoption of the IHRA definition of antisemitism until safeguards are in place; the full text and vote are not public.

Reasonable people disagree about those positions. The structural point is simpler: a union's first duty is to its members, and it should discharge it. That is exactly why the profession's collective voice cannot be the only guardian of its humanity — and why the NHS, as employer, has to be.

What rebuilds the bond

The GMC has already named what doctors need: autonomy, belonging and competence. Belonging is the one the modern NHS has taken away most completely, and it is the cheapest to restore: the same people working together long enough to know one another's names, a consultant who knows who their trainees are, a ward that expects you back.

The lighthouses in Volume 8 and in the care annex share exactly that feature — Buurtzorg's small stable teams, North Yorkshire's one team that stays with the young person, the Esther network's named relationships.

The June 2026 settlement contained the right kind of signal, if it is followed through. Beyond pay, it moves 4,000–4,500 locally employed doctors onto substantive contracts, reimburses the first two attempts at Royal College examinations, covers portfolio and college fees from April 2027, and creates a joint industrial relations committee.

Each of those says the same thing to a doctor: you belong here, and your future here matters to us. Whether it is followed by a restored ladder and restored teams will decide whether the truce becomes a relationship.

And the lesson for everyone else in the system is the one this document keeps arriving at. Humanity does not flow downhill from policy. It is passed on, person to person, by people who have been treated humanely themselves.

Staff who feel cared for are the only reliable route to patients who are. An organisation that wants its doctors and nurses to go the extra mile for a frightened stranger has to be seen, every week, to go it for them.

Why being human is good for health

Is being human — meaning, connection, kindness, warmth, being known — good for health? Weighed honestly, yes: the evidence is consistent, often modest, strongest at the two ends of life, and some popular claims do not survive checking.

The one thing to take away: humanity is part of the treatment, for patients and for the people who care for them.

Why this annex exists

The rest of this document argues that a humanity-first care system is a better care system. This annex asks a different and older question: are the human qualities themselves good for health? Not the organisation of care, but meaning and purpose, connection with other people, kindness, and being cared for by someone who is recognisably a human being — who listens, is warm, and knows who you are.

If they are, then humanity is not a soft extra layered on top of treatment. It is part of the treatment, for patients and for the people caring for them.

The honest answer is yes — in direction, consistently, across very different kinds of evidence; but the effects are smaller and less proven than the popular claims, and larger than sceptics assume. This annex sets out what is established, what is plausible, and what is overstated, because the argument is stronger when the reader can see that the evidence has been weighed rather than collected.

1. Meaning and purpose: eudaimonia

The OECD has put a name to the part of wellbeing that is about how you live, not how you feel. Its 2013 Guidelines on Measuring Subjective Well-being set out three components — how people evaluate their lives, their moods, and eudaimonia, "a sense of meaning and purpose in life, or good psychological functioning". At the time the OECD was cautious, calling the evidence base for eudaimonic measures "less clear".

By December 2024 it had moved on: a working paper proposed defining eudaimonia as "behaving and thinking in ways that are typically conducive to long-term personal well-being and the well-being of others", and noted, citing the research literature, that it is "a strong predictor of many outcomes … including health".

Two things are worth holding onto. The OECD's own definition now reaches beyond the self to other people. And the health claim is the literature's, which the OECD cites rather than tests — so the literature is where to look.

People with more meaning and purpose live longer, consistently, in very large studies.

• In the English Longitudinal Study of Ageing, 9,050 older adults followed for eight and a half years, those in the highest quarter for eudaimonic wellbeing had a 30% lower risk of dying than those in the lowest, after adjusting for age, health, depression and much else (Steptoe, Deaton and Stone, The Lancet, 2015). The authors are careful: the findings "do not unequivocally show" cause.
• A meta-analysis of ten prospective studies and 136,265 people found purpose in life associated with 17% lower all-cause mortality and 17% fewer cardiovascular events (Cohen, Bavishi and Rozanski, Psychosomatic Medicine, 2016).
• The largest analysis so far — 25 samples, 488,765 people, 48,928 deaths — found a 24% lower risk, falling to 15% after adjustment for behavioural and clinical risk factors and to 9% after adjustment for depression: "due in part but not completely" to those factors (Sutin and colleagues, Psychological Medicine, 2026).

The honest caveats matter here. The association shrinks as more is adjusted for, and a 2026 re-analysis of one American cohort found much of it disappeared when baseline health was measured properly — people who are already ill may simply find it harder to feel purposeful.

And meaning is not the same as mood: the Million Women Study of 719,671 UK women found that happiness itself had no direct effect on mortality once health was taken into account (The Lancet, 2016). What the evidence supports is the narrower claim — that meaning and purpose are consistently associated with longer, healthier lives, and that this association survives more scrutiny than happiness does.

The UK already measures it. One of the ONS's four national wellbeing questions asks how far "the things you do in your life are worthwhile" — the eudaimonic question. In the year to March 2023 the average was 7.73 out of 10, still below its pre-pandemic level.

2. Connection: the strongest evidence of all

Being connected to other people is as consistently linked to survival as anything in public health. A meta-analysis of 148 studies and 308,849 people found that those with stronger social relationships had 50% higher odds of survival over the study periods; for the fullest measures of social integration the figure was 91% (Holt-Lunstad, Smith and Layton, PLoS Medicine, 2010).

A second analysis, of 70 studies and more than 3.4 million people, found social isolation associated with a 29% higher risk of death, loneliness with 26%, and living alone with 32% (Perspectives on Psychological Science, 2015).

It is now recognised at the highest level. The WHO Commission on Social Connection reported in June 2025 that loneliness affects one person in six worldwide and is "linked to an estimated 100 deaths every hour — more than 871,000 deaths annually".

The US Surgeon General's 2023 advisory called it an epidemic and linked poor social connection to 29% higher risk of heart disease and 32% higher risk of stroke. These are associations, not counted deaths — but they are consistent, large, and repeated across countries.

Which matters for care because hospitals and care homes are where the isolated end up. A health and care system can reconnect people or complete their isolation. Every ward that has no time to talk, every fifteen-minute home visit, every rota that means a patient never sees the same face twice, is on the wrong side of this evidence.

3. Kindness is good for the person who gives it

People who help others live longer, and the effect does not look like a household artefact. Across five cohort studies, volunteers had 22% lower mortality (Jenkinson and colleagues, BMC Public Health, 2013).

A linkage study of 617,466 married people in Northern Ireland found the same — volunteers' risk of death was 22–23% lower — while their non-volunteering spouses, who shared their homes and circumstances, had no advantage at all (International Journal of Epidemiology, 2017).

In Detroit, stressful life events predicted death among people who did not help others, but not among those who did (American Journal of Public Health, 2013).

But kindness as a prescription has not yet been proven. The small randomised trials of volunteering found no consistent health benefits, and a meta-analysis of 27 kindness experiments found a small-to-moderate gain in the doer's wellbeing from studies that were mostly underpowered (Curry and colleagues, 2018); a 2024 review of preregistered experiments found "surprisingly little support" for random acts of kindness as a wellbeing strategy.

The fair reading is that kindness as a way of life goes with better health; kindness as a two-week exercise does not reliably produce it. That is an argument for building it into how people live and work, not for leaflets.

4. The genuine encounter: being cared for by a human being

The idea is old, and it was first described in words rather than numbers. Scandinavian caring science, following Katie Eriksson, and the Icelandic nurse-scholar Sigríður Halldórsdóttir, described a spectrum of ways of being with a patient, from the "biocidic" — life-destroying — to the "biogenic", life-giving (1991).

American family doctors, drawing on Martin Buber's distinction between treating someone as an "It" and meeting them as a "Thou", described the healing relationship in primary care in the same terms (Annals of Family Medicine, 2008).

This literature is qualitative and philosophical: it describes the genuine encounter precisely, but no study measures it as a single variable. The measurable evidence sits under other names — warmth, empathy, the therapeutic relationship, continuity — and it says three things.

First, warmth changes outcomes even when nothing else changes. In a Harvard trial of 262 people with irritable bowel syndrome, everyone received the same sham treatment. Among those left on a waiting list, 28% reported adequate relief after three weeks; among those treated in a brisk, businesslike way, 44%; among those treated with warmth, attention and confidence, 62% (Kaptchuk and colleagues, BMJ, 2008). The treatment was identical. The human being delivering it was not.

Second, across many trials the effect is real but small. A meta-analysis of 13 randomised trials of the patient–clinician relationship found a small but statistically significant effect on health outcomes (Kelley and colleagues, PLoS ONE, 2014); another, of 28 trials and 6,017 patients, found empathic communication modestly reduced pain and anxiety and improved satisfaction, with positive, encouraging messages having a larger effect on psychological outcomes (Howick and colleagues, Journal of the Royal Society of Medicine, 2018).

Small effects, applied to every patient in every consultation, are not small in total.

Third, and most powerfully, being known by the same doctor over time saves lives. In Norway, a study of 4.55 million people found that those who had kept the same GP for more than fifteen years were 25% less likely to die, 28% less likely to be admitted to hospital as an emergency, and 30% less likely to use out-of-hours services than those who had been with their GP for a year — with the benefit rising steadily with every additional year (Sandvik and colleagues, British Journal of General Practice, 2022).

A systematic review found lower mortality with greater continuity in 18 of 22 studies across nine countries (Pereira Gray and colleagues, BMJ Open, 2018), and a 2025 review judged with moderate certainty that continuity probably reduces mortality, admissions and emergency attendances by around 10–15%.

In England, higher continuity was associated with 6% fewer admissions that good primary care should prevent (Barker and colleagues, BMJ, 2017).

Continuity is not a technology. It is a relationship — the thing the modern NHS rota most reliably destroys.

And patients' experience of being treated well goes with safer, more effective care. A review of 55 studies found consistent positive associations between patient experience and both patient safety and clinical effectiveness (Doyle, Lennox and Bell, BMJ Open, 2013). Being treated as a person and being treated safely turn out not to be separate things.

What this annex does not claim. Several of the most widely repeated findings in this field do not hold up and are not used here. "Forty seconds of compassion reduces anxiety" came from healthy volunteers and cancer survivors watching a scripted video, and a 2024 trial in real oncology patients found no effect.

The finding that empathic doctors achieve better diabetes control rested on doctors rating their own empathy and was not replicated in a larger study in 2019. The claim that empathic practitioners shorten colds comes from subgroups of patients who rated their clinician perfectly, not from the randomised comparison. The honest evidence is enough without them.

5. And for the carer

Caring with compassion appears to protect the carer; caring by absorbing distress does not. In laboratory studies at the Max Planck Institute, training people in empathy — feeling another's pain — increased their own negative feelings, while training that followed in compassion — warm concern and the wish to help — increased positive feelings and reduced negative ones (Klimecki, Singer and colleagues, 2013–2014).

The studies were small and not in clinical settings, so this is plausible rather than proven, but it matters: it suggests the answer to burnout is not to care less, but to care differently.

Staff who attended Schwartz Rounds regularly saw psychological distress fall from 25% to 12%, against 37% to 34% among those who did not — a small, self-selected group, but a striking difference. And across the NHS, trusts with more engaged staff have lower patient mortality (West and Dawson, King's Fund, 2012).

The doctors-and-nurses annex describes the other side of this coin: moral injury, when staff are prevented from giving the care they know is right.

6. Across a life: older people and children

The evidence is firmest at the two ends of life, where people depend most on others. The strongest trials in this whole annex are about frail older people and about babies and young children. That is not a coincidence. The more dependent a person is, the more their health rests on the attention of someone else.

Older people

For frail older people in hospital, attentive, personal care prevents delirium. The Hospital Elder Life Program pairs staff and trained volunteers with patients over seventy to keep them oriented, moving, eating, drinking, sleeping and in company.

In its first study, of 852 patients, delirium fell from 15.0% to 9.9% (Inouye and colleagues, New England Journal of Medicine, 1999 — a matched study, not a randomised one).

Across 14 later studies and 3,605 patients, the odds of delirium were roughly halved and falls were reduced, at an estimated saving of $1,600–3,800 per patient (Hshieh and colleagues, American Journal of Geriatric Psychiatry, 2018).

It is a bundle of practical care, not warmth on its own. But its core is someone paying attention to the person in the bed.

In care homes, seeing the person rather than the diagnosis reduces distress — modestly, and without costing more. A randomised trial across 69 English care homes and 847 residents with dementia found that training staff in person-centred care improved quality of life and reduced agitation and other distressing symptoms, and cost less than usual care (Ballard and colleagues, PLOS Medicine, 2018). The effects were small, about a fifth to a third of a standard deviation.

An Australian trial found person-centred care reduced agitation against usual care, though falls rose in that arm (Chenoweth and colleagues, Lancet Neurology, 2009).

And an earlier English trial found that training and supporting staff in person-centred approaches cut antipsychotic prescribing from 42% to 23% of residents without making agitation worse (Fossey and colleagues, BMJ, 2006). Where antipsychotic use was already low, the 2018 trial found no further reduction.

Being cut off matters — but it is contact, more than the feeling of loneliness, that the evidence tracks. In the English Longitudinal Study of Ageing, about 6,500 people over fifty-two, social isolation was linked to a 26% higher risk of death after adjustment for health and circumstances; loneliness on its own was not (Steptoe and colleagues, PNAS, 2013).

For a care system the lesson is practical: the visit, the conversation, the familiar face are what count.

At the end of life, early palliative care improves how people live. In a randomised trial of 151 people with advanced lung cancer, those who also saw a palliative care team from the start had better quality of life, less depression (16% against 38%), less aggressive treatment at the end (33% against 54%) and a longer median survival — 11.6 months against 8.9 (Temel and colleagues, New England Journal of Medicine, 2010).

A Cochrane review of seven trials confirmed small gains in quality of life and symptoms but not the survival benefit (Haun and colleagues, 2017). Better lives, then; longer lives, not proven.

Some well-meant ideas have not shown results, and it is fair to say so. The one controlled evaluation of the Eden Alternative care-home model found no benefit, and more falls (Coleman and colleagues, Journal of Gerontology, 2002).

A large trial of open visiting in Brazilian intensive care units did not reduce delirium, though relatives were less anxious and depressed (Rosa and colleagues, JAMA, 2019). Befriending schemes produce a small reduction in depressive symptoms, from few studies (Mead and colleagues, British Journal of Psychiatry, 2010).

Dignity and family presence are right in themselves; claims that they improve health should wait for the evidence.

Children and young people

For the smallest babies, a parent's body saves lives. In a World Health Organization trial of 3,211 babies weighing between 1 and 1.8 kilograms, starting continuous skin-to-skin contact with the mother immediately after birth cut deaths in the first 28 days from 15.7% to 12.0% — a quarter fewer — and the trial was stopped early because the benefit was clear (New England Journal of Medicine, 2021).

A Cochrane review found kangaroo mother care reduced deaths by 40%, severe infection by about two-thirds and hypothermia by more than two-thirds (Conde-Agudelo and Díaz-Rossello, 2016). Part of the mechanism is warmth and breastfeeding. That is what human closeness is.

Treating parents as partners in neonatal care helps babies grow. In a trial across 26 neonatal units and 1,786 premature babies, bringing parents into the daily care of their baby improved weight gain, raised breastmilk feeding at discharge from 63% to 70%, and lowered parents' stress and anxiety (O'Brien and colleagues, Lancet Child & Adolescent Health, 2018).

The strongest evidence of all: a child needs a person, not an institution. In the Bucharest Early Intervention Project, 136 young children living in Romanian orphanages were randomly assigned either to stay or to move to foster families.

The children in foster care developed significantly better, with the largest gains for those placed youngest (Nelson and colleagues, Science, 2007), and the gains were still there at sixteen (Wade and colleagues, PNAS, 2022).

It is one of very few randomised tests of whether being cared for by someone who is yours changes a child's development. It does.

For children in care, stability matters. In a study of 729 American children in foster care, unstable placements raised the probability of behavioural problems from 22% to 36% (Rubin and colleagues, Pediatrics, 2007).

It is observational, but it points the same way as the care annex's account of children moved again and again between placements.

Connection protects as children grow, and its absence follows them. In a US survey of 12,118 adolescents, feeling connected to family and to school was protective against almost every risk behaviour studied (Resnick and colleagues, JAMA, 1997).

Adults who recalled six or seven positive childhood experiences — feeling safe at home, supported by family, cared for by at least two non-parent adults — had 72% lower odds of depression or poor mental health than those who recalled two or fewer (Bethell and colleagues, JAMA Pediatrics, 2019). Both studies are observational, and the second relies on memory.

In the clinic, the effects are real but small. Distraction reduces children's pain from needles (Birnie and colleagues, Cochrane, 2018, on low-certainty evidence), and in young people's mental health care the strength of the relationship with the therapist is modestly linked to how well they do (Karver and colleagues, Psychotherapy, 2018).

And here, too, not everything works. The Family Nurse Partnership, an intensive home-visiting programme for teenage mothers, showed no benefit on any of its main outcomes in an English trial of 1,645 mothers (Robling and colleagues, The Lancet, 2016).

And a parent's presence at a child's anaesthetic did not reduce the child's anxiety (Cochrane, 2015).

Letting parents stay with their children in hospital — the change that followed James and Joyce Robertson's films and the Platt Report of 1959 — was an ethical reform, not a clinical one, and right on those grounds.

What this means for the argument. The people with the least power to ask for humane care — the frail old person with dementia, the premature baby, the child in an institution — are exactly the ones for whom the evidence that it matters is strongest. They are also the people the care annex shows the system most often failing.

What it adds up to

No single study proves that humanity heals. Taken together, the evidence points one way. People live longer when their lives have meaning and when they are connected to others. They feel and do better when the person treating them is warm and attentive.

They are less likely to die, or to end up in hospital, when the same doctor has known them for years. The people who give care survive the work better when they give it with compassion rather than by absorbing distress. And none of this has been found to do harm, and almost none of it costs money.

So a system that crushes these qualities is working against its own outcomes. The rota that means no one knows the patient, the fifteen-minute visit, the target that leaves no time to talk, the culture that burns out the people who care — each of these removes something the evidence says is good for health, from the patient and from the carer.

Bevan's promise was not only a way of paying for care. It was, as the founding annex puts it, a way of treating people. The evidence is now catching up with him: the way people are treated is part of whether they get well. As Chris Turner put it, "the way we behave towards each other counts."

Loneliness, the silent killer

Loneliness is usually pictured as an old person's problem. The evidence now says young people report it most, and it is rising. It is linked to earlier death, heart disease, stroke, dementia and depression, and it touches patients, staff and the public alike.

The one thing to take away: authentic human connection is not a nicety but a health intervention, and a health and care system can either build it or destroy it every day.

Young people are now the loneliest generation

The group most at risk is not who most people think. In September 2026 the Institute for Public Policy Research reported that the share of young people with no close friends had risen from about one in 100 in 2014 to one in 20 in 2026 — five times as many — and that 47% of young women and 36% of young men aged 16 to 24 met the threshold for psychological distress.

"Kids aren't snowflakes," said its associate director Avnee Morjaria. "This is a rational response to a bleak world we have left for them." (IPPR's release does not name the survey behind the friendship figure; press reports say it is Understanding Society. The exact source and age band should be confirmed before the figure is quoted elsewhere.)

The official statistics point the same way. In England's Community and Engagement Survey for 2025/26, 7% of adults said they were lonely often or always — higher than the 5–6% recorded between 2013 and 2018 — and the rate was highest among 16–24-year-olds, at 9%, against 4% for those aged 65 to 74.

Worldwide, the WHO Commission on Social Connection (2025) found that about one person in six is lonely, rising to 17–21% of people aged 13 to 29, with teenagers highest of all; one adolescent in four is socially isolated. Across 142 countries, Meta and Gallup found 27% of 19–29-year-olds lonely against 17% of over-65s.

And in the United States, young people aged 15 to 24 spent about 150 minutes a day with friends in 2003 and about 40 in 2020 — a fall of nearly 70%, though the final year was a pandemic year. The OECD's summary is blunt: men and young people, "groups previously considered at lower risk", have seen some of the largest deteriorations.

One honest qualification. The surveys do not all agree on chronic loneliness. The Health Survey for England 2024 found 16–24-year-olds had the lowest rate of chronic loneliness, at 4% — but 29% of them were lonely at least some of the time, against 15% of 65–74-year-olds.

The fair summary is that young people consistently report the most loneliness overall, and on most official measures the most chronic loneliness too. The belief that loneliness belongs to old age is out of date.

Why it is happening is contested, and the document does not pretend otherwise. The World Happiness Report 2026 found that adolescents using social media for seven hours or more a day have clearly lower wellbeing than light users, with platforms built for connection associated with higher happiness and algorithmic feeds with lower — but asked whether social media explains everything, its answer was "Of course not."

Other researchers, writing in Nature in 2024, find a mix of no, small and mixed associations, and argue that poor mental health may drive heavy use as much as the reverse. What is not contested is the direction of travel.

The head of the official review agrees the rise is real. On 21 September 2026, ahead of publication, Professor Peter Fonagy — the clinical psychologist leading the independent review of mental health, ADHD and autism that Wes Streeting ordered as health secretary — told the BBC that the review will conclude the rise in mental distress since the 1990s is real, with about one in four 16–24-year-olds now struggling, and that it "dismisses the idea of a snowflake generation". "It's harder to be young now," he said.

He pointed to a rise in loneliness among young people since 2000, linking it to phones and social media cutting face-to-face contact and disrupting sleep, and blamed cuts to youth services for leaving fewer places where young people can mix and feel they belong.

The same report notes that the number of children and young people in contact with NHS mental health services in England has risen by 50% in five years, and that one in three people referred to mental health services are still waiting for help after a year; Fonagy called the service "not fit for purpose".

His remedy is a human one: "Saying to someone, 'I notice you're not quite the same as yesterday. Is there anything the matter?' Knowing there is someone who knows makes an enormous difference." (The review had not been published when this was written.

The figures are as reported by the BBC, which notes that the data are contested because some rely on self-reporting and greater awareness may mean more people coming forward — a point Fonagy accepts in part.)

Why it is called a silent killer

The health effects are consistent and large, though mostly observational. Across 90 studies and 2.2 million people, social isolation was linked to a 32% higher risk of early death and loneliness to 14% (Wang and colleagues, Nature Human Behaviour, 2023).

Earlier meta-analyses found 29% higher risk of coronary heart disease and 32% of stroke (Valtorta and colleagues, Heart, 2016), and a 2024 analysis of more than 600,000 people found 31% higher risk of dementia.

The WHO estimates that loneliness is linked to about 871,000 deaths a year — some 100 an hour — a modelled estimate rather than a count.

And the effects start young: across 63 studies of children and adolescents, loneliness predicted depression up to nine years later, with how long it lasted mattering more than how intense it was (Loades and colleagues, 2020). The WHO adds that lonely teenagers are 22% more likely to get lower grades.

What the evidence does not say. The familiar line that loneliness is "as bad as smoking 15 cigarettes a day" is a rough comparison drawn from all measures of social connection, not from loneliness alone, and its own author says so; this document does not use it.

Nor do these studies prove that loneliness causes the harm — illness can cause isolation as well as the other way round. The consistency of the findings across countries, ages and methods is what makes them hard to dismiss.

Patients: lonely people come to us, and we can make it worse

Loneliness brings people through the door. A review of 25 studies and more than 100,000 people found that lonely people make more visits to their GP (Sirois and Owens, 2021). In one study of 300 hospital inpatients aged over 65, around six in ten were lonely, and those who said they had no close relationships were 4.6 times more likely to die within a month.

A health service that sees the lonely more often than anyone has more chances than anyone to reconnect them.

But the main national answer has not yet worked. England's social prescribing link workers, rolled out across more than 1,200 primary care networks, brought small gains, but a 2025 national evaluation found they were "unable to support the NHS aim for patients to feel more connected, less lonely".

The lesson is not that connection cannot be built but that it cannot simply be referred on. The evidence in the annex on why being human is good for health points the same way: what helps frail older patients in hospital is people keeping them company, oriented and known; what helps everyone is being seen by the same clinician over time.

Staff: the carers are lonely too

Belonging is one of the three things staff most need, and the modern NHS works against it. The GMC's review Caring for doctors, caring for patients named belonging — "to be connected to, cared for, and caring of others" — as a core need, alongside autonomy and competence.

It described "ward hopping" creating "feelings of isolation, alienation and vulnerability", rotations that undermine doctors' "ability to be a member of any team", and the loss of the doctors' mess.

A review of 24 studies of nurses found loneliness closely tied to burnout. And resident doctors are mostly in their twenties — the age group reporting most loneliness — moved between hospitals and teams every few months. That combination is this document's inference, not a measured finding, but it deserves attention.

There is a measurement gap here worth closing. No UK survey measures loneliness among NHS or social care staff directly. The NHS Staff Survey asks 766,285 people a year how they are treated; it does not ask whether they feel alone. Adding the question would cost almost nothing.

The public, and the policy

Britain led the world on this, and then went quiet. In January 2018 the UK appointed the world's first minister for loneliness, and in October 2018 published what it called "the world's first government strategy for tackling loneliness", A connected society. Japan followed with its own minister in 2021.

But the government's loneliness pages have not been updated since June 2023, the last annual report was published in March 2023, and no current strategy or named minister could be found.

The benefit of authentic human connection

Connection is protective at every age. The largest meta-analysis found that people with stronger social relationships had 50% higher odds of survival over the study periods. The Harvard Study of Adult Development, which has followed its participants since 1938, reduces eighty years of findings to one sentence: "Good relationships keep us happier and healthier." Its director, Robert Waldinger, is blunter: "loneliness kills".

What works is real but modest, and it is mostly relationship rather than programme. A 2023 review for the UK government found a range of interventions effective at relieving loneliness in the short term, with the strongest results for psychological and social-skills approaches — but only seven studies included under-18s.

Evidence for befriending is inconclusive, and for Men's Sheds encouraging but weak. The more robust evidence is about the quality of ordinary encounters: warmth, continuity, time, being known.

That is why this belongs in a document about healthcare. The ward, the surgery, the care home and the team room are among the few places left where people meet face to face every day.

What this means for humanity-first healthcare

For patients: continuity, time to talk and family involvement are treatments for loneliness as well as for illness; a service that rotates staff and shortens visits is removing connection from the people who most need it.

For staff: belonging is a patient-safety measure. Stable teams, somewhere to meet, and leaders who know their people's names are the cheapest protection against both burnout and loneliness — and the evidence in Volume 6 says care reaches the patient only if it reaches the carer first.

For the public, and especially the young: the NHS is the largest employer in Europe and has contact with almost every family. If it chose to be a place of authentic connection — for the young people it employs, the young patients it treats and the lonely people who keep coming back — it would be doing, at scale, what the evidence says protects health.

That is what Tredegar's miners built in 1890: not only a doctor, but a community that would not let its neighbours face illness alone.

Joining up health and social care

England runs health and social care separately, and people get stuck in the gap. The government now wants a National Care Service "highly integrated" with the NHS. This annex asks what the evidence says about joining the two.

The one thing to take away: coordination around the person helps, slowly and modestly; merging the organisations does not, on its own, make anything better — and Northern Ireland, which merged them in 1973, is the clearest proof.

Why it matters now

The direction of travel is towards integration. In a speech on social care on 29 July 2026 the Prime Minister, Andy Burnham, asked Baroness Casey to bring her commission forward by a year and set out the aim: "How do we deliver a national care service in England that sits alongside the National Health Service, works in a similar way that is highly integrated?"

He added that it "has to be very integrated with the rest of the system. It has to be preventative in its character", with "a principle of home care wherever possible".

The idea is an old one for him. His 2010 white paper promised "a comprehensive National Care Service" and "a new duty for NHS bodies and local authorities to deliver integrated care"; in 2013 he argued for "One person. Three care services" to become "one budget, one service co-ordinating all of one person's needs".

The 2026 speech proposes a service alongside the NHS rather than a merger, and names no funding model and no free-care pledge. So the question this annex answers is live: what does joining up actually achieve?

Northern Ireland: fifty years of one system

On waiting times Northern Ireland is the worst-performing of the four UK nations, by a wide margin.

• In 2019 the Nuffield Trust found that a person in Northern Ireland was "more than 3,000 times as likely as a citizen of England to have been waiting more than a year for healthcare", and nearly 50 times as likely as someone in Wales.
• The Northern Ireland Audit Office found that in December 2022, 26.3% of the population was on a waiting list, against 12.4% in England, and that 50.9% of those waiting had waited more than a year, against 5.4% in England. It called this "extremely poor value for money".
• At 31 March 2026, 510,680 people were waiting for a first outpatient appointment, and 54.1% of them had waited more than a year.
• In January to March 2026, 30.5% of emergency department patients were seen within four hours and 23.5% waited more than twelve — in the Royal College of Emergency Medicine's words, the worst quarterly figures ever recorded there. "It's chaos in our departments," its Northern Ireland lead said.

One caution on comparisons. The four nations use different targets and definitions, and Northern Ireland published no comparable data for part of 2023–24 while it installed a new patient record system.

That affects precise league tables, not the overall picture, which every independent source shares.

No one presents it as a model — the analysts say the opposite. The Nuffield Trust found integration there "often described as statutory rather than delivered in practice", with "health continues to dominate", and warned that "combining health and social care within the same organisation does not necessarily result in social care being accorded a higher priority".

One senior figure told its researchers that social care had its "pocket picked" to support hospitals. The care emergency annex sets out what followed: home-care packages for older people cut from 26.2 to 20.6 per 1,000 people over sixty between 2014 and 2024.

It is fair to record the positives. Multidisciplinary teams in primary care — physiotherapists and mental health staff working alongside GPs — have been piloted and spread; pharmacists and optometrists have taken on work from hospitals.

And the Department of Health reported in June 2026 that the longest waits had begun to fall, with four-year outpatient waits down 54%, backed by £80m ($108m) of ringfenced funding. Those gains came from targeted effort and money, not from the structure, which has been in place for fifty years.

Where joining up has helped

Torbay: the most cited English example. Torbay created a single care trust in 2005. By 2009/10 the average number of occupied hospital beds had fallen from 750 to 502 over a decade; emergency bed days for people over 65 were 1,920 per 1,000, against a regional average of 2,698; delayed transfers of care fell to "negligible" levels; and more people were supported at home rather than in care homes.

The King's Fund's own caution matters: "establishing cause and effect is difficult", and a small area with a long culture of joint working came before the merger. Torbay is best read as people who already worked together being given a structure that let them — not a structure that made them.

Greater Manchester: joined-up public services, modest but real gains. After health and social care were devolved in 2016, life expectancy rose 0.2 years more than expected, and by 0.39 years in the most deprived areas (Lancet Public Health, 2022).

Alcohol-related admissions fell 11.1% and satisfaction with adult social care rose 17.6% — but A&E waits lengthened 12.2% (Social Science & Medicine, 2024). The researchers attribute the gains to a combination of devolution, wider public-service powers and a new mayor, not to integration alone.

Scotland: early promise, then decline. Audit Scotland found local gains in the first years of integration joint boards — Angus freed an average of 37 hospital beds a day, and a Red Cross home-from-hospital service in East Ayrshire saved 1,730 bed days — but by 2024 it reported "a general decline in performance of services and outcomes", a 9% real-terms fall in funding, and no evidence of the intended shift from hospital to community.

England's national schemes: slow and small. The National Audit Office found in 2017 "no compelling evidence to show that integration in England leads to sustainable financial savings or reduced hospital activity": under the Better Care Fund, emergency admissions rose by 87,000 when a fall of 106,000 had been planned.

The Health Foundation found that integrated care programmes and multidisciplinary teams took between three and six years to reduce emergency admissions, if they did at all — Mid-Nottinghamshire's "Better Together" programme saw reductions only in years five and six.

What the systematic reviews say. Joined-up care "may lead to increased patient satisfaction", "may increase perceived quality of care" and "may increase patient access to services", but evidence that it reduces hospital use or cost is limited and inconsistent (Baxter and colleagues, 2018).

For older people it "tends to have a positive impact on hospital admission rates", with no effect on mortality (Liljas and colleagues, 2019).

What it adds up to

Coordination helps; merger on its own does not. Everything that has worked shares three features: people from health and care working around the same person, the same faces over time, and years of patience.

Leeds and Manchester's children's services, set out in Annex 5 · The care emergency: social and children's care, show the same thing: they turned round inside separate structures by keeping the same people around the same children.

Northern Ireland shows the reverse — one structure for fifty years, and the money still flows to hospitals while home care is cut.

So a National Care Service will succeed or fail on what it does, not how it is organised. If it protects the care worker's time with the person, joins up the people around the patient, and holds leaders to account for how people are treated, the evidence says it can help.

If it is another reorganisation of the boxes, Northern Ireland has already shown the result. That is the argument of this whole document, applied to its most important live decision.

Why we all love the Welsh

The NHS did not begin in 1948, or even in Tredegar in 1890. It began with a relay that ran for two centuries through one small nation: a parson who taught the poor to read, chapels where working people learned to govern themselves, miners who paid for their own libraries and doctors, and a miner's son who took all of it to Westminster.

At almost every handover, the person carrying the idea was called a nuisance by the people in charge. The one thing to take away: humanity compounds. A few people choose to act for the good of others, they connect with a few more, and what they create becomes the ground the next generation stands on. That is the Champions Pathway — choice, connection, creation — and Wales is its clearest proof.

This annex is a lens, not a proof. Other places had chapels, coal and poverty without producing a Bevan. What follows is the author's reading of a well-documented history, and every fact in it has been checked; where a popular story could not be traced, it has been left out.

The first ripple: a parson who taught a nation to read

It began with a shepherd boy who became a parson. Griffith Jones was born in 1683 at Penboyr in Carmarthenshire, worked as a shepherd as a boy, and in 1716 became rector of Llanddowror. He found that preaching to people who could not read left nothing behind once the sermon ended.

So he took the school to the people. From 1731 his "circulating schools" sent a teacher to a parish for about three months, usually in winter when farm work was slack, with night classes for adults who worked by day, and taught in Welsh, the language people actually spoke.

By his death in 1761, 3,495 schools had been held and more than 158,000 scholars taught in twenty-five years; one Cardiff historian estimates that over 200,000 people learned to read, in a country of about 480,000.

In 1764 a commissioner acting for Catherine II of Russia reported on the schools. Wales became, in that historian's words, "one of the most literate countries in the modern world".

He did not do it alone. His patron, Bridget Bevan — no relation to Aneurin — funded and defended the schools, carried them on after his death until her own in 1779, and left £10,000, a fortune at the time, to keep them going.

It is the first "power of two" in this story: one person with the idea, another who made sure it outlived him.

And he was a nuisance. Between 1745 and 1752 five pamphlets attacked Jones and his schools, largely for his links with the new Methodist preachers. He has been called "the morning star of the Methodist revival in Wales". The poor he taught to read would become the congregations that built Welsh Nonconformity.

Connection: chapels that taught people to govern themselves

Reading met a revival. On 7 March 1739 George Whitefield, the most famous preacher of his age, met the Welsh evangelist Howell Harris in Cardiff for the first time, and in 1743 the English and Welsh Methodists held their first joint association at Watford, near Caerphilly, with Whitefield as moderator.

Across the next century Wales left the established church for its own chapels: at the 1851 religious census, 80% of people in Wales who attended a place of worship went to a Nonconformist chapel.

The chapels were schools for democracy. In many of them, especially the Baptists and Independents, the congregation chose its minister, raised its own money, kept its own accounts and argued its business out in meetings.

Working people were chairing committees, managing budgets and voting on their leaders generations before most of them could vote for Parliament.

Then the state called them backward, and they answered. In 1847 three commissioners, Anglicans who spoke no Welsh, reported on education in Wales. "The Welsh language is a vast drawback to Wales," they wrote, "and a manifold barrier to the moral progress and commercial prosperity of the people"; they blamed Nonconformity and cast Welsh women as immoral.

The reports caused a furore and became known as Brad y Llyfrau Gleision, the Treachery of the Blue Books. The chapels, the Welsh press and a generation of writers answered in print. A people who had been taught to read now had something to prove.

Even the name tells the story. "Welsh" comes from the Old English wealh: a foreigner, a stranger, a Briton — on the Britons' own island.

The Welsh name for themselves, Cymry, comes from an old British word meaning fellow-countrymen. The settlers named them by difference; they named themselves by what they shared.

The annex that follows, Annex 11 · Why a top-down model fails, tells what the commissioners of 1847 represented: a way of governing people from above that would do lasting harm far beyond Wales.

Creation: the valleys build their own

When coal came, a reading, arguing, self-organising people poured into the valleys — and built for themselves. Miners paid for their own workmen's institutes and libraries, usually through a deduction of a penny or two in the pound from their wages.

By the Second World War the Tredegar library was circulating 100,000 volumes a year, and the committee that chose its books was headed by Aneurin Bevan.

They built their own health service too. In 1890 four local funds merged to become the Tredegar Workmen's Medical Aid Society. Members paid three pence in the pound, and in return the town had its own doctors, a cottage hospital from 1904 and a central surgery from 1911, covering some 20,000 people.

Bevan joined its hospital management committee in 1924 and chaired it in 1929–30. Nobody gave Tredegar its health care; its people organised it, paid for it and ran it.

Outsiders helped most when they worked alongside. After the 1926 strike, Quakers — the same tradition as Elizabeth Fry — opened Maes-yr-Haf at Trealaw in the Rhondda in 1927, an educational settlement for unemployed miners and their families.

It was run on Quaker lines: not charity handed down, but classes, workshops and practical help shared.

The Bevan household: everything the relay had built, under one roof

He was born into all of it. Aneurin Bevan was born on 15 November 1897 at 32 Charles Street, Tredegar, the sixth of ten children; only eight survived infancy and six reached adulthood.

His father, David, was a miner and a Baptist, "fond of books and music", who was won to socialism by Robert Blatchford's paper The Clarion. His mother, Phoebe, a Methodist, was the daughter of a local blacksmith.

Two strands of Welsh Nonconformity, a love of reading and a radical newspaper: the whole inheritance sat around one kitchen table.

He grew up in the afterglow of a revival. In November 1904 Evan Roberts, a 26-year-old former collier from Loughor, set off the Welsh Revival. It ran until early 1906; around 100,000 converts were claimed, with estimates from 80,000 to 162,000, and in Glamorgan convictions for drunkenness fell by almost half between 1902 and 1907.

Then it burned out. Roberts broke down and withdrew from public life.

Bevan, seven when it began, grew up in a Wales that had felt what collective conviction could do — and seen how quickly it faded when nothing was built to hold it. That, on this annex's reading, is the lesson his generation drew: fervour is not enough; it has to be built into institutions.

He had to fight for his voice. Bevan left school at thirteen for the pit and had a severe stammer, which he worked to master; the library, the lodge and the council chamber became his university. He won a miners' scholarship to the Central Labour College in London in 1919, and was elected to Tredegar council in 1922.

Then came the moment that made it personal. In February 1925 his father died of pneumoconiosis, the coal-dust disease of miners' lungs — in Aneurin's arms, as he later recalled. Because the disease was not then recognised under the compensation law, the family received nothing. The community had given him the answer; the law had given him the anger.

The squalid nuisance

In the war, Bevan refused to stay quiet. While the parties governed in coalition, he became the sharpest critic of Churchill's conduct of the war. In the censure debate of 2 July 1942 he told the Commons: "The Prime Minister wins Debate after Debate and loses battle after battle. The country is beginning to say that he fights Debates like a war and the war like a Debate."

Churchill won the vote by 476 to 25, but the diarist Harold Nicolson recorded that Bevan had opened "with a brilliant offensive" and that "the impression left is one of dissatisfaction and anxiety".

Not all Bevan's wartime demands were wise; his insistence that Parliament keep asking questions in wartime was.

Churchill gave him the title. On 6 December 1945, with Bevan now Minister of Health, Churchill warned the House that unless he changed course "he will be as great a curse to this country in time of peace, as he was a squalid nuisance in time of war".

The insults ran both ways: in July 1948 Bevan called the Tories "lower than vermin", and Churchill replied that he might better be called the "Minister of Disease". Bevan's temper could cost him.

But his opponent knew what he was. When Bevan resigned from the government in 1951 over charges for NHS teeth and spectacles, Churchill found Jennie Lee in the Commons smoking room and told her, as she later recorded his lisp: "Do not undereshtimate your husband."

After Bevan's death in 1960 she wrote: "He was a great humanist whose religion lay in loving his fellow men and trying to serve them."

In this story, "nuisance" is a badge of honour. Griffith Jones was a nuisance to the bishops, the Welsh chapels were a nuisance to the commissioners of 1847, and Bevan was a nuisance to Churchill. Each was a nuisance to an establishment that thought things were fine as they were — and each left Britain better for it.

Ripples through the generations: the Champions Pathway at work

Each generation chose, connected and created — and handed the next a little more. Read as a relay, two hundred years of Welsh history follow the Champions Pathway almost step for step:

The ripples compound. None of these people could have done what they did without the generation before. Bevan did not invent the idea that neighbours should share the cost of illness; he inherited it, argued for it in the voice the chapels had taught him, and scaled it up.

"All I am doing," he is reported to have said, "is extending to the entire population of Britain the benefits we had in Tredegar."

Seventy-eight years later, some 1.5 million people in England go to work in the NHS every day. Every one of them stands on ground first laid in a winter classroom in eighteenth-century Carmarthenshire.

What this means for healthcare now. Three lessons run through this story and through this document.

First, humanity compounds in both directions: habits of care, like habits of neglect, are handed down.

Second, every generation that succeeded built its values into something that would outlast it — a school, a bequest, a chapel constitution, a poundage, an Act of Parliament. The one thing the NHS never built in was the humanity itself.

Third, the nuisances matter. The consultants at Chester, the families at Durham and the whistleblowers in every inquiry in this document are today's nuisances. A health service that remembers where it came from would thank them.

That is why we all love the Welsh. A small nation, called foreigners on their own island and backward by their own government, taught itself to read, to argue and to look after its own — and then gave that to all of Britain.

The English, it seems, have noticed: in a 2022 survey by Golley Slater and Beaufort Research, 55% of English people described the Welsh as welcoming, against 38% who said the same of the English.

As Bevan understood, it was never about being Welsh. It was about being human, and building it in.

Why a top-down model fails

Wales shows what happens when people build care for themselves. This annex tells the opposite story, from the same century and the same island: what happens when a confident elite decides what is good for people it neither knows nor listens to, and then begins to rank them by worth.

It started in benevolence, hardened into control, and was bent by the racial thinking of the age into something that dehumanised the Welsh, the Irish, Jewish people, colonial subjects and, in the end, disabled people in Britain's own hospitals.

The one thing to take away: a system that decides about people instead of with them will, sooner or later, stop seeing them as people at all — and the consequences are disastrous.

This is a warning, told fairly. Not every top-down reform failed; some of the greatest public health gains in history were imposed from the centre, and the last section of this annex says so. The failure described here is narrower and more dangerous: power over people without relationship with them.

Every fact has been checked; claims from the source material that could not be supported have been corrected or left out.

The benevolent founders

It began with people doing great good. The late eighteenth and early nineteenth centuries were an age of moral energy in Britain. Evangelicals of the Clapham Sect, with William Wilberforce at their head, led the campaign that abolished the British slave trade in 1807.

Quaker industrialists such as the Darbys at Coalbrookdale, where Abraham Darby first smelted iron with coke in 1709, and later the Cadburys at Bournville, showed that business could be run on conscience.

But the same reformers were sure they knew best. Frightened by the French Revolution, many feared any independent organising by working people. Wilberforce called trade unions "a general disease in our society" and favoured a Combination Act against them in 1799; he defended the sedition laws of 1795, voted to suspend habeas corpus in 1817, "most unwillingly", and approved the repressive Six Acts of 1819 "as a temporary necessity".

To this way of thinking, the poor needed improving, not a voice: moral guidance from above, not a share in decisions.

That is the seed. Benevolence that listens builds a Tredegar. Benevolence that decides for people, however sincere, builds a machine that the next generation inherits — with the authority to intervene, and without the humility that sometimes checked the founders.

The inspectors: Wales, 1847

The inquiry was born of fear. On 10 March 1846 William Williams, MP for Coventry, asked Parliament for an inquiry into education in Wales, "especially into the means afforded to the labouring classes of acquiring a knowledge of the English tongue", citing "the Newport and Rebecca riots".

The unrest of a poor, chapel-going, Welsh-speaking population was read not as a protest about poverty but as a threat to order.

The men sent to judge Wales did not know it. The three commissioners were Anglican laymen, two of them young barristers, and none spoke Welsh: Ralph Lingen, 26, a Balliol fellow, who covered Carmarthen, Glamorgan and Pembroke; Jelinger Symons, 36, a Cambridge graduate who had worked on earlier government inquiries, who covered Brecknock, Cardigan and Radnor; and Henry Vaughan Johnson, who covered North Wales.

They worked through Welsh-speaking assistants, and leaned heavily on local gentry and Anglican clergy, while the testimony of the Nonconformist majority counted for less.

They judged a whole people. Their reports, submitted in July 1847, called the Welsh language "a vast drawback to Wales, and a manifold barrier to the moral progress and commercial prosperity of the people", blamed the chapels and cast Welsh women as immoral.

The backlash was fierce and lasting. In 1850 Ieuan Gwynedd launched Y Gymraes, a magazine for Welsh women, whose opening pages, by the poet Elen Egryn, urged them to rise "above shame and hateful mockery".

The system rewarded the inspectors. Lingen went on to run the Treasury as its Permanent Secretary from 1869 to 1885 and was made a peer. Nothing in the method — outsiders who did not speak the people's language, relying on the establishment's witnesses, passing judgment on a community — counted against him.

One popular belief needs correcting. The "Welsh Not", the board hung round the neck of a child caught speaking Welsh, did not come from the Blue Books. It was in use in some schools from at least the 1790s, it was a local practice rather than state policy, and it was unusual by the 1880s.

The historian Martin Johnes concludes it arose from misunderstanding about how children learn and from parents' own desire for English, and that it hindered the very learning it was meant to achieve.

It is still a painful memory in Wales, and part of the same attitude: that a people's own language was something to be shamed out of them.

From improving people to ranking them

The same decades turned cultural contempt into a hierarchy of peoples. In 1835 Thomas Babington Macaulay's Minute on Indian Education declared that "a single shelf of a good European library" was worth the whole native literature of India and Arabia.

In 1850 the anatomist Robert Knox, in The Races of Men, described the Celt as given to "furious fanaticism; a love of war and disorder; a hatred for order and patient industry" and claimed the Celtic race "never could be made to comprehend the meaning of the word liberty".

The Welsh, the Irish and the Highland Scots were being written into a ranking of human worth with the Anglo-Saxon at the top.

In Ireland, the ranking cost lives. During the Great Famine of 1845–52 about a million people died and more than a million emigrated, while relief was run on strict free-market principles by the Treasury official Charles Trevelyan, who could write that if small farmers left "we shall at last arrive at something like a satisfactory settlement of the country".

Jewish people were kept outside the nation. In 1847 Lionel de Rothschild was elected MP for the City of London but could not take his seat without a Christian oath; the Commons voted to let him, the Lords refused, and it took until 1858 for a Jewish MP to sit.

Half a century later the Aliens Act of 1905, Britain's first modern immigration control, was aimed largely at Jewish refugees fleeing persecution in Eastern Europe. (There is no evidence that the 1847 commissioners themselves held anti-Jewish views; the point is the shared assumption of the age, that belonging depended on fitting one mould.)

Then the ranking claimed the authority of science. In 1883 Francis Galton coined the word "eugenics"; in 1907 the Eugenics Education Society was founded in London.

As Home Secretary, Winston Churchill wrote to the Prime Minister in December 1910 that the "Feeble-Minded and Insane classes" were "a national and race danger which it is impossible to exaggerate", and supported their segregation.

The idea that some people were a danger to the race had moved from the lecture hall into government.

Where it led in health and care

It was written into law. The Mental Deficiency Act of 1913 provided for people judged "mentally defective" to be segregated in institutions; Parliament rejected compulsory sterilisation, but not segregation.

The numbers held in such institutions are estimated to have grown from about 10,000 to about 60,000 between 1920 and 1946. People who did not fit the mould were no longer to be improved; they were to be removed from sight.

The NHS inherited those institutions, and their culture. In 1948 the new health service took over the long-stay hospitals. One of them was Ely Hospital in Cardiff, which had begun in 1862 as a Poor Law school for orphaned children and became a workhouse before it became a hospital for people with learning disabilities.

In 1967 a newspaper published allegations of ill-treatment there; the inquiry chaired by Geoffrey Howe reported in March 1969, one of the first great scandals of the NHS, and led to the 1971 white paper Better Services for the Mentally Handicapped.

And the pattern did not end with the long-stay hospitals. At Winterbourne View, a private hospital near Bristol, abuse of people with learning disabilities was exposed in 2011, and eleven staff pleaded guilty.

At Whorlton Hall in County Durham, abuse was exposed in 2019, and four support workers were found guilty of ill-treatment in 2023.

The annex on the care emergency sets out how far this reaches today. The thread from 1913 is not a conspiracy; it is a habit of mind — that some people are cases to be managed rather than people to be known — and it has proved very hard to kill.

Medicine was not immune. The Lancet Commission on Medicine, Nazism, and the Holocaust found that health professionals' part in Nazi crimes reflected dangers "inherent to modern, scientific medicine". Eugenics was an international movement, and Britain was one of its homes.

Volume 6 sets out why medicine, of all professions, needs to be formed in humanity and not only in skill.

The honest qualification: top-down is not always wrong

Some of the greatest gains in health were imposed from the centre. Edwin Chadwick's Sanitary Report of 1842 led to the Public Health Act of 1848.

After the "Great Stink" of summer 1858, Joseph Bazalgette built London's sewers — 82 miles of main sewers and 1,100 miles of street drains, opened from 1865 and completed in 1875. The NHS itself was created by an Act of Parliament and run from Whitehall.

So the dividing line is not top-down against bottom-up. It is whether power is used for people, as people, or on them, as problems. Bazalgette's sewers served everyone alike and asked nothing of anyone's worth. The NHS scaled up what Tredegar's miners had already built and chosen for themselves.

The 1847 commissioners, the eugenicists and the long-stay institutions did the opposite: they decided about people without them, measured them against a mould, and treated those who did not fit as less than human.

That is the top-down model this annex means, and it fails — sometimes slowly, sometimes catastrophically.

What it means for healthcare now

The method of 1847 is still with us. Outsiders who did not speak the people's language, relying on the testimony of those in charge, judged a community from above.

At County Durham, the Aubrey review found that the board "accepted reassurance rather than seeking genuine assurance" while patients were harmed. At Chester, the doctors who raised the alarm were treated as the problem.

The families in this document were so often the last to be listened to. Each is the same failure in a modern form: judgment passed on people without relationship with them.

Dehumanising starts small. Volume 5 describes how ordinary people come to harm others step by step, along a continuum: labels, distance, the sense that some people count less.

The Victorian ranking of peoples did not begin with eugenics; it began with contempt for a language and a chapel. The lesson for any health service is to notice the first steps — the shorthand that turns a patient into a case, the joke about a group of staff, the assumption that a family "doesn't understand".

The antidote is simple humanity, built in. Wales is the counter-story: people who were looked down on built schools, chapels, libraries and a medical aid society for themselves, and that is what became the NHS.

Humanity-first healthcare asks the service to govern that way — every person counts; caring for the carers; it's who we are — so that power is always exercised with people, and never simply over them.

Humanity-First Healthcare: applying lessons from the past

In 2026 one of the largest NHS trusts in England was found by its regulator to have a "fundamental failure in quality governance".

A decade earlier its flagship hospital, Salford Royal, was among the most admired in the country. Same buildings, same city, many of the same staff.

The one thing to take away: excellence that rests on one leader is not excellence, it is borrowed time — what endures is what is built into the fabric of an organisation, and the last four hundred years show how that is done.

Three pillars that do not depend on who is in charge

Humanity-first healthcare is the art of enablership: making people able, and building a capability that outlasts the people who built it.

If excellence that hangs on individuals is borrowed time, then the design question is not "who should lead?" but "what will still be true when this leader has gone?"

Three pillars answer it. The specific form of governance, management and leadership will vary from place to place; the three principles do not.

Read it from the top down. The bar across the top is the test, and it is the only one that matters here: what will still be true when this leader has gone?

The three columns are what holds it up — the individual, the community, and the leadership that answers for both.

The band underneath is what all three cost: no reorganisation, no new legislation, no heroic appointment.

Each column is taken in turn below.

The first pillar is the individual: the Champions Pathway. It is literacy in what it means to be human — choice, connection and creation — and it rests on the claim that we choose what we become.

None of us chooses the circumstances, the rota or the pressure.

Under pressure a person still chooses whether to see the patient in front of them as a human being.

That capacity can be taught, practised and grown, and at its best it becomes instinctive — the fourth of the Champions Four Stages of Humanity, set out in Volume 10.

The second pillar is the community: Champions Circles. Small, self-governing groups, closer to a Welsh chapel or a YPO forum than to a committee: a handful of people who meet regularly, in confidence, and who know each other well enough to tell the truth.

They are the social glue that a rota-driven hospital has lost, and they cost almost nothing.

The decisive feature is that the network sits beyond and independent of the employer.

A circle that answers to the organisation it is meant to challenge is a committee. A circle that belongs to an external network can do both of the things that matter: call it forward, when something is being done well and nobody outside has noticed, and call it out, when something is wrong and the internal route has failed.

The evidence says this is the missing piece rather than a nicety.

There were 1,519 Freedom to Speak Up guardians across 820 organisations in March 2026, but as the Thirlwall Inquiry records, "until very recently, all FTSU Guardians were appointed by hospital Trusts from within their existing staff", most doing the job alongside another one, and the National Guardian's Office had "limited resources (16 staff) and limited powers" — "neither the power nor resources to hold hospitals or other healthcare settings to account".

In the 2025 NHS Staff Survey the speaking-up score fell to a five-year low, and only 55.49% of staff were confident their organisation would address concerns about clinical safety.

The National Guardian's Office closed on 30 June 2026 and its remaining functions moved into NHS England.

Thirlwall recommended the opposite direction of travel: that they go to the Parliamentary and Health Service Ombudsman, with the power to investigate complaints that whistleblowing had not been dealt with properly.

The third pillar is accountable leadership: trust, forgiveness and being held to account. All three, or none of them work.

• Trust, so that people tell the truth early.
• Forgiveness, so that an honest mistake does not end a career and therefore does not have to be hidden.
• Being held to account, so that forgiveness is not mistaken for indifference.

The Thirlwall Inquiry put the failure of the last twenty-five years starkly: "The move towards a 'no blame' culture in the NHS, starting in 2000, was a mistake.

The focus on system faults rather than the failings of individuals, coupled with a learning culture that omitted acknowledging responsibility or blame, meant difficult conversations about conduct were avoided.

This undermined patient safety." Her sharpest line is the one to keep: "Instead of being safe, patients are at greater risk of harm in a 'no blame' culture."

And her distinction is the one this pillar rests on — "whilst blame is always negative, responsibility is not".

She is also sceptical of the cure that is now fashionable, and this document takes the warning seriously.

Of "just culture" she writes: "I question the value of identifying a desirable culture and then adding a single adjective in front of the word 'culture' with the intention and hope of influencing people to behave in accordance with that description … It is behaviours that need to change."

Naming a culture is not building one. The annex on radical leadership sets out the mechanisms that do the building, drawn from aviation, from DuPont and from the NHS's own patient-safety rules.

The official answer so far is narrower than the inquiry asked for. The UK government has confirmed a barring scheme run by the Health and Care Professions Council, covering board-level directors and their direct reports, to be legislated when parliamentary time allows.

Thirlwall asked for a barring system for all managers, clinical and non-clinical, by September 2027, and for an individual duty of candour on every manager.

Either way, barring removes a person after the event. It does not build the capability that would have prevented it.

The rest of this annex is the evidence for why those three, and not a fourth reorganisation. What follows is a case still unfolding in Greater Manchester, where a hospital that had everything lost much of it within a few years of the person who held it leaving — and then four hundred years of institutions that wrote their values into the fabric and kept them.

This annex is written from the record as it stands on 25 September 2026. There is no public inquiry into the Northern Care Alliance. What exists is regulatory action, external review and investigative journalism, and where something is an allegation it is described as one.

That is what makes the case useful: this is what the early stage looks like, years before an inquiry report.

What has happened in Greater Manchester

A trust that inherited an Outstanding rating is now under a regulator's warning notice. The Care Quality Commission served a Section 29A warning notice on the Northern Care Alliance NHS Foundation Trust on 13 August 2026, "due to significant concerns about the trust's management of risk, patient safety, complaints, freedom to speak up processes and workforce wellbeing and inclusion".

It covered the whole trust, not one service.

NHS England had already acted. In June 2026 it took enforcement action after finding the trust "currently unable to provide assurance … that the organisation has a clear and consistent quality governance structure" and that it "does not have effective systems and processes to identify and address risks affecting the quality and safety of the surgical wards".

A review by the Good Governance Institute, commissioned by the trust, ran to 43 recommendations. The Department of Health and Social Care said patients in Salford had been "severely let down".

None of this was sudden. The trust had already been placed in segment 4 — the lowest — of the NHS Oversight Framework, and reported in December 2025 that it was "under close review" against 22 indicators.

The staff had been saying so for a year. A survey of 400 consultants across the trust, taken in November 2025 and later leaked to Channel 4 News and The Independent, found that nearly 81% said patient safety was worse than in pre-Covid times and 75% did not believe the trust was well led.

In January 2026 doctors threatened a vote of no confidence in the chief executive and the chief medical officer unless their concerns were addressed within three months.

The allegations that followed are clinicians' accounts, not regulator findings, and this document treats them as such. Staff described the longest wait for a bed as "four days on a corridor", and the conditions as "unsafe and … inhumane".

On the handling of safety incidents, Channel 4 News reported: "During this investigation, we have repeatedly been told of management downgrading them, asking that they be closed, or closing them with no action."

The trust replied that it has "clear processes to investigate incidents and learning from these is an area we need to prioritise".

The harm has patients attached to it. In May 2026 it was reported that more than 60 patients had been harmed by delays in a gynaecology service at Salford Royal, after a trust audit of hundreds of cases linked to one consultant, with some cancer diagnoses delayed.

The trust said: "We're truly sorry for any distress we've caused."

A further independent external review of gynaecology surgery was announced in August 2026. A consultant gynaecologist who resigned in April described a case in which shortened appointment times contributed to a missed cancer and the patient died within months — "probably a preventable death".

And the workforce was in dispute while all of it was going on. More than 350 operating theatre staff — nurses, clinical support workers and housekeepers — struck across the trust's four hospitals in August 2026 over how contractual overtime was paid, with some losing more than £80 ($108) a shift.

The trust declared a critical incident on 24 August so that urgent surgery could go ahead. A walkout planned for 14 September was suspended for fresh talks.

The region noticed before the regulators did.

A group of former staff emailed the mayor's office in July 2024, asking that the trust's membership of the Greater Manchester Good Employment Charter be revoked, and writing that "trust-wide systems which permitted so many patients to be harmed for decades within the current NCA are probably unlikely to be due to an individual 'rotten apple' or confined to just one department".

The trust no longer holds member status of the Charter. Greater Manchester has had devolved health powers since 2016, but a foundation trust answers to national regulators, so local leaders could apply pressure and not much else.

The merger that was meant to be a rescue

The trust was built to spread good practice, and it spread the pressure instead. In 2016 the CQC rated Pennine Acute Hospitals NHS Trust inadequate, and Salford Royal was asked to take over its leadership.

On 1 October 2021 Pennine Acute was dissolved and its staff, property, services and liabilities were transferred to Salford Royal, which was renamed the Northern Care Alliance.

The trust now has around 20,000 staff and income of about £2.0bn ($2.7bn), on which it ran a deficit of £70.8m ($95.6m) in 2024/25, and it serves more than a million people across Salford, Oldham, Bury and Rochdale.

Salford Royal was rated Outstanding by the CQC in March 2015 and again in August 2018.

Salford Royal Hospital was rated Requires improvement in December 2022, and again in February 2026.

The evidence on mergers was discouraging before this one was made. Studying a decade of English hospital mergers, Gaynor, Laudicella and Propper concluded that "there seems to be little hard evidence that this attempt at government planning of hospital care has achieved much more than simply reducing hospital admissions".

In December 2014 the Health Foundation found that "mergers have occurred for decades in the NHS, although the evidence, such as it is, suggests the benefits they set out to achieve tend not to be realised".

The newer fashion for hospital groups is no better evidenced.

Nearly a quarter of NHS trusts have adopted or committed to a group model; reviewing them in August 2025 the Health Foundation found the evidence base "underdeveloped", that "cultures have proven harder to change than structures", and summed the position up in six words: "We're in the realm of belief, rather than evidence."

It is worth being accurate about what the Dalton Review actually said, because it is often quoted the wrong way round. Published in December 2014, it set out federations, joint ventures, service chains, management contracts and foundation groups, and it said that for one category of struggling trusts "an acquisition by another organisation is likely to be the best route to sustainability".

Its warning was not against takeovers but against uniformity: "there should be no national blueprint or one size fits all", and "there are no 'right' or 'wrong' organisational forms — what matters is what works".

What Greater Manchester ended up with was a full statutory transfer at the largest scale, which is one of the forms, imposed in the one way the review cautioned against.

The hero problem

Salford Royal's excellence was real, and it was carried by people rather than built into the institution. Sir David Dalton led the hospital for 18 years, retiring on 31 March 2019, and from April 2016 led Pennine Acute as well.

The CQC's 2015 inspection recorded what the place felt like: "Information on performance was prominently displayed in the units for all to see. This included staffing levels and information on incidences of harm such as falls, the number of new pressure ulcers and the number of catheter associated urinary tract infections."

One unit's board read "56 days since the last fall and 500 days since a patient had developed a hospital acquired pressure ulcer". Over three years the trust cut pressure ulcers by 73%.

The style was deliberate, and it was not a tour of inspection.

The same report notes that "the executive team 'worked with' staff at the frontline rather than 'walk about'", giving the example of the chief executive shadowing a community nurse. In fairness, it also records that on some units most staff "could not recall seeing a board member", so even there the practice was uneven.

Copying the ritual without the purpose makes it worse than useless. Across 16,797 returned surveys, work settings reporting leadership walkrounds with feedback had substantially higher safety culture scores, better engagement and less burnout; walkrounds without feedback did not.

A study of walkrounds in the English NHS found that "local, pragmatic adjustments to the walkrounds approach could radically alter its character", so that staff came to see them as monitoring and evaluation rather than as a way of surfacing concerns.

The rule to hard-wire is not the walk. It is the feedback and the visible action afterwards.

The measure of a leader is what is still standing afterwards. Dalton did try to build beyond his own organisation: with Professor Sir Norman Williams he wrote Building a culture of candour for the Secretary of State in March 2014, which asked every registered organisation to train and support staff to disclose and apologise, to improve reporting "so that this information is used as the basis for organisational learning and not for criticism of individuals", and to close the audit loop by applying and publishing what was learned.

The statutory duty of candour for NHS bodies came into force in November 2014.

What he could not do was make Salford's way of working independent of the people doing it.

Within a few years of his departure the organisation had absorbed a failing neighbour, and the hospital that had twice been rated Outstanding was rated Requires improvement.

That is the pattern this document keeps meeting: what is written into structure survives, and what is left to individuals leaves when they do.

What endures is what is built in

Two hundred years ago, Wales answered this question by building institutions rather than appointing heroes. A parson's schools taught a nation to read; the people who learned built chapels they governed themselves, then libraries and institutes, then a medical aid society in Tredegar that no single leader owned.

When the founders died the institutions carried on, because the rules, the buildings and the habits held what the founders had started. That is the whole argument of the annex on Wales, and it is the argument here in a different century.

A French order of nuns has now held a culture of gentleness for four hundred years, because they wrote it into the entry rules. The Order of the Visitation was founded at Annecy on 6 June 1610 by Francis de Sales and Jane de Chantal, at a time when convents took only the physically robust.

The Visitation deliberately admitted the people other orders turned away: widows, "the aged, provided they were of right mind; the crippled, provided they were sound in mind and heart; even the sick".

And it removed the competitive harshness of the age — "austerities of the cloister, like rising at night, sleeping on hard surfaces, were suppressed".

Francis de Sales set out the substitution plainly: "humility towards God and gentleness towards your neighbour must in your houses take the place of the austerity of others". Gentleness was not preached and hoped for; it was designed into who could join, what was demanded and what was measured. Around 130 Visitation monasteries survive today.

It is worth adding that gentleness was not softness. The Catholic Encyclopedia describes Jane de Chantal herself as a woman of "firmness and great vigour", who "did not like to see her daughters giving way to human weakness".

The two things held together: a merciful rule and an exacting standard. That combination is exactly what healthcare has failed to hold.

The opposite approach has its own long record, and the annex on why a top-down model fails tells it. Benevolence that decides about people without them — the commissioners who judged Wales in 1847, the ranking of human worth that followed, the long-stay hospitals the NHS inherited — ends in harm, however sincere it starts.

A merger designed in Whitehall and a boardroom dashboard in Salford are milder versions of the same habit: deciding about people at a distance from them.

What this would mean at the Northern Care Alliance

None of it requires a reorganisation, which is just as well, because a reorganisation is what caused much of this. The trust is in emergency mode: clearing a gynaecology backlog, steadying an emergency department, mending a dispute with its theatre staff.

Those are the right immediate priorities. The question is what is built while the emergency is being handled, because that is what decides whether the next chief executive inherits a recovery or a relapse.

Three things follow from the three pillars, and each is affordable this year. Give staff the pathway, so that the humanity the organisation wants is taught and practised rather than exhorted. Support circles that belong to a network outside the trust, so that there is a route for concerns and for praise that no executive can close.

And publish, in plain terms, what is protected and what is not: an honest mistake, reported promptly, is protected; silence, downgrading an incident and retaliation against the person who raised it are not.

And change the measure of success. A chief executive's performance indicator should not be the waiting list alone. It should be how well the ward-to-board loop works when the chief executive is not in the room — whether concerns raised last month were answered, whether the people who raised them are still there, and whether anyone below board level can stop an unsafe process without asking permission.

That is the test Salford would have passed in 2015 and failed in 2025, and it is the only one that tells you whether the humanity is in the person or in the fabric.

A note on sourcing. The Northern Care Alliance material here comes from the CQC's own record, NHS England, the trust's board papers and reporting by The Independent and Channel 4 News (19 August 2026).

The leaked consultant survey is reported as broadcast: The Independent's earlier account of a related poll of about 426 consultants gives different proportions, and who commissioned the survey is not public.

Claims that could not be verified at source — including a named recovery director, specific deficit figures circulating online, and a reported rise in stillbirths at one of the trust's hospitals — have been left out.

Radical leadership: honesty, accountability, trust and forgiveness

Every safety-critical industry has to hold two things at once: people must tell you when something has gone wrong, including when they did it, and people must be answerable for what they do.

• Lean too far towards punishment and the truth disappears.
• Lean too far towards forgiveness and nobody owns anything.

Aviation found the line and wrote it down. DuPont found it, built an empire of safety on it, and then partly lost it.

The NHS has swung from one end to the other without ever writing the line down.

The one thing to take away: honesty, accountability, trust and forgiveness only work as a set of four, and each one has to be built into a rule, not asked for in a speech.

The problem in one sentence

A system that punishes error hides it; a system that forgives everything learns nothing. The Thirlwall Inquiry put the second half bluntly: "The move towards a 'no blame' culture in the NHS, starting in 2000, was a mistake … a learning culture that omitted acknowledging responsibility or blame meant difficult conversations about conduct were avoided.

This undermined patient safety."

And: "Instead of being safe, patients are at greater risk of harm in a 'no blame' culture."

The inquiry is equally impatient with the replacement slogan.

Of "just culture" the chair writes: "I question the value of identifying a desirable culture and then adding a single adjective in front of the word 'culture' … It is behaviours that need to change."

That is why this annex is about mechanisms. Each of the four below can be written into a policy, a contract or a piece of equipment, and each can be checked.

What "no blame" sounds like from inside a boardroom is worth hearing. Shortly before he retired in 2018, the medical director of the Countess of Chester, Ian Harvey, addressed the hospital board.

As reported by the Daily Mail in 2023: "There is no doubt this team has been tested. It's all Tony's fault," he joked, referring to the chief executive, Tony Chambers. "I think a couple of years ago he said, 'Well, it's all been fairly easy, hasn't it? We haven't really been tested'. We have."

The babies had died in 2015 and 2016, the consultants' warnings had been met with a demand that they apologise, the police had been called in 2017, and a nurse had been arrested that July.

The word "tested" describes the executives, not the families.

Responsibility passed round the table as a joke. That is the culture this annex is trying to design out — not cruelty, but a settled habit of treating catastrophe as something that happened to the leadership.

Radical honesty

Aviation made protection conditional on telling, and put a clock on it. A pilot who breaks a rule can file a report with NASA's Aviation Safety Reporting System, which is deliberately held outside the regulator.

The Federal Aviation Administration's own rule is precise about what that buys: "although a finding of violation may be made, neither a civil penalty nor certificate suspension will be imposed" — provided the violation "was inadvertent and not deliberate", did not involve a criminal offence, an accident or a lack of competence, the person has no enforcement finding against them in the previous five years, and the report was filed "within 10 days after the violation, or date when the person became aware or should have been aware of the violation".

The scale is the point.

The programme has taken in more than 2.3 million reports since 1976 and issued 8,069 safety alerting messages to people in a position to fix the problem. The FAA "will not use any reports submitted to NASA under the ASRS (or information derived therefrom) in any enforcement action", and NASA does not give the FAA the reporter's identity.

Airlines run the same bargain in-house through the Aviation Safety Action Program, where "safety issues are resolved through corrective action rather than through punishment or discipline", and reports are judged by a committee of a manager, a union representative and an FAA inspector, deciding by consensus.

The bargain has teeth at both ends: reports involving "gross disregard for, or deliberate indifference to, safety" are excluded and referred for investigation, and where there is evidence beyond the report itself, "the FAA investigates apparent violation(s) based on evidence independent of an ASAP report".

Protection is strongest when the report is the only evidence there is. Silence is the one move that costs you everything.

The NHS has the learning half of that bargain and not the conditional half. The Patient Safety Incident Response Framework, published in 2022, replaced the Serious Incident Framework and shifted the emphasis to "understanding how incidents happen rather than attributing blame or liability", while removing the requirement to investigate every incident that crossed a threshold.

NHS England has since withdrawn its 2018 just culture guide as "no longer fit for purpose", replacing it with the Being fair tool.

What no NHS document does is state the exchange in the aviation form: report promptly and you are protected; stay silent and, if we find it another way, you are not.

It is worth being honest about the limits.

A systematic review of incident reporting in healthcare found "little evidence of either improvements in outcomes or changes in the latent managerial factors involved in error production", and concluded that such systems work best when they are "owned and led by clinical teams rather than centralized hospital departments".

Reporting on its own changes nothing.

Reporting that reaches someone who can act, and is answered, is what changes things — which is the same finding as the research on walkrounds.

Radical accountability

The line is drawn by behaviour, not by how badly it turned out. The distinction the safety literature settles on separates human error (a slip), at-risk behaviour (a shortcut taken because everyone takes it) and reckless behaviour (knowingly running an unjustifiable risk), with support for the first, coaching and a fix for the second and no tolerance for the third.

The NHS already owns a version of this: the National Patient Safety Agency's Incident Decision Tree, built on James Reason's culpability model, asks four questions — deliberate harm, incapacity, foresight, substitution — to establish whether an individual or the system is the right subject of attention.

The same line is drawn in aviation, and the exclusions are where it bites: criminal activity, substance misuse, intentional falsification, acting "while knowing that such conduct is contrary to a regulation", and repeated non-compliance by the same person for the same reason.

Everything outside those categories is a system problem; everything inside them is personal.

Accountability also runs upwards from the bedside, and that part needs a rule of its own. Aviation's two-challenge convention is often described as a law requiring a junior pilot to seize the controls.

It is not: a briefed crew procedure allows a pilot who is not flying, having stated a concern twice without a response, to take the controls, while the law still says the pilot in command is "directly responsible for, and is the final authority as to, the operation of that aircraft".

Healthcare's adaptation is the one to copy. In the AHRQ's TeamSTEPPS programme: "It is your responsibility as the challenger to assertively voice your concern at least two times to ensure that it has been heard", then escalate; three scripted words — "I am Concerned, I am Uncomfortable, this is a Safety issue" — carry it; and "any team member should be empowered to 'stop the line' if they sense or discover an essential safety breach".

After training in this method, junior doctors in a simulator challenged an anaesthetist in 72% of cases, against 16% before.

And managers have to be inside the accountability, not above it. The UK government has confirmed a barring scheme for board-level directors and their direct reports, run by the Health and Care Professions Council, to be legislated when parliamentary time allows.

The Thirlwall Inquiry asked for more: a barring system for all managers, clinical and non-clinical, by September 2027, and an individual duty of candour on every manager, on top of the organisation's statutory duty.

Barring is a punishment at the end. The duty of candour is the rule that operates every day, and it is the one this document would put first.

Radical trust

Trust is a structure, not a feeling: it is what the frontline is allowed to do without asking permission. Toyota's andon cord is the most quoted example in healthcare and the most often misdescribed.

Toyota's own account: "the line will not stop each time the Andon cord is pulled. The Andon System is not a line stop, it is more of an alert system." Pulling it summons the team leader immediately, and the line stops only if the problem is not solved within that team member's cycle.

That is a better model than the myth. The worker is guaranteed help within minutes, and stopping is what happens automatically if the help does not work.

The healthcare version exists and has been measured. Virginia Mason Medical Center's Patient Safety Alert system "requires any employee who encounters a situation that is likely to harm a patient to make an immediate report and to cease any activity that could cause further harm", supported by a policy statement, executive commitment, dedicated resources and a 24-hour hotline.

Reports rose from an average of three a month in 2002 to 285 a month in 2006, 6,112 in total, most processed within 24 hours and resolved in two to three weeks.

Be careful what that proves. Those are process results, not proof of less harm, and a systematic review of Lean in healthcare found "no statistically significant association with patient satisfaction and health outcomes" and "potential, yet inconsistent, benefits on process outcomes".

The mechanism worth copying is the guarantee — anyone can summon help, and something happens — not the branding around it.

The patient's version of the same guarantee already exists in England, and it is working. Under Martha's Rule, patients, families and staff have round-the-clock access to a rapid review by a critical care outreach team that is not the team treating the patient.

It is a right to an independent second look, not a right to overrule the doctor: the reviewing team assesses, escalates where needed and works alongside the treating team.

In its first 18 months, to February 2026, there were 12,301 calls, of which 1,080 came from staff; 4,047 identified acute deterioration; 2,310 patients had their treatment changed; and 524 were transferred to intensive care in what NHS England describes as potentially life-saving moves.

It is now in every acute inpatient site in England. Only about a third of the public had heard of it a year ago, which is the next problem to solve.

What DuPont got right, and what it later got wrong

The founder put his own house inside the blast zone. When E. I. du Pont built his gunpowder mills on the Brandywine from 1802, he built the family home among them — in the company's own words, "determined that he and his family would share any risk with employees, he built his house right there in the blast zone".

The written rules followed in January 1811, and one of them is still the sharpest safety rule ever written: "No employee may enter a new or rebuilt mill until a member of top management has personally operated it."

The point is not heroism. It is that the person who signs off the risk is the person who takes it first.

Between 1910 and 1931 the company's fatality rate fell from 5.66 per million hours of exposure to 0.09.

Later DuPont turned the culture into a product: the STOP observation programme, in which managers and peers watch work as it happens and intervene, and the Bradley Curve, a four-stage model of safety maturity — reactive, dependent, independent, interdependent — attributed to a DuPont manager in the mid-1990s.

It is a consultancy model rather than a validated one, and it should be used as a way of describing where an organisation is, not as evidence of anything.

Then the reputation outran the reality, and that is the most useful part of the story for the NHS. In January 2010 there were three chemical releases in 33 hours at DuPont's Belle plant in West Virginia; a worker exposed to phosgene died the next day.

The US Chemical Safety Board found "numerous safety deficiencies including lack of safe equipment design, ineffective mechanical integrity programs, and incomplete investigations of previous near misses", and its chairman said that it happening "at a company with DuPont's reputation for safety should indicate the need for every chemical plant to redouble their efforts".

In November 2014 four workers were killed at DuPont's La Porte plant in Texas.

The Board's final report found that the site "did not effectively implement" the corporate process safety management system "nor did it formally evaluate its process safety culture", and — the finding that should be read twice in any hospital — that the employee bonus structure "may have disincentivized workers from reporting injuries, incidents, and 'near misses'".

That is the same failure alleged at the Northern Care Alliance, in a different industry and with different metrics: when the numbers that decide pay, promotion or regulatory standing are the numbers that reporting makes worse, reporting stops.

No amount of stated culture survives an incentive pointing the other way.

Radical forgiveness

Forgiveness is what makes honesty affordable. If a nurse who reports her own drug error can expect support, a fair process and her job, she reports it in the first hour, when the patient can still be helped.

If she expects an investigation aimed at her, she reports it late or not at all, and the ward learns nothing.

NHS England's Being fair tool exists to make that the norm rather than the exception, and the evidence on staff who are involved in serious incidents — the so-called second victims — says the organisation's response to them is what decides whether they stay in the profession.

Forgiveness in this sense is a discipline, not a mood.

The Order of the Visitation kept a culture of gentleness for four centuries by writing it into its rule rather than relying on the kindness of superiors, as the companion annex describes, and the same order was exacting about standards. The two belong together: mercy for the person, seriousness about the work.

And forgiveness has a boundary, stated in advance. It covers the honest error, the shortcut taken under pressure, the judgment that turned out wrong.

It does not cover silence, the downgrading of an incident to protect a figure, or retaliation against the person who raised a concern.

Those are the behaviours that break the bargain for everyone else, and they are the ones every inquiry in this document finds.

The four rules, in plain words

Radical honesty: tell it early, and protection follows. Protection is conditional and the condition is prompt disclosure — with a clock on it, as in aviation. Reports go somewhere that can act, and the person who reported is told what happened.

Radical accountability: the line is behaviour, not outcome. Human error is supported, at-risk behaviour is coached and designed out, reckless behaviour and cover-up are treated as what they are. The line applies to managers exactly as it applies to clinicians, with an individual duty of candour on both.

Radical trust: anyone can summon help, and anyone can stop an unsafe process. The guarantee is that a call brings someone with the authority to act, within minutes, every time — for staff as at Virginia Mason, and for patients and families as under Martha's Rule.

Radical forgiveness: an honest mistake costs the telling, and nothing more. The organisation carries the person through the process, and reserves its severity for the conduct that keeps the truth from coming out.

And all four exist for one moment: when telling the truth will cost the teller something. That is what holding the line means, and it is the point at which every mechanism in this annex is either real or decorative.

The companion annex sets out what it takes, and why the people who do it need someone holding the line for them. Annex 17 · Holding the line

None of the four is expensive, and none of them works alone. Honesty without forgiveness is a trap. Forgiveness without accountability is the mistake the NHS has spent twenty-five years making. Trust without honesty is sentiment.

Accountability without trust is the fear culture that produced the silence in the first place. Together they are what this document means by humanity-first healthcare with teeth: a system built on the assumption that the people inside it are human, fallible and worth protecting — and that precisely because they are, the truth has to be safe to tell and impossible to bury.

A note on sourcing. Several popular versions of these stories are wrong and have been corrected here: an aviation self-report waives the penalty rather than granting immunity, and the window is ten days; the two-challenge rule is a briefed procedure, not a law; pulling an andon cord alerts a team leader rather than stopping a factory; the Bradley Curve is a proprietary model with no published validation; and a Martha's Rule call brings an independent review rather than a takeover of the patient's care.

Figures come from the FAA advisory circulars, NASA's programme briefing, the US Chemical Safety Board, Toyota's own account of its production system, NHS England, the Thirlwall Inquiry report and the peer-reviewed studies cited.

Healthcare: why the Art of Enablership is critical

Healthcare is too important to rest on the strengths of a few remarkable people.

Remarkable people are wonderful when you have them, but you cannot plan on always having them. That is simply part of life.

That is the power of Enablership: hardwiring the ability to create the right kinds of people, their behaviours into the foundations, the fabric of what makes an effective health organisation. Read on to find out more.

The founders of the American republic faced the same problem in a harder form and answered it by design: a constitution written for the officeholders you might actually get, not the ones you hope for.

The one thing to take away: enablership is the work of making an organisation able — building excellent care into the fabric, the rules and the habits, so that it survives the departure of whoever built it.

Where the phrase comes from. The Art of Enablership: Rethinking Leadership for a Complex World was published by Barrie Watson on 12 April 2026.

Its argument, in the book's own description, is that the obsession with leadership may be holding organisations back, that "over-reliance on leaders can weaken decision-making and silence valuable voices", and that the answer lies in "distributed thinking, shared responsibility, and collective intelligence" — a shift "from control to enablement".

Its closing line is the one this annex is built on: "You don't need better leaders. You need better conditions for people to succeed."

This annex applies that argument to healthcare, where the stakes are patients rather than performance, and tests it against the evidence.

The nearest published relatives of the idea are Jim Collins's distinction between clock building and time telling, and the literature on enabling leadership.

Footnote: clock building and time telling. The phrase is Jim Collins's, from Built to Last (Collins and Porras, 1994). His image is of a person so gifted that they can look at the sky and tell you the exact time: impressive, but useless the moment they leave the room.

Building a clock is less glamorous and worth far more, because the clock goes on telling the time after its maker has gone. Applied to organisations, time telling is a leader who personally supplies the standard — the answer, the energy, the judgement — and clock building is designing an organisation in which the standard holds without them. The section below applies the distinction to hospitals.

Healthcare cannot be run on heroes

Salford Royal (Manchester, UK) is the clearest case in this document, and it is a fair one because nobody did anything wrong. Under Sir David Dalton, who led it for 18 years until 31 March 2019, it was rated Outstanding by the Care Quality Commission in March 2015 and again in August 2018.

In 2016 it was asked to take on the leadership of a neighbouring trust that had been rated inadequate; in October 2021 that trust was dissolved into it.

Salford Royal Hospital was rated Requires improvement in December 2022 and again in February 2026, and in August 2026 the enlarged trust was served a warning notice covering risk, patient safety, complaints, speaking up and workforce wellbeing.

That is not a story about a bad successor. A pandemic, a merger and a decade of financial strain would have tested anyone.

It is a story about what was portable and what was not: the buildings, the staff and the reputation transferred, and the way of working did not.

The fuller account is in the companion annex on applying lessons from the past. Annex 12 · Humanity-First Healthcare: applying lessons from the past

The NHS changes its leaders faster than any culture can set. In 2018 the King's Fund found the median tenure of a substantive NHS trust chief executive was three years, with a mean of four; 17% had been in post less than a year and only 11% for more than ten; and more than half were in their first chief executive role.

In trusts rated inadequate, 72% of executive directors had been appointed within the previous year, against 20% in trusts rated outstanding — an association that runs in both directions, since churn follows poor ratings as well as causing them.

If the typical leader has three or four years, then any improvement that needs a decade of the same person to hold it in place is not an improvement. It is a loan.

And the answer is not to pretend that leaders do not matter, because the evidence says they matter more than they used to. Studying 18,467 company-years from 1950 to 2009, Quigley and Hambrick found the share of performance variance attributable to the chief executive rose across successive eras — from 4.1% to 10.9% to 16.2% on one method, and from 8.6% to 20.3% to 26.4% on another.

Leaders are consequential. That is precisely why a system that depends on drawing a good one is a system waiting for its turn.

The honest position is that nobody has properly studied what happens next. A systematic review of healthcare programme sustainability in 2023 found that committed leadership "play[s] a critical role in sustaining interventions", and then found that the literature barely looks at the question this annex is about: only 8 of the studies reviewed — 6.5% — reported on programmes that had been discontinued, and fewer than a quarter looked beyond five years.

Healthcare celebrates turnarounds. It does not go back five years later to see whether they held.

Clock building, not time telling

The most useful sentence in management writing on this subject is Jim Collins's. "Leading as a charismatic visionary — a 'genius with a thousand helpers' — is time telling; shaping a culture that can thrive far beyond any single leader is clock building."

The distinction is exact. A brilliant chief executive who personally holds the standard is telling the time. An organisation in which the standard holds without them is a clock.

Collins's related finding is uncomfortable for healthcare, where the hero narrative is strong.

Comparing companies that sustained excellence with those that did not, he reported that "in more than three-quarters of the comparison companies, we found executives who set up their successors for failure, chose weak successors, or both".

His Level 5 leaders are not modest in ambition — they display "a powerful mixture of personal humility and indomitable will", with the ambition directed "first and foremost for the cause, for the organization and its purpose, not themselves".

What the Americans did about the same problem

The most systematic attempt ever made to build an institution that survives its officeholders starts from a low view of human nature, honestly stated. James Madison, in Federalist 51 in 1788: "If men were angels, no government would be necessary. If angels were to govern men, neither external nor internal controls on government would be necessary."

The design principle follows immediately: "Ambition must be made to counteract ambition. The interest of the man must be connected with the constitutional rights of the place."

Madison is explicit that structure exists to cover for the absence of virtue — "supplying, by opposite and rival interests, the defect of better motives" — and that this "might be traced through the whole system of human affairs, private as well as public".

In Federalist 10 he puts the operational version in nine words: "Enlightened statesmen will not always be at the helm."

That is the sentence every NHS board should have on the wall in place of its values poster, because it is the one that tells you what the values poster is for.

The design was not a case against strong leadership, and it would be a misuse of the analogy to pretend otherwise. Alexander Hamilton argued in Federalist 70 that "energy in the Executive is a leading character in the definition of good government", and that a single executive acts with more "decision, activity, secrecy, and despatch" than a committee.

His safeguards were not weakness but answerability: "a due dependence on the people" and "a due responsibility", which are easier to enforce on one person than on a group who can share out the blame.

A hospital needs an energetic chief executive for the same reason. It needs one who can be held to account, and whose authority stops where patient safety begins.

And then there is what George Washington did, which mattered more than anything he wrote. In December 1783 he resigned his commission rather than keep an army at his back — "Having now finished the work assigned me, I retire from the great theatre of Action" — and in 1796 he declined a third term.

He understood exactly what he was doing. Writing in January 1790 he said: "In our progress towards political happiness my station is new; and, if I may use the expression, I walk on untrodden ground. There is scarcely any part of my conduct wch. may not hereafter be drawn into precedent."

The familiar line that "no man is indispensable" is not in the Farewell Address, and this document does not use it.

What Washington did is stronger than any slogan: the most powerful individual in the new republic twice chose to make himself unnecessary, so that the institution rather than the man became the thing people relied on.

That is the act healthcare almost never sees. Leaders are rewarded for indispensability, and the reward destroys the very thing that would outlast them.

The analogy has limits, and they are worth saying out loud. A hospital is not a state; separation of powers can produce deadlock as well as safety; and the American system has plainly not immunised itself against its own crises.

The principle that transfers is narrower and sturdier than the metaphor: build for the officeholder you might get, connect each person's interest to the duties of their post, and never let one part of the organisation be the sole judge of its own conduct.

What enablership means inside a hospital

Enablership is not delegation, and it is not empowerment as a slogan. To enable someone is to equip them, to authorise them and to protect them: to give them the capability to act well, the standing to act without asking permission, and cover when they do.

• Delegation without capability is abandonment.
• Capability without authority is frustration.
• Authority without protection is a trap, and the inquiries in this document are full of people who found that out.

Three branches make an organisation able, and none of them is the chief executive's to switch off.

The first is the individual: the Champions Pathway, which builds the capacity to act for the good of others under pressure, so that humanity is a trained capability rather than a personal trait some staff happen to have.

The second is the community: Champions Circles, small groups that belong to a network beyond the employer, so there is always a route that management cannot close — to call it forward when something is being done well, and to call it out when it is not.

The third is accountable leadership: trust, forgiveness and being held to account, with the line between honest error and unacceptable conduct written down in advance, as the annex on radical leadership sets out. Annex 13 · Radical leadership

Put in Madison's terms, the first branch supplies capability, the second supplies the check that does not depend on goodwill, and the third supplies the answerability that makes both real.

Put in Collins's terms, all three are clock parts. Not one of them requires a reorganisation, new legislation or a heroic appointment.

Dalton's ward boards at Salford are the test case for the difference. Displaying every ward's harm and staffing figures where patients and families could read them — "56 days since the last fall and 500 days since a patient had developed a hospital acquired pressure ulcer" — was structural.

It did not depend on the chief executive being in the building, and anyone walking past could see the truth.

The parts of Salford's excellence that depended on the chief executive personally working alongside staff were not structural, however admirable, and they left with him.

The question for any leader is which of their practices would still be running, unprompted, two years after they had gone.

The constitution test for a board

A board can audit its own dependence on individuals in an afternoon, with four questions.

1. What here would still be true six months after the chief executive left?
2. Which of our safety checks can an executive switch off, defer or reword to meet a target this quarter?
3. Who, below board level, can stop an unsafe process without permission, and when did one of them last do it?
4. And when the internal route fails, where does a member of staff go — to somewhere inside this organisation, or to somewhere that does not report to us?

An organisation that cannot answer those four is running on the character of the people currently in post. That may be working today.

It is not a plan, and every inquiry in this document is what it looks like when the luck runs out.

The practical consequences are modest and immediate. Write the handful of rules that cannot be changed in-year to hit a number, and publish them.

Fund circles that sit outside the trust.

Make succession a duty of the post rather than a courtesy to it, and evaluate a departing leader on what still works a year later.

And ask regulators to look for built-in capability rather than compliance — because the manager barring scheme now being legislated, whatever its merits, acts after the harm, and no barring list has ever made a ward safe.

The NHS has done the structural half of this before, and knows it can be done. Bevan wrote the money into law in 1948: free at the point of need, funded collectively, and that half has survived seventy-eight years and every reorganisation since.

The humane half was left to the professions, to training and to chance, and it is the half that appears in every inquiry report.

Enablership is the name for finishing the job — building the human half into the fabric the way the financial half was built in, so that good care does not depend on who happens to be in charge this year.

A note on sourcing. Madison, Hamilton and Washington are quoted from the Federalist Papers (Avalon Project), Washington's Farewell Address and his letter of 9 January 1790; his resignation words are as recorded by Mount Vernon.

The Collins quotations are from his own published concepts and from Harvard Business Review, January 2001. NHS chief executive tenure figures are from the King's Fund, Leading in today's NHS, 2018; the chief-executive effect from Quigley and Hambrick, Strategic Management Journal, 2015; the sustainability evidence from Zurynski and colleagues, Implementation Science, 2023.

The Salford and CQC material is sourced in the companion annex. The framing of enablership is taken from Barrie Watson, The Art of Enablership: Rethinking Leadership for a Complex World (12 April 2026); the wording quoted here is from the book's published description rather than from its text, and the application to healthcare, with every claim and figure in it, is this document's own.

Humanity in healthcare: examples

Humanity built into the fabric of a health organisation is not an aspiration.

It has been done, in a Welsh mining town, a Swedish county, a Dutch neighbourhood, an Alaskan city, an American clinic founded by a doctor from Salford, and in two English NHS trusts and a converted east London chapel.

This annex asks a narrower question than Volume 8 does: in each case, what exactly was built in, what happened when other people tried to copy it, and which part actually travels.

The one thing to take away: the mechanism rarely transfers, the principle usually does — and the copy only works when the people who have to run it own it.

A note on what this annex adds. Volume 8 sets out the bright spots as evidence that humane care is achievable and often cheaper. This annex is the practical companion: it reads the same cases, and others, as design examples, and it is honest about the ones that did not survive the journey. 8 · Bright spots

Why copying fails, stated first

The evidence on spreading good practice is blunt about it. "A technology or pathway that works smoothly in setting A will operate awkwardly (or not at all) in setting B," wrote Trisha Greenhalgh and Chrysanthi Papoutsi in the BMJ in 2019, adding that innovations "rarely achieve widespread uptake even when there is robust evidence of their benefits".

The Health Foundation's own review of the problem is equally direct: "when initially successful interventions are spread to new settings they may fail to achieve the same impact, or indeed any impact at all".

Implementation research offers the distinction that makes this usable.

An intervention has core functions — "the core purposes of the change process" — and forms, "the specific strategies or activities that may be customized to local contexts".

Copy the function; redesign the form. The Health Foundation puts the same point as a question every board should answer before importing anything: which components are core to making this work, and "what needs to remain invariant versus what can be changed".

The people who built the most admired system in this annex say it themselves. Reviewing Alaska's Nuka system, the King's Fund reported that "Southcentral's leaders would not recommend simply 'copying and pasting' their model".

Tredegar: a town that owned its own health service

The oldest example in this document is also the most complete, because the people who used the service owned it. The Tredegar Workmen's Medical Aid and Sick Relief Fund was formed in 1890 from four existing funds.

Members paid three pence in the pound of their wages — a contribution proportionate to income — "for medical attention, medicine and appliances".

The society opened the Park Cottage Hospital in 1904 and the Central Surgery in Church Street in 1911. Under Walter Conway's secretaryship, from 1915 to 1933, it covered some 20,000 people, about 95% of the town, employing doctors, dentists, nurses and support staff.

Aneurin Bevan sat on its hospital management committee from 1924 and chaired it in 1929–30.

He later said he was "extending to the entire population of Britain the benefits we had in Tredegar for a generation or more".

The historian Steve Thompson's caution is worth carrying: the claim that the NHS was modelled on Tredegar "is not 100% correct — the truth is a little more complex", since the NHS also drew on Poor Law medicine, the 1911 National Insurance scheme, voluntary hospitals and municipal medicine. Tredegar was inspirational rather than a blueprint.

What was built in was ownership, and that is the part that travels. The subscription, the committee and the cottage hospital were forms, and they belong to their century.

The function was that the people who depended on the service governed it, and could be neither ignored nor patronised.

The annex on Wales traces where that habit came from: a century of chapels, libraries and institutes that the same families built and ran for themselves. Annex 10 · Why we all love the Welsh

Elizabeth Fry: proximity first, institution second

Modern nursing began with a woman going in person to a place nobody else would enter. Elizabeth Fry first visited Newgate Prison in 1813; the conditions she found there began a life's work.

Nursing came later and almost sideways: in 1840 she founded what was first called the Protestant Sisters of Charity and then, after Anglican objections and at Queen Adelaide's suggestion, the Institution of Nursing Sisters.

In September 1841 Guy's Hospital agreed to take two trainee sisters, and probationer placements continued through the 1840s and 1850s.

The Institution's own report of 1848 states its purpose in one sentence that could be the mission statement of this whole document: "It is the spirit of the Institution of Nursing Sisters that the poor should be attended in their houses with the same kindness and care bestowed upon the more wealthy part of the community."

Be careful with the Nightingale connection, because the links are real but loose. The two women moved in overlapping circles: Lady Inglis, who presided over the ladies' committee of Fry's Institution, was, in the historians' words, "a friend of the Nightingale family", and knew of Florence Nightingale's move to the Harley Street institution in 1853.

The Dictionary of National Biography records that Nightingale's horizon was widened "possibly by a chance meeting in girlhood with Mrs. Elizabeth Fry" — a tradition, carefully hedged even there. And the German connection runs through Fry herself: when Theodor Fliedner, the Lutheran pastor who founded the deaconess school at Kaiserswerth, came to London, his most important friendship there was with Elizabeth Fry, and her work at Newgate was, in his biographer's phrase, a revelation to him.

Fry's own institution drew on Kaiserswerth in turn, and Nightingale trained there in 1851. Fry shaped the school that trained Nightingale.

What cannot be shown is direct contact or acknowledgement. The standard study concludes that "written proof of direct contact, although likely, is lacking", and the claim that Fry's nurses sailed with Nightingale in October 1854 does not survive checking: the party of thirty-eight was drawn from St John's House, Miss Sellon's Anglican sisters, the Bermondsey and Norwood convents and the London hospitals.

Two nurses trained at the training institution in Devonshire Square — the address of Fry's Institution — do appear in the register of nurses sent to the East, but later, in 1855. By 1857 all ninety of the Institution's trained sisters were at work in private nursing.

The honest summary is that Fry built the first model a generation earlier, that the two were connected by circle rather than by collaboration, and that, as the same study puts it, Nightingale's fame "inadvertently dimmed those who preceded her".

The lesson is the order of events. Fry did not set out to reform healthcare. She went close to people the system had written off, let that proximity change what she thought was tolerable, and only then built the institution that outlived her.

Proximity first; structure second; and the structure is what makes the proximity survive its founder.

Whether Fry and Nightingale ever met matters less than the fact that they were cut from the same cloth. Humanity is there to be found by anyone who looks: it has one source and comes out in different flavours — a Quaker's prison visiting, a statistician's sanitation tables, a Dutch nurse's caseload, an Alaskan community's clinic. The variety is the pleasure of it, and it is what makes a community healthy as well as a hospital.

Mayo: the doctor from Salford, England inspires the world's best hospital (in USA)

The most admired clinic in the world was founded by a doctor from Salford. William Worrall Mayo was born in Salford, near Manchester, on 31 May 1819, and emigrated in 1846.

Mayo Clinic's own account of its values traces them to that English upbringing, including the influence of the scientist John Dalton, who it says instilled "Quaker values of tolerance and working together for the common good".

The clinic itself began in response to a natural disaster. After the 1883 tornado in Rochester, Minnesota, the Sisters of Saint Francis, led by Mother Alfred Moes, proposed to build a hospital if the Mayos would staff it; Saint Marys opened in 1889 — a day early, because a patient needed surgery.

The institution's primary value, "the needs of the patient come first", is Mayo's own distillation of what William J. Mayo told the Rush Medical College graduates in 1910: that "the best interest of the patient is the only interest to be considered".

What was built in is structural, not exhortatory, as Volume 8 sets out: salaried physicians, with no financial reason to do more or less than the patient needs.

The function that travels is not the salary scale but the principle behind it — find the incentive that competes with the patient's interest, and remove it.

There is an irony worth noting, since this document has spent two annexes in Greater Manchester. The values Mayo exported to Minnesota came from Salford; the hospital that carries Salford's name was, until recently, the most admired in England, and is now under a warning notice.

Institutions are not owed their inheritance. They keep it by building it in.

Esther, Sweden: one question, with authority behind it

A whole county reorganised itself around a question about one old woman. The Esther model began in 1997 in the Högland area of Jönköping County, led by Mats Bojestig, after the journey of an elderly patient with breathing difficulties through a fragmented system — in the Commonwealth Fund's account, "Esther saw a total of 36 different people and had to retell her story at every point".

Esther became a persona, and one question was put into every meeting, handover and budget discussion: what is best for Esther?

The question was given hands. From 2006 frontline staff across hospitals, primary care and municipal social services were trained as Esther Coaches — eight days of structured training over eight months in problem analysis, improvement and client focus — and 102 were in place by 2011, with authority to redesign pathways across organisational boundaries.

The published results are before-and-after figures rather than a controlled study, and should be quoted that way: medical department admissions fell from about 9,300 in 1998 to 6,500 in 2013; hospital days on medical and geriatric wards from 48,222 in 2002 to 44,769 in 2013; rehabilitation stays from 19.2 to 9.2 days between 2009 and 2014.

England imported the question and not the authority. Kent and Medway adopted Esther from 2013, hired a Swedish coordinator in 2016, and by 2019 had trained 88 Esther Coaches with more than 500 ambassadors across a population of 1.8 million.

No quantified outcome from the English programme has been published.

A coach who can ask what is best for Esther, but cannot override a discharge that suits the bed numbers, is asking a rhetorical question.

Buurtzorg: the model most copied, and least transferred

Buurtzorg is what happens when you delete the management layer and trust the nurses. Founded in the Netherlands in 2006 by Jos de Blok, it works through self-managing teams of 10 to 12 nurses, each looking after 50 to 60 patients in one neighbourhood, with coaches rather than managers.

Overheads ran at about 8% of costs against a Dutch sector average of 25%; client satisfaction is 9.2 out of 10 across more than 15,000 published ratings; it has repeatedly been voted the best employer in the Netherlands, with sickness absence and turnover around half the sector average.

Two cautions, because the famous numbers are usually misquoted.

Ernst & Young found that Buurtzorg met patients' needs using about 40% of the authorised care hours, which is not the same as saving 40% of a national health budget.

KPMG found 108 hours of care per patient a year against 168 elsewhere, and a lower average home-care cost — but once the whole pathway was counted and adjusted for case mix, Buurtzorg's total cost per client sat just below the national average.

The English copies failed in instructive ways. In Tower Hamlets, East London, a neighbourhood nursing team ran from June 2017 to August 2018; the published evaluation concluded that "some of the core features of the Buurtzorg model were difficult to put into practice in the National Health Service (NHS) because of significant cultural and regulatory differences", that "an embedded professional hierarchical culture made it challenging for nurses", and that the organisation "showed reluctance in relinquishing responsibility" over decisions such as budgets.

In West Suffolk, a test team peaked at six staff with a caseload of 16 to 20 people and was down to fewer than two whole-time equivalents by autumn 2018; the King's Fund found genuinely "outstanding holistic care for patients/clients", alongside staff who had not been given the leadership training or support that self-management requires, and who were sharing one computer.

An international review of 25 Buurtzorg-derived models reached the same verdict: the obstacles were "the self-managed working culture, the organizational framework, or national healthcare policies", and implementing such a model "is complex, challenging, and requires adaptations on several levels".

What travels is continuity and time; what does not is autonomy granted inside a hierarchy that has not changed. Self-management is not a team-level intervention. It is a change in what the organisation above the team is allowed to do, and unless that changes, the pilot is a bubble waiting to be reabsorbed.

Virginia Mason: an obligation, not a permission

Seattle's answer was to make stopping the line a duty rather than a right. Virginia Mason adopted its production system in June 2002. Under the Patient Safety Alert system, any member of staff who sees something likely to harm a patient must report it immediately and stop the activity.

Reports rose from an average of three a month in 2002 to 285 a month in 2006, 6,112 in all, most resolved within two to three weeks; the most serious alerts escalate to a senior executive or the on-call administrator with resolution required within 24 hours.

The habit held for a decade and a half, and the numbers kept climbing. The 50,000th alert was recorded in September 2014, running at about 879 a month. By 2018 the total had passed 73,000, and by September 2020 — the last count the hospital published — it stood at about 120,000.

The outside scorecards moved with it. Virginia Mason was named a Leapfrog Top Hospital in most years from 2007 onwards and has held an 'A' safety grade in every Leapfrog round since the grades began in 2012.

It was not a clean run. In June 2016 the Joint Commission placed the hospital on contingent accreditation after finding 29 standards out of compliance, including a dialysis unit being run by a respiratory therapist; around 650 patients were notified of possible hepatitis B exposure. No infections were found, and full accreditation was restored in September 2016.

And the honest counterweight belongs here too. In October 2023 the Washington State Nurses Association picketed the hospital over staffing and safety, reporting medical-surgical assignments of six and sometimes seven patients, turnover above 100% in several departments in a single year, and that 40% of registered nurses had less than eighteen months' experience.

Virginia Mason merged with CHI Franciscan in January 2021 and its long-serving chief executive stepped down at the start of 2022; the enlarged system says the production system is being trained across the organisation, though no independent evaluation confirms it survived intact.

The lesson is not that the alert system failed. It is that a mechanism for raising concerns does not by itself supply the staff to act on them, and that no single tool immunises an organisation against the ordinary pressures described throughout this document.

A note on the evidence. The headline Virginia Mason figures — the alert counts, the professional liability premium reduction, the savings claimed from the production system — are self-reported and have not been independently audited.

The premium claim has drifted in the retelling, from 56 per cent in earlier accounts to 76 per cent in later ones, and the hospital has never published an infection or mortality series alongside the alert counts.

The alert system is well evidenced as a habit sustained over fifteen years. It is not evidence, on its own, of a particular clinical outcome.

Nuka, Alaska: ownership as the design brief

The most complete example of humanity built into a system is the one where the patients own it. Under the 1975 Indian Self-Determination Act, Alaska Native people took over their own services: Southcentral Foundation was established in 1982 and assumed primary care and other services from the Indian Health Service in 1998, with co-ownership of the Alaska Native Medical Center in 1999.

Patients are "customer-owners", and the word is not decoration — they are the people the organisation answers to.

Each primary care team — a doctor, a nurse case manager, a certified medical assistant and administrative support — looks after about 1,400 people, with pharmacists, dieticians, midwives and behavioural health consultants attached to each clinic.

The published results are among the strongest in any health system: hospital admissions down 53%, emergency department use down by more than 45%, specialist referrals down by more than 60%, primary care visits down 36%, four-week waits replaced by same-day access, and 96% of customer-owners saying they were involved in decisions about their care.

Staff turnover fell from 37% in 2008 to 17% in 2011. Southcentral won the Malcolm Baldrige National Quality Award in 2011 and again in 2017.

The caveats matter as much as the figures. Southcentral has a single payer, a defined population, per-head funding above the NHS average, and a workforce drawn largely from the community it serves; it also covers 55 remote villages reachable only by air or sea.

Its own leaders, as the King's Fund records, would not recommend copying and pasting the model.

What transfers is the design question: who does this organisation belong to, and whose preferences count as the brief?

Gundersen, Wisconsin: a town that learned to talk about dying

Gundersen's transferable achievement is that a whole community learned to talk about dying before it had to. Its Respecting Choices advance care planning programme, established in 1990, was not run as a hospital initiative.

It was built across La Crosse County with the other local health providers and the area's nursing homes: trained facilitators who had the conversation in ordinary time, a shared document, and an agreement that a person's stated wishes would travel with them wherever they were treated.

It worked to a degree no other American community has matched. A study of deaths in the mid-1990s found 85% of decedents had a written advance directive, 96% of those documents were in the medical record, and treatment matched the stated preferences in 98% of cases; a repeat study a decade later found 90%, 99% and 99%.

This is the argument of the whole annex in miniature, which is why it belongs here. The forms are local and replaceable — the facilitators, the paperwork, the county-wide agreement. The function is that a conversation about how a person wants to be treated at the end of their life belongs to the community and happens in ordinary time, rather than in a corridor at two in the morning.

Respecting Choices has since been licensed and taught in many other places, and the part that spread most easily was the paperwork. The part that made La Crosse different was that the town had already had the conversation.

One further Gundersen habit belongs here, because it is a rule about money. Under Jeff Thompson, chief executive from 2001 to 2015, the Envision programme treated the health of the local population and environment as a test that capital investment had to pass.

After an energy audit in 2008 the system cut its energy use by about 40%, saving in the region of $2m (£1.5m) a year, and on 14 October 2014 it became the first American health system to produce more energy than it consumed — while expanding its buildings by a quarter.

What travels is the rule, not the wind turbines. A capital decision that must demonstrate benefit to the community it serves, and an advance care planning programme that belongs to the town rather than to the hospital, are both statements about who the institution is for — and both outlasted the chief executive who introduced them.

East London: improvement as an identity, with the caveats stated

The best-documented English example of building humanity into method is East London NHS Foundation Trust. It began its improvement programme in February 2014 with the Institute for Healthcare Improvement and trained staff at every level in it.

Across six wards in Tower Hamlets, physical violence fell from 12.1 incidents per 1,000 occupied bed days in 2014 to 7.2 in 2015, and as the work spread the trust moved from the top quartile for ward violence to the bottom quartile.

Footnote: what east London means here. Tower Hamlets, the borough most of this work sits in, is not the London of the postcards. It is among the most deprived local authority areas in England and has the highest rate of child poverty in the country: around half of its children live in poverty once housing costs are counted (Trust for London, from Department for Work and Pensions data, 2024/25).

It also has one of the highest rates of household overcrowding in England, and fewer than a quarter of its residents describe themselves as white British (Census 2021). For readers outside the UK, and for anyone who knows London only by its centre: this is not Chelsea.

Whatever is achieved here is achieved in one of the hardest places in the country in which to achieve it.

Waiting time from referral to first visit in community services fell 23% over the decade. The trust was rated Outstanding by the CQC in 2016, 2018 and again in 2022 — the first community and mental health trust to be rated Outstanding three times in a row.

Coproduction is built into the method rather than added to it: the trust's own analysis of more than 500 completed projects found that projects with full service-user involvement were 2.8 times more likely to achieve their aim, and a record 224 people with lived experience now work in trust roles such as peer support.

The caveats are part of the example, not a footnote to it. A focused CQC inspection published in July 2026 rated the trust's mental health crisis services and health-based places of safety as Requires Improvement, with five regulatory breaches; and in the 2024 NHS Staff Survey the trust's staff engagement score sat just below its peer-group average, recovering slightly in 2025.

An identity is not a permanent state. It has to be renewed, in exactly the way this document argues values must be.

Northumbria: treating staff as the strategy

Northumbria Healthcare has been rated Outstanding twice, in 2016 and 2019, and its distinguishing feature is how it treats its own people. Its staff rated it the best acute and combined acute and community trust in England in the 2020 and 2021 NHS Staff Surveys, and in 2025 it still matched the best result in its peer group on compassion and inclusion, on health and safety, and on morale — on a 74% response rate against a peer median of 47%.

The mechanisms are unglamorous and mostly cheap: an in-house psychology and counselling service with fast-track physiotherapy within three working days, gyms at £1 ($1.35) a month, a partnership with a local credit union offering savings and affordable loans, eleven staff networks, and its own staff bank.

Its specialist emergency care hospital, opened in June 2015, was independently evaluated in the Emergency Medicine Journal in 2020: 60-day mortality odds of 0.94 and readmission risk of 0.90 to 0.92 against the previous model, with the largest benefit in patients over 80.

One correction to a claim often made about the trust: its facilities company is a wholly owned subsidiary of the trust, not an employee-owned business.

Bromley by Bow: the community builds it, again

In 1984 Andrew and Susan Mawson arrived at an east London church with a handful of members, and did what Tredegar did. They handed the building to the people around it.

The Bromley by Bow Centre now runs health and wellbeing services, social prescribing, advice, employment and enterprise support from the same site as a GP practice — run by a separate partner organisation, Bromley by Bow Health, which is worth stating precisely since the Centre is often described as running the surgery.

Footnote: what social prescribing means. Social prescribing is the practice of referring a patient to non-medical support in their own community when the thing making them unwell is not primarily medical.

In the NHS it usually runs through a link worker attached to a general practice: the GP refers, the link worker spends the time a ten-minute appointment does not allow working out what would actually help, and then connects the person to whatever exists locally.

Typical examples are a walking, gardening or choir group for someone isolated after a bereavement; a debt or benefits adviser for someone whose anxiety is really money; a housing caseworker for the child whose asthma is really damp; cookery and exercise classes for someone newly diagnosed with type 2 diabetes; English classes or an employment adviser for someone whose health is stuck behind their circumstances.

The logic is straightforward: if the cause is not medical, a prescription will not fix it, and the money may be better spent on a bus pass and a phone call. The evidence, set out below, is that it works best where the community infrastructure already exists and the practice has changed in order to use it — and thins out where a link worker has been bolted onto a surgery that has not changed at all.

In 2024/25 the Centre worked directly with 5,128 people, 94% of them from Tower Hamlets, delivering 8,391 one-to-one sessions and 13,505 hours of support, with six social prescribing programmes reaching 1,658 people.

Its outcome figures — 74% saying their concerns had eased, 63% feeling more connected — are self-reported, and there is no recent independent evaluation.

It is fairer to call it one of the earliest and most influential English examples of social prescribing than its birthplace: the King's Fund traces the idea back to the Peckham Experiment of the 1920s.

The national evidence shows what happens when the idea is scaled without the community that made it work. By the end of 2023, about 5.5 million GP consultations in England had led to a social prescribing referral.

The national evaluation published in the British Journal of General Practice in 2025, covering more than four million survey responses, found that adding link workers was associated with small improvements in patients' confidence in managing long-term conditions and in feeling supported by local services — and that "our findings were unable to support the NHS aim for patients to feel more connected, less lonely, or less isolated".

A national wellbeing analysis published in 2026 found meaningful before-and-after improvements in wellbeing and anxiety after referral, and a systematic review in 2024 concluded that social prescribing showed "some improvements across a range of outcomes although the quality of evidence remains poor".

The realist evaluation of the same programme gave the whole argument its title: link workers were bolting on, fitting in, or belonging. Bromley by Bow belongs.

A link worker attached to a practice that has not changed is bolted on — and a third of them were considering leaving.

Martha's Rule: the newest example, and already the largest

England's own contribution is barely two years old and is already running in every acute inpatient site. A patient, family member or member of staff can call for an urgent review by a critical care outreach team that is independent of the team treating the patient.

In its first eighteen months, to February 2026, there were 12,301 calls; 4,047 identified acute deterioration, 2,310 patients had their treatment changed, and 524 were transferred to intensive care. The annex on radical leadership sets out how it works and what it does not do. Annex 13 · Radical leadership

What the examples have in common

Five features recur, and none of them is a structure. Each place has an identity stated in the first person — this is who we are — rather than a values statement in the third.

Each converts humanity from a permission into an obligation: someone must stop, ask, report or call.

Each has somewhere to go that is not the line of command.

Each protects time for the relationship, whether that is a Buurtzorg nurse's hour or a Nuka team's continuity.

And each has at least one rule about money that cannot be waived to make a quarter's numbers.

And each was built by the people who had to live with it. Tredegar's miners, Fry going into Newgate herself, an Alaskan community taking over its own clinics, a Swedish county inventing Esther, a Dutch nurse designing the job he wanted to do: in every case the design belonged to the people doing the work or receiving the care.

That is why the imported copies thin out.

Ownership is not a nice feature of these examples. It is the mechanism.

So the rule for anyone importing one of these is simple. Name the function you are trying to reproduce, not the form. Redesign the form for your own place, with the people who will run it. Change whatever sits above the team so that the new way of working is not quietly reabsorbed.

And measure the thing you actually wanted — not the tool you installed.

A note on sourcing. Tredegar from Gwent Archives, the British Journal of General Practice (2018) and Steve Thompson (Aberystwyth, 2018); Elizabeth Fry from Huntsman, Bruin and Holttum, Medical History (2002), quoting the Institution's own report of 1848; Mayo from the clinic's own heritage pages.

Esther from the Commonwealth Fund's case study and working paper, the 2023 BMC Health Services Research paper and the International Journal of Integrated Care (2019) for Kent; Buurtzorg from the Commonwealth Fund (2015), Buurtzorg's 2024 quality report, Lalani and colleagues in BMC Health Services Research (2019), the King's Fund's West Suffolk review (2019) and Hegedüs and colleagues (2022); Virginia Mason from Furman and Caplan (2007), Virginia Mason's own materials and the Washington State Nurses Association (2023); Nuka from the King's Fund (2015) and the Baldrige award profiles (2011, 2017); Gundersen from its own releases and the Respecting Choices programme history; ELFT from CQC reports, the trust's annual report and the Institute for Healthcare Improvement; Northumbria from CQC reports, NHS Staff Survey benchmark reports and Price and colleagues, Emergency Medicine Journal (2020); Bromley by Bow from the Centre's own annual review and the King's Fund; social prescribing evidence from Wilding and colleagues, BJGP (2025), Bu and colleagues (2025 and 2026) and O'Sullivan and colleagues, Clinical Rehabilitation (2024).

Figures that could not be verified at source — including the widely quoted "40% saving" attributed to Buurtzorg and a claim that 96% of one American county's residents hold advance directives — are corrected above rather than repeated.

Added in this revision: the Virginia Mason alert totals, Leapfrog recognition and the June 2016 Joint Commission contingent accreditation from Virginia Mason's own published materials, Leapfrog Hospital Safety Grade records and contemporaneous reporting; the La Crosse advance directive studies from Hammes and colleagues; the Tower Hamlets figures from Trust for London (Department for Work and Pensions data, 2024/25) and the 2021 Census; the description of social prescribing from NHS England's guidance on social prescribing link workers.

When the sheriff gets it wrong

Every system of assurance includes someone from outside who is trusted to call it as it is, to form an unbiased and authoritative view, something others can rely upon. In short, their voice matters. But what happens when this fails?

When that person gets it wrong, everyone downstream inherits a false reassurance and acts on it — the board, the regulator, the families.

This annex takes three cases: the Royal College of Paediatrics and Child Health review at the Countess of Chester in 2016; County Durham, where the same kind of check did eventually work and still took thirteen years; and the accounting firm that was built on audit independence and destroyed by the loss of it.

Maybe there is a case for Martha's Rule applied more widely. It gives families and staff the right to call for an urgent second opinion from a team that is not the one treating the patient, and in September 2026 it was extended to emergency departments.

The same right, pointed at an organisation rather than a bedside, is close to what this annex argues for.

Everything in the Chester section comes from the Thirlwall Inquiry's own account. Chapter 20 of Volume II of the report sets out the review from the first phone call to the last letter, and the quotations here are the Inquiry's words or the evidence it records.

What the review was, and what it could never have been

The RCPCH ran a paid service, not an investigation, and it says so. The Royal College set up its Invited Review Service in 2012. For a fee — in this case £19,000 ($25,650) plus VAT — a small team visits a paediatric service, interviews people, reads documents and reports with recommendations.

As the Inquiry puts it, drawing on the College's own evidence, "it is not a regulatory process or a forensic investigation". Its purpose is peer review to improve a service.

That distinction was not a technicality. It meant that the one question the Countess most needed answered — were these babies being deliberately harmed — was the one question the review could not answer. What matters is what happened to that limitation once the review was commissioned: it disappeared.

None of that is a defence, and this annex should not be read as offering one. A review team that has been told, before it arrives, that consultants suspect babies are being deliberately harmed is not relieved by the wording of a contract of the obligation to say so.

The fee, the scope and the terms of reference explain how the review came to be the wrong instrument. They do not explain why people who had the facts in front of them failed to act on them — which is the failure the Inquiry actually found, and which the College itself has since accepted without qualification.

How it was commissioned

The medical director who did not want to call the police chose the reviewer, the scope and the timing.

Ian Harvey emailed the RCPCH on 28 June 2016 asking whether they reviewed "individuals practice or for departments where there are concerns".

He confirmed in evidence that the review was his suggestion. The College's Head of Invited Reviews, Sue Eardley, summarised the case in her own briefing note: "Outlier for neonatal deaths over last 12-18 months. Done a thematic review and nothing highlighted — no pattern."

At some point before the team was assembled, Harvey did tell her about the doctors' suspicions of deliberate harm and about a nurse being on shift for the deaths.

Her evidence was that he did it "with almost a passing remark", and that it was "not intimated to me that that was a serious allegation taken seriously by the Medical Director".

Asked whether she thought he considered it serious, she gave the answer that recurs throughout this inquiry: "I think if he had thought it was a serious allegation he would have called the police sooner." And so, in her words, she did not think it was serious either.

The Inquiry's verdict on the setting-up is blunt: "The Terms of Reference were agreed too quickly. Haste led to predictable results, including an assertion by Mr Harvey that a high powered team had not called out a criminal act, when they could never have done so."

Who went, and what they were not told

The reviewers arrived without the one fact that would have changed what they did. The team that visited on 1 and 2 September 2016 was Dr Milligan as lead reviewer, a retired consultant paediatrician and neonatologist; Dr Graham Stewart as second reviewer; Claire McLaughlan as lay reviewer; and Alexandra Mancini as nursing reviewer, nominated by the Royal College of Nursing.

Sue Eardley supported the team and ran the project. Dr David Shortland was the service's clinical lead, based at the College rather than on the visit.

Eardley accepted that she had not shared her knowledge of the consultants' concerns with the reviewers. The College's president, Professor Stephen Turner, said in evidence that she should have; the Inquiry agreed, noting that it "would have led to an early discussion of whether an invited review was appropriate at all".

Dr Milligan had nonetheless seen a table of staff on duty at the time of eleven deaths, and observed that one nurse was present for all but one or two of them.

They were told again, in person, and still did not say the obvious thing. At the feedback session on 2 September the team advised management to involve the police if there was suspicion of criminality — which, as the Inquiry observes, is odd, "since he knew there was suspicion of criminality — all the RCPCH review team knew that; and Dr Stewart's own notes refer to 'foul play'".

Dr Milligan recalled asking Harvey whether he thought the police should be involved, and being told that he preferred to handle it internally for now.

Neither reviewer pressed him. The Inquiry: "All that was necessary was to say we think you should call the police."

What they did with what they heard

The letter that followed downgraded seven consultants into one complainant. Eardley's letter of 5 September 2016 described the steps taken against the nurse as having been taken "on the basis of an allegation made by one member of medical staff, supported by his medical colleagues".

The Inquiry is unsparing: "There was no excuse for the RCPCH downplaying the nature of the concerns in this letter … Even the reference to 'one member of medical staff' understated the position and experience of the Lead Clinician for the unit and ignored the fact that seven consultant paediatricians were expressing concerns."

Then it recommended a human resources investigation. Under the heading 'Action required — HR Investigation', the letter told the trust to formalise what it was already doing with the nurse.

The Inquiry's finding: "To recommend an HR investigation into concerns that babies were being deliberately harmed and dying was misconceived … It was worse than inappropriate, it was egregious."

The College accepted that it was wrong. Asked why the team had gone down the HR path rather than recommending the police, the lay reviewer "could not explain"; asked "why not call the police first?", she "was not able to provide any satisfactory answer".

The second recommendation — a forensic case note review covering every death, including details of all staff with access to the unit in the four hours before each one — asked doctors to do police work.

Both Eardley and McLaughlan accepted in evidence that this was a matter for the police. It became, as the Inquiry notes, "something else to be done instead of calling the police". More than two months had already passed since the deaths of Baby O and Baby P.

And in the report itself, the doctors' case was recorded in the language of their opponents. The confidential report said the consultants' allegation rested on the nurse being on shift each time an infant died, "combined with 'gut feeling'", and that there was "no other evidence".

The phrase "gut feeling" was the nurses' description of the doctors' suspicions; by adopting it the review team, in the Inquiry's words, "undermined the seriousness of the doctors' concerns". It also recorded, wrongly and without asking the doctors, that the consultants had refused supervised practice and required the nurse's removal.

What the report said, and what it was used for

It found staffing problems, and that is what the hospital wanted. The review found a shortage of consultants and problems with staffing, while acknowledging that nursing levels were closer to target than the network average.

Asked whether the unit had clear and engaged leadership and good team working, the answer given was: "Yes, generally but there were some areas where communications could be strengthened." The final report carried 21 recommendations.

There were two versions: a confidential report and a redacted "dissemination" version, published on 8 February 2017, from which the material about the nurse had been removed. In the dissemination version, under the heading 'Findings', there are no findings at all — only recommendations.

The decisive sentence is the Inquiry's own. "The RCPCH took on a task for which it was ill-equipped, the product of which was to be relied on much later as exonerating Letby when it did nothing of the sort."

And, on the reports themselves: "When the dissemination report was eventually published, none of its findings explained any of the deaths; nor did the findings in the confidential report."

A review that cannot look for deliberate harm, does not look for it, and says nothing about it, was read as evidence that there was none. That is how a false negative works. It does not lie. It answers a different question, and lets other people treat the answer as the one they wanted.

The families, and the timing

The parents received the version with the concerns taken out. One mother told the Inquiry that a letter arrived by black taxi half an hour before the report went online: "I was absolutely mortified … it beggars belief."

Another said: "We thought the report was going to give us some answers … There was no answers in it at all." A third said she felt they had been "kept in the dark", given "a whitewash gloss-over of a report".

The Inquiry's finding is one line: "The families were failed by the Countess and by the RCPCH, from the commissioning of the report to its publication."

No baby was harmed on the unit after the review, and that is not to the review's credit. The nurse had been moved to non-clinical duties in July 2016, before the reviewers arrived; the convictions relate to June 2015 to June 2016.

What the review bought was time for the hospital: the police were not called until May 2017, nearly a year after the move, and in the meantime the report was used to press the consultants to apologise and to contemplate a return to the unit. Delay was the harm.

The part that deserves credit

When it was finally examined, the College did the thing almost nobody else in this story did. Professor Turner opened his evidence by apologising to the parents for the College's part in delaying their finding out what had happened, and to the Countess paediatricians for how their professional body had treated them.

The Inquiry records: "Unlike other organisations, the starting point for the RCPCH was an uncomplicated recognition that their review team had made serious mistakes, for which the RCPCH acknowledged accountability and responsibility, including its contribution to the delay in the police being called to investigate. The apologies were genuine and reflective. The changes made since then by the RCPCH have been far reaching."

The service had already been examined once. A review commissioned by the College in 2021 found that it rarely turned work away and had no risk-based approach to deciding which requests to accept — findings the Inquiry says the evidence strongly supported.

That is the honest shape of this case: a body that got something badly wrong, and then, unusually, owned it.

Who actually held the line

The external reviewer did not stop this. Seven consultants did, at considerable cost to themselves. After the review reported, the doctors who had raised the alarm were required to apologise to the nurse they had reported and to take part in mediation with her, and, as the Inquiry records, "reference was made to avoiding a GMC referral".

They were refused sight of the report they had been promised until after the trust had made its public statement — a refusal the Inquiry calls "unreasonable and unprofessional".

They wrote to the chief executive on 30 January 2017 agreeing to apologise and asking, reasonably, what the board believed had caused the increase in deaths. They got no reply for more than a fortnight.

And they did not stop. One of their number, Dr McGuigan, wrote in an email read out at a meeting in March 2017: "My opinion is that this can never be investigated properly without a police led investigation … I have read the RCPCH service review and I don't see anything that explains the high mortality rates."

That is a working paediatrician reading the sheriff's report and saying, correctly, that it answers nothing.

On 27 March 2017 they said it to the chief executive's face, with NHS England in the room. The regional team's Ms McCabe described the meeting: "There was a high level of anxiety … I remember the clinicians trembling, my recollection was that the clinicians felt like they were putting their careers on the line, but they were prepared to do this given the seriousness of the matter."

She recorded that when the clinicians confirmed they suspected criminality, "Tony Chambers said that if that was the case then they should call the police. This sounded like an ultimatum."

Dr Brearey's account is shorter: "[W]e directly told or asked the Chief Executive Tony Chambers to go to the police."

After that it took outsiders who were not being paid by the trust. On 27 April 2017 the chair of the Pan-Cheshire child death panel and Detective Chief Superintendent Nigel Wenham met the medical director and two of the consultants.

When the conversation turned to the rotas and the one member of staff on shift for each collapse, the panel chair remembered thinking: "What, what are we being told here? This, this is gravely concerning."

It became "clear that the reviews that had taken place so far hadn't ruled out anything untoward", and the detective said it was a matter for his officers. A barrister instructed by the trust advised that if reasonable people were concerned, the police should be consulted — advice the medical director did not pass on in full when he described the opinion to others.

The chief executive then wrote to the Chief Constable of Cheshire, and meetings followed on 5 and 12 May 2017.

Even then the trust's instinct held: after the second meeting the medical director emailed that the police were "minded not to hold an investigation", and described the consultants' document listing their concerns as "a very prejudiced view, effectively pointing the finger at one nurse".

The recipient recorded that she "was very concerned", and asked him to make sure the police spoke to the two consultants before deciding anything.

The police met Dr Brearey, Dr Jayaram and Dr Holt on 15 May 2017, and announced an investigation the same day.

That sequence is the argument of this annex in miniature. The paid external review produced a document that was used to close the question.

What reopened it was people with no financial relationship to the trust — two doctors who would not sign off a return to the ward, a regional NHS team, a child death panel and a detective — and, before them, clinicians who took the professional risk of saying the same thing over and over for eleven months.

The system did not correct itself. It was corrected by people who refused to stop being a nuisance, which is the pattern this whole document keeps finding.

The counter-case: Durham, where the outside check did work

County Durham is the same story with the opposite ending, and it is not a happy one either. There the external check did eventually do the job it exists for. It took thirteen years.

The first warning was structural. In 2012 the General Medical Council agreed to withdraw junior doctor training placements from the breast surgery service; the Aubrey review records that this was "attributed to inadequate supervision and poor clinical governance", a judgement the Royal College of Surgeons confirmed thirteen years later.

Then the clinical networks began to see what the trust could not. In 2018 the Northern Cancer Alliance flagged the trust as a regional and national outlier on re-excision and benign lesion excision rates.

In the same year the Care Quality Commission raised formal concerns. In 2019 the national Getting It Right First Time programme reported low rates of reconstructive surgery and high re-excision rates.

On the 2018 alliance warning, Aubrey's finding is one sentence: "Although this was formally flagged and comparative data was available, no real action was taken to modernise clinical practice, review multi-disciplinary team (MDT) pathways, or address the identified variation."

The people who would not let it go were clinicians in other organisations. In March 2021 two external consultant oncoplastic surgeons raised concerns at the Newcastle breast MDT about the high operative rate for benign cases — including surgery being performed on normal breasts, and "the potential risk of being perceived as complicit should a future incident arise". They escalated formally to the cancer alliance.

In the same month the executive medical director of another trust formally raised two Durham patients whose proposed surgery, on second opinion, was "inappropriate" and "below nationally accepted standards".

Neither episode was escalated through Durham's own governance structures, and neither was reported through its incident process at the time.

When the alliance finally looked at the operations themselves, it found what the doctors had been describing. In May 2024 it led an external review of breast surgery carried out in 2023. Its peer review report of June 2024 identified care falling "significantly below expected standards", including unnecessary surgical procedures that caused physical and psychological harm, and clear non-compliance with national guidance.

The trust did not start patient safety investigations or the duty of candour process until February 2025 — nine months later, and only after an index case of metastatic breast cancer and another delayed diagnosis. That was the point at which it commissioned the 2023 look-back.

And the royal college did what a royal college should do. The Royal College of Surgeons carried out an invited service review of breast surgery, breast radiology and breast non-surgical oncology from 15 to 17 January 2025.

Its informal feedback alone — before any report — raised concerns serious enough that it recommended immediate external support, which is why the trust commissioned the independent governance review in March 2025. The College's final report, on 11 April 2025, confirmed longstanding governance failures.

That is the same instrument that failed at Chester, used properly: a college that said what it saw, early, and said it loudly enough to force action.

Set the two cases side by side and the design lesson is exact. At Chester the college was handed the one fact that mattered and did not act on it. At Durham the college acted on what it found, and the delay came from everything upstream of it: nobody acting on the 2018 outlier data, concerns from other trusts never reaching the governance structures, nine months between a peer review finding avoidable harm and the first duty of candour conversation.

It also matters who did the finding. Not the regulator — though the CQC raised formal concerns in 2018 and went on rating the trust well-led — and not the board. It was other clinicians: a cancer alliance, a national improvement programme, surgeons in Newcastle who could see across organisations while the trust could only see itself.

And there was money running the other way. Substantial public funds went to private-sector contracts for breast care — Spire Healthcare Ltd and Durham Surgical Services LLP, about £1.8m ($2.4m) a year in 2024/25.

Aubrey's finding on the conflict is blunt: "The consultant benefitted from additional LLP activity while operational leaders who were under pressure to meet performance targets were responsible for managing these contracts."

The review also records nine breaches of the trust's Standing Financial Instructions and a £761,555 ($1.0m) loss on the LLP.

The transferable rule is about sight-lines, not about better reviewers. No single organisation can see that it is an outlier. Only a network can. Durham's failure ran for thirteen years inside a trust that was looking at itself, and ended when people outside it compared it with everywhere else — which is an argument for funding the comparison, publishing it, and requiring a board to answer it in public. 2 · The inquiry record · Annex 17 · Holding the line

A note on sourcing for this section. All of it is from the Aubrey review (the independent external governance review of breast surgery services at County Durham and Darlington NHS Foundation Trust, commissioned March 2025), with quotations as that report gives them.

Figures circulating elsewhere — including a total of about £6m paid to the private clinics — could not be verified in the report and are not used here.

Arthur Andersen: the same failure, in another industry

A firm can be founded on independence and still lose it, and the loss looks like success while it is happening. Arthur Andersen founded his accounting firm in Chicago in 1913, and it was built on a phrase he took from his mother: "Think straight, talk straight".

The story the firm told about him for decades — and it should be labelled as that, a story circulated after his death in 1947 — has a company president flying in to demand that an audit report be changed, and Andersen replying that "there is not enough money in the city of Chicago to induce me to change the report".

He lost the client and kept the reputation. The principle he is credited with is the one that matters: the auditor's duty runs to the people relying on the accounts, not to the management paying the fee.

That principle is not folklore.

The US Supreme Court put it in 1984: the independent auditor "assumes a public responsibility transcending any employment relationship with the client … This 'public watchdog' function demands that the accountant maintain total independence from the client at all times, and requires complete fidelity to the public trust."

What changed was not the code of ethics. It was what the firm rewarded. Consulting grew far faster than audit through the 1970s and 1980s.

By 1998 the firm's US managing partner had introduced what he called the "2X" strategy: a partner should bring in twice their own practice revenue in work outside it, so that an auditor billing $2m for policing a company's books should sell $4m of other services to it — and partners were measured against it in their performance reviews.

The auditors became salesmen to the people they audited.

The consequences arrived before Enron did. Settling with the Securities and Exchange Commission over Waste Management in June 2001 — a $7m penalty, the largest then levied on a major accounting firm — the SEC's enforcement director said: "Arthur Andersen and its partners failed to stand up to company management and thereby betrayed their ultimate allegiance to Waste Management's shareholders."

A private settlement over Sunbeam followed, and another, of $217m, with investors in the Baptist Foundation of Arizona.

Then Enron. In 2000 Andersen earned $52m from Enron, of which only $25m was for the audit; the rest was for other services. The firm also provided Enron's internal audit — so it was, in substance, reviewing its own work — an arrangement a Senate committee singled out afterwards.

In October and November 2001 documents were destroyed on a large scale until, on 9 November, an email went round: "Per Dave — No more shredding … We have been officially served for our documents."

The firm was indicted in March 2002, convicted of obstruction of justice that June, and told the SEC it would stop practising before it by 31 August 2002.

A firm of 85,000 people worldwide, 28,000 of them in the United States, was finished within months.

The ending is the part people forget. In May 2005 the Supreme Court overturned the conviction unanimously, holding that the jury instructions "simply failed to convey the requisite consciousness of wrongdoing".

It was too late to matter. And the legislative response, the Sarbanes-Oxley Act of 2002, did not ask auditors to try harder: it created the Public Company Accounting Oversight Board, ended self-regulation of public-company audit, banned nine categories of non-audit service — including internal audit outsourcing — for audit clients, required audit committees to approve anything else in advance, and forced partner rotation.

It changed the conditions rather than exhorting the people, which is the argument of the annex on enablership applied to a different profession. Annex 14 · Healthcare: why the Art of Enablership is critical

What the two cases have in common

In both, the reviewer's scope, fee and briefing came from the organisation being reviewed. At Chester the medical director chose the reviewer, framed the question, agreed the terms of reference in haste, and described the concerns to the College in terms that made them sound like one doctor's hunch.

At Andersen the client bought the audit and much else besides, and the partner's own income depended on selling the "much else". Neither arrangement requires anybody to be dishonest for the outcome to be predictable.

In both, the failure was invisible at the time, because a clean report looks the same as a clean organisation. That is what makes a failed external review more dangerous than no review at all. It converts an open question into a closed one, and it does so with the authority of a name that nobody in the building is willing to argue with.

And in both, the people who did the right thing were inside, junior to the problem, and paid a price for it. Chester's consultants were told to apologise. Andersen's own Professional Standards Group, according to one legal analysis, was overruled by the client on technical questions, with a member barred from working on the account for being "too rule-oriented".

What would have to change

Five rules would have caught both cases, and none of them is expensive. First, the organisation under review does not get to write the question alone: terms of reference for any review of unexplained harm should be agreed with someone outside it, and should state in terms what the review cannot detect.

Second, every external review needs a stop rule — if the reviewers hear an allegation of deliberate harm, the review pauses and the matter goes to the police or the designated safeguarding officer that day, whatever the terms of reference say.

Third, the reviewer's duty runs to patients, not to the commissioner, and should be written into the engagement letter in those words.

Fourth, reports go to the people affected in full, and if a redacted version exists the existence of the other one is disclosed.

Fifth, somebody checks, a year later, what was done with the report — the Countess review had no follow-up that made any difference, and Andersen's failures were visible in the prior settlements to anyone who read them.

The deeper point is the one about incentives, and it is uncomfortable. As long as the reviewer is chosen and paid by the organisation it reviews, independence depends on the reviewer's character.

Sometimes character is enough. Chester and Andersen are what happens when it is not, and the answer in both is structural: change who sets the scope, who the duty runs to, and what happens automatically when the reviewer hears something the review cannot handle.

A note on sourcing. The Chester account is taken from the Thirlwall Inquiry report, Volume I, Chapter 12 and Volume II, Chapters 19, 20, 23, 25 and 26, with all quotations as the report gives them. The review team is named as the report names it — Dr Milligan, Dr Graham Stewart, Claire McLaughlan and Alexandra Mancini, with Sue Eardley supporting and Dr David Shortland as the service's clinical lead; other lists of names circulating online are wrong, and the wording of the consultants' letter of apology is not quoted here because it could not be verified at source.

The Andersen material comes from the Supreme Court's opinions in Arthur Andersen LLP v. United States (2005) and United States v. Arthur Young & Co. (1984), SEC press releases of June 2001 and June 2002, the Senate Governmental Affairs Committee's Financial Oversight of Enron (2002), the Sarbanes-Oxley Act 2002, and contemporaneous reporting for the "2X" strategy.

The founder's Chicago reply is firm lore rather than documented fact and is presented as such; the claim that it involved a railway is not supported by the best source.

Holding the line

Every failure in this document had a moment when somebody could have stopped it, and in every case the cost of stopping it landed on the person who tried.

Holding the line is what happens when someone meets that moment and does not look away — while their job, their training number, their reference, their mortgage and their family are all quietly arguing for silence.

The one thing to take away: holding the line is a choice, it is almost never held alone, and the people who hold it get targeted, worn down and sidelined — which is why champions need champions.

The moment, and what is on the other side of it

Holding the line is not heroism, and treating it as heroism is part of the problem. A hero is by definition unusual, which lets everybody else off. What the record actually shows is ordinary people, frightened, making the same choice over and over for months, mostly without recognition and often at real cost.

The pressure that makes people look away is rarely dramatic. It is personal, and it is legitimate: the mortgage, the training number, the reference you will need, the colleagues you have to work beside tomorrow, the partner who has asked you not to make yourself a target again. Nobody should pretend those are small things to risk.

And every alternative to holding the line has a respectable name. Wait for more evidence. Escalate it and let others decide. Follow the process. It is not your patient, not your ward, not your department. Keep your head down. Revert to type. Each sounds reasonable on its own; together they are how a breach becomes normal, step by acceptable step. 5 · Incivility and the continuum

What makes a line a line is that holding it costs you something now and letting it go costs you nothing now. The cost of letting go arrives later, and it lands on somebody else — usually a patient, sometimes a baby, always someone who never knew the choice was being made.

That asymmetry is the whole design problem. An organisation that leaves holding the line to individual courage has decided, in effect, to pay for its safety out of the personal reserves of whichever member of staff happens to have the most to lose.

Chester: seven consultants, eleven months

The clearest case in this document of a line held is also the case where the institution did everything it could to make the holders stop. The paediatricians at the Countess of Chester raised concerns from June 2016.

The hospital bought an external review, and the review's reassurance was then used to close the question down.

After it reported, the doctors who had raised the alarm were required to apologise to the nurse they had reported and to take part in mediation with her, and, as the Thirlwall Inquiry records, "reference was made to avoiding a GMC referral".

They were refused sight of the report they had been promised until after the trust had issued its public statement — a refusal the Inquiry calls "unreasonable and unprofessional".

The chief executive's own note of May 2017 sets out what was planned for them: "1. GMC 2. Actions from grievance 3. Mitigation from SoS/whistleblowing 4. Action plan to manage out".

They did not stop. One of them, Dr McGuigan, wrote in an email read out in March 2017: "My opinion is that this can never be investigated properly without a police led investigation … I have read the RCPCH service review and I don't see anything that explains the high mortality rates."

On 27 March 2017 they said it to the chief executive's face, with NHS England in the room. The regional team's Ms McCabe described the meeting: "There was a high level of anxiety … I remember the clinicians trembling, my recollection was that the clinicians felt like they were putting their careers on the line, but they were prepared to do this given the seriousness of the matter."

Dr Brearey's account of the same period is one sentence: "[W]e directly told or asked the Chief Executive Tony Chambers to go to the police." The police met Dr Brearey, Dr Jayaram and Dr Holt on 15 May 2017 and announced an investigation the same day.

Two details matter for anyone tempted to file this under exceptional courage. There were seven of them, not one: the line was held by a group, and a group is a different thing from a hero. And they held it for eleven months, through an apology they should never have been asked to make, which is the part that wears people down.

The full sequence, and the review whose reassurance was used against them, are set out in the two annexes on this case. Annex 16 · When the sheriff gets it wrong · Annex 1 · Organisational failure

And sometimes the people who hold the line are not even in the organisation. At County Durham it was two consultant oncoplastic surgeons in Newcastle who would not let go of what they were seeing at a shared multidisciplinary meeting in March 2021 — operations on normal breasts, and the risk, as they put it, of "being perceived as complicit should a future incident arise".

They escalated it formally to the cancer alliance. In the same month a medical director at another trust put two Durham patients in writing, whose proposed surgery a second opinion had found inappropriate.

Neither had any authority over Durham. Both used the only thing they had: a willingness to say it formally, in their own names, about somebody else's service. It was another three years before the alliance reviewed the operations themselves. Annex 16 · When the sheriff gets it wrong

Bevan: three years of it, against his own side and against the profession

The NHS exists because one minister held a line for three years while the people he needed most told him to give way. The first pressure came from inside his own government. On 5 October 1945 Bevan put his hospital plan to Cabinet; on 12 October Herbert Morrison, Lord President of the Council, circulated a memorandum against it, admiring "a brilliant and imaginative paper" while opposing the removal of hospitals from local authorities, because "if we wish local government to thrive — as a school of political and democratic education as well as a method of administration — we must consider the general effect on local government of each particular proposal". On 18 October the Cabinet approved Bevan's approach in principle.

In December 1945, with the profession threatening not to take part, The National Archives records that some of Bevan's Cabinet colleagues suggested this might delay the launch. He was not moved: he would, he said, simply "squeeze them out" with better services, and the Archives' account quotes him telling colleagues, "Here is our chance to do something big".

Then the profession balloted against him, repeatedly. The British Medical Association held a series of plebiscites; the one announced on 21 February 1948 was reported by the BMJ under the headline "38,534 Doctors Disapprove", an 84 per cent poll in which, in the Journal's own words, "90% of all those voting" rejected the Act.

Bevan answered in the Commons on 9 February 1948: "From the very beginning, this small body of politically poisoned people have decided to fight the Health Act itself and to stir up as much emotion as they can in the profession."

He closed: "I hope the House will not hesitate to tell the British Medical Association that we look forward to this Act starting on 5th July, and that we expect the medical profession to take their proper part in it."

A fourth plebiscite followed in April. The BMA had set its own threshold — if the majority against included approximately 13,000 general practitioners it would go on advising doctors "not to enter the National Health Service on July 5".

The threshold was not met. On 28 May 1948 a Special Representative Meeting accepted the Council's recommendation to co-operate, and the service opened on the appointed day.

And he conceded, which is the point. On 7 April 1948 Bevan told the Commons that a whole-time salaried service would not be brought in by regulation and would require fresh legislation, and offered new entrants a fixed element alongside capitation.

He gave way on the mechanism and not on the principle: free at the point of need, funded collectively, starting on 5 July. Holding the line is not refusing to move. It is knowing in advance which one thing you will not trade, and trading almost anything else to keep it.

Two corrections, because this story attracts embellishment. The line "I stuffed their mouths with gold" survives only as a later recollection — it is usually sourced to Brian Abel-Smith's The Hospitals 1800–1948 (1964), and is elsewhere placed at a dinner in 1955 or 1956 — and there is no contemporaneous record of it.

And the Treasury is often said to have fought him: what the record shows is Hugh Dalton's Budget of 9 April 1946 setting out the money and describing "the bold and constructive National Health Service Bill of my right hon. Friend the Minister of Health". The pressure Bevan had to hold out against before the appointed day came from his own Cabinet and from the profession, not from the Exchequer. The Treasury squeeze — ceilings, and the power to charge for prescriptions in 1949 — came afterwards.

Chris Turner: the line held, and the cost landed

The third case is the one that shows what happens when someone holds the line and nobody holds it with them. Dr Chris Turner joined the emergency department at Mid Staffordshire in 2007 as a junior doctor and gave evidence to the public inquiry.

He described a "culture of bullying and harassment", staff "utterly demoralised", nurses threatened "on a near daily basis" with losing their jobs, and "no significant medical leadership".

He paid for it in the ordinary currency. "First I stopped sleeping and eventually eating"; he lost two and a half stone in nine weeks, slept three hours a night, and "after two years I broke under the pressure and resigned".

He went on to co-found Civility Saves Lives, and his single sentence about that hospital is the shortest statement of this document's argument available: "I worked in mid staffs, and in mid staffs the exec team was measured on finance, and pretty much nothing else."

The person best placed to describe what that culture does to clinical work is someone the department removed. That is not a happy ending. It is a warning about what an organisation loses when the people who hold the line have to do it alone.

Champions need champions

People who hold the line get targeted, worn down and sidelined, and the evidence is in this document's own numbers. There were 1,519 Freedom to Speak Up guardians across 820 organisations in March 2026, most doing the job alongside another one and, until very recently, all appointed by the trusts they were meant to hold to account.

The National Guardian's Office had 16 staff and, in the Thirlwall Inquiry's words, "neither the power nor resources to hold hospitals or other healthcare settings to account". It closed on 30 June 2026.

In the 2025 NHS Staff Survey the speaking-up score fell to a five-year low, with only 55.49% of staff confident their organisation would address a concern about clinical safety. Half of doctors surveyed report moral injury.

Chester's consultants were made to apologise; Mid Staffordshire's witness resigned nine weeks into losing two and a half stone.

So the line cannot be a solo act — and the three things that make it holdable are the three the Champions Pathway is built on.

Choice. Nobody can make the decision for you, and no policy can make it for you either. Under pressure a person still chooses whether to see the patient in front of them as a human being, and whether to say the thing that will cost them something. That capacity can be taught, practised and grown.

Connection. Seven consultants, not one. People who know what you know, who will say it alongside you, and who are not employed by the organisation you are challenging. That is the whole argument for circles that belong to a network beyond the employer: a route that management cannot close, used to call it forward when something is being done well and to call it out when it is not.

Creation. Objecting is not enough, because an objection can be absorbed. Something has to be built: a record, a route, a change, a service. Chester's consultants wrote their concerns down and kept a document; Turner built a campaign; Bevan built an Act of Parliament. The built thing is what survives the person.

And somebody has to hold the line for the people holding the line, which is what "champions need champions" means in practice. Not a mentor for their career, and not a case manager for their concern, but someone whose responsibility is the person: outside the chain of command, in contact while it is happening, and still there afterwards.

Every case in this document shows what its absence costs, and none of the reforms now being legislated provides it.

What has to be built, so that the line is not held by luck

Turn holding the line from a permission into an obligation. Where it works, staff are not allowed to look away: at Virginia Mason anyone who sees something likely to harm a patient must stop the activity and report it; under Martha's Rule a patient, family member or member of staff can call for an urgent review by a team that is not the one treating the patient.

An obligation protects the person who acts, because they are doing what the rules require rather than making themselves conspicuous. Annex 15 · Humanity in healthcare: examples

Write the line down in advance. An honest mistake, reported promptly, is protected. Silence, downgrading an incident and retaliation against the person who raised it are not. People hold a line they can see; nobody gambles their career on a value statement. Annex 13 · Radical leadership

Give concerns and praise a route that no executive can close. The evidence on speaking up is that the internal route fails exactly when it is most needed, which is why the circle has to belong to a network outside the employer. Annex 14 · Healthcare: why the Art of Enablership is critical

And judge leaders on it. The question is not whether the chief executive says the right things about candour. It is whether anyone below board level can stop an unsafe process without asking permission — and what happened to the last person who did.

And the closing point is the simplest one. Holding the line is a team effort. It takes the choice, which is yours alone; the connection, which is other people; and the creation, which is what you build so that the next person does not have to be as brave as you were.

A note on sourcing. The Chester material is from the Thirlwall Inquiry's report and the evidence it records, and is set out in full in the annexes on organisational failure and on the external review.

Bevan: Cabinet papers of October 1945 (CAB 129/3) as quoted in Powell and Greener, Social Policy & Administration (2024); The National Archives' account of the foundation of the NHS for December 1945; the British Medical Journal of 21 February, 1 May and 5 June 1948 for the plebiscites; Hansard for 9 February 1948 (vol. 447), 7 April 1948 and the Budget statement of 9 April 1946.

Chris Turner's evidence is from the Mid Staffordshire public inquiry and his own published accounts. Freedom to Speak Up figures, the 2025 NHS Staff Survey and the moral injury finding are sourced in the volumes.

You cannot regulate people to care, but you can teach them

The central argument of this document is made elsewhere. This annex illuminates it from one angle. It starts with a thought experiment, and then asks who, after the Thirlwall Inquiry, was able to say they had got it wrong — and who could not. The pattern is old, it is not peculiar to healthcare, and there is a body of practice that answers it.

The simple truth is you cannot regulate people to care. It has to come from within. But you can teach and enable this capability, within individuals and a wider community that is passed down through generations.

In fact, this is the origin story of the NHS. It is how a small community in the Welsh valleys established a system, a process, a community built on human and moral values.

Other sectors have learnt the hard way how to overcome the natural human inclination towards self-protection, arrogance or top-down, hero-centric leadership — typically the pathway that leads to cataclysmic errors. Aviation and the nuclear industries are good examples, where the cost of failure is too great.

As we've seen elsewhere, the processes themselves are seldom replicable, but the principles can be applied.

The key lesson is the need for brutal honesty, rooted in proximity — closeness to those involved. Distance, remoteness from the day-to-day reality, from the effect on those involved — staff and patients — only serves to numb the normal human instinct to act.

Start with a thought experiment: imagine it had happened in Tredegar

Before any of the evidence, picture the same thing happening in the town that invented the idea.

From 1890, the workers of Tredegar paid a few pence from every pound into a medical aid society, so that no neighbour faced illness alone. By the 1920s it covered some 20,000 people. Their money, their doctors, their committee.

Now imagine fourteen babies harmed in the service that town had built for itself. Seven of them dead.

The families live on the same streets as the committee. The men who run the fund pass the bereaved parents at the chapel, at the school gates, in the queue at the shop, at the pit head on Monday morning.

Now run the same experiment inside a family.

Now imagine a child in your own family harmed, tragically, and then imagine learning that adults in the house had been told twice and ignored the warnings.

The answer to both would be deep soul searching. What, why, who, what could, what should I have done to prevent it?

There would be deep compassion for the bereaved and the injured, those who have to carry the pain of loss and injustice every moment, every day.

Back to the present, to real life.

The point is the normal, human reaction in the face of such terrible loss is exactly this, deep soul searching. But there is a near-total absence of it in the case of Chester.

These are the facts and feelings of many of those involved, directly or indirectly, although not all. Why?

There are three possible answers, maybe more:

1. Vested interests: the implications of facing the reality, the consequences is too costly.
2. Their own 'pain' is too dominant — people have become numb to the pain of others.
3. There is insufficient proximity, too much distance between daily reality and decision-makers.

The lack of soul-searching begs further questions. Whatever the reasons, it is indicative of a culture, a system that has lost or is losing its humanity.

Chester

What follows in this annex is the record of how little soul searching there was or is.

Not from everybody — the doctors who had been ignored were tormented by it, and one professional body apologised without conditions. But from the people who ran the unit and the hospital, almost none.

And when a group of decent people fail a child and then do not search themselves, that is not a character flaw in those particular individuals. It is a fact about the distance the system had put between them and the people they were failing.

The variable that changed between Tredegar and Chester is proximity.

Tredegar's committee could not get away from the people it served. A modern board can, and mostly does: the harm arrives as a data item, the family arrives as a complaint, and the person who raised the alarm arrives as an HR problem.

Every one of those is a form of distance. And distance is what makes it possible to be a good person and still look away.

Who could say they got it wrong, and who could not

The truest test of an institution's humanity is who is able to say sorry once the truth has landed.

On that test the Thirlwall Inquiry divides the people involved almost exactly along the line this document argues matters: not doctors against managers, but those who could still see a person and those who could only see an institution.

The people who showed the sharpest remorse were the ones who had been right all along.

Dr John Gibbs, one of the consultants who raised the alarm, was candid: "I regret that we or I didn't go to the police at this time after the thematic review."

On the meeting that failed to act, he reflected that the outcome "should have been a trigger that we bypass managers and went to the police. And we failed to do that."

That is remorse from the people who had already paid for speaking. Dr Stephen Brearey and Dr Ravi Jayaram "both received letters that made reference to referral to the GMC and both were criticised following the grievance procedure, to the extent they were required to apologise to Letby".

One professional body managed an apology without conditions, and the Inquiry said so.

Professor Stephen Turner, President of the Royal College of Paediatrics and Child Health, apologised at the start of his evidence to the parents, and to the Countess paediatricians for how their own professional body had treated them.

The Chair's assessment: "Unlike other organisations, the starting point for the RCPCH was an uncomplicated recognition that their review team had made serious mistakes, for which the RCPCH acknowledged accountability and responsibility, including its contribution to the delay in the police being called to investigate. The apologies were genuine and reflective. The changes made since then by the RCPCH have been far reaching."

The Nursing and Midwifery Council is the counter-example, and the tell was the language.

The Inquiry found the NMC "rightly acknowledged a number of mistakes and failures. However, these were most unfortunately undermined by the written closing submissions … which reverted to the language of things that 'could have been' done differently".

The Chair singled out the phrase "identified opportunities to learn and improve" as "[a] particularly egregious example of language avoiding responsibility".

That is the mechanism in one line. Institutional language is not a neutral medium. It is how an organisation avoids meeting the person it harmed.

Not one nurse on the unit said, with hindsight, that something had felt wrong.

The Inquiry approached 30 nurses involved in the care of the indicted babies, and sent questionnaires to 20 more neonatal nurses and to all 14 midwives. "None of the nurses or midwives had any suspicions or concerns about Letby's conduct in relation to harming babies while she worked on the neonatal unit."

"No incidents were reported by nurses." And: "Almost no one thought it possible that a nurse was harming babies."

One HR manager did what none of the clinical nursing staff did: she "accepted that there were safeguarding issues here and she should have spoken up".

What replaced curiosity was loyalty, and the Inquiry named it for what it is.

The unit manager described Letby as the "crème de la crème", told her grievance interview that she believed Letby "is 100% innocent", and wrote in a note for the grievance hearing that "we did not or never had believed for an instant that LL was guilty of ANY wrong doing".

The Chair: "I have no doubt that Ms Powell spoke up for Letby vociferously. Of course she did. She thought that the concerns were misconceived. She was very angry about them and very supportive of her nurse."

On 7 April 2016 the same manager moved Letby to day shifts, explaining the decision was "driven by a 'well-being approach'" because Letby "had been involved in so many of the recent deaths that that must have a profound effect on her well-being".

The Chair's reading of that decision is the whole problem in a sentence: "It is likely that part of the reason for moving Letby to days was to see whether anything changed — specifically, whether or not the deaths continued at night."

A hypothesis was being tested on a neonatal unit, and it was not written down as one.

The senior nurse who blocked the doctors was found to have lost her judgement entirely.

Of the divisional head of nursing the Chair wrote: "her hostile approach was deeply unfortunate, and it would continue for many months".

And, of an allegation she made against a paediatrician: "That Ms Rees was comfortable making such an extreme allegation against a paediatrician without any basis demonstrates that she had lost all judgement in the face of what she had been told."

She was later part of a WhatsApp group with Letby that carried over 750 messages between April 2017 and July 2018, and she accepted she "had become too close to Letby".

The executive lead for safeguarding did not refer the matter for nearly two years, and when she did the referral was wrong.

"The failures in safeguarding occurred at every level of the hospital … However, particular opprobrium attaches to Ms Kelly, for her repeated and serious failures to ensure appropriate safeguarding responses were initiated. As the Executive Lead for Safeguarding, she should have ensured that the concern about Letby harming babies was referred to the LADO and indeed to the police. She did not refer the matter until March 2018, and when she did, the referral contained inaccuracies and was misleading."

The case advanced on her behalf was that she had understood the consultants' concern to be pastoral — concern for Letby's welfare. The Chair rejected it outright.

And the chief executive told the suspect the institution was behind her.

At a meeting with Letby and her parents in February 2017 he said: "Lucy, don't worry, we have got your back."

He called it clumsy language used to de-escalate. The Chair: "To say 'we have got your back' is unambiguous. If it was to 'avoid any possible escalation', it was misconceived and dishonest. If he said it because he meant it, it was a gross misjudgement."

When he left, the Trust's announcement said his departure was "not a judgement on his ability as a CEO but more a reflection of his integrity as a leader". The Chair: "That was pure spin. He was stepping down to avoid a vote of no confidence. It had nothing to do with integrity and everything to do with the preservation of his career."

The Inquiry's own word for the thing underneath all of this is tribalism.

The report records a "them and us" approach between doctors and nurses, and that nurse managers accepted relationships "could be described as 'nurses against doctors'".

The Chair: "a well-known cultural phenomenon: sometimes referred to as tribalism, it is unthinking loyalty to one's team or profession. The senior nurses demonstrated unquestioning loyalty to Letby throughout."

And tribalism has a national-scale version, which the report also names.

"[I]t is inescapable that there is a culture of avoiding dealing with poor performance."

Failing managers were moved on with the active assistance of NHS England, in a process called rehabilitation. The chief executive himself said it is referred to as "the donkey sanctuary".

"This state of affairs is accepted with almost a collective shrug of senior shoulders. It should not be tolerated."

A collective shrug is not a policy failure. It is a group of people who have stopped being able to see the person at the end of the decision.

None of this is peculiar to the NHS

Every institution that has lost its humanity has done it the same way: the group became sacred, and the person in front of them stopped counting.

Puritan Massachusetts was founded by people fleeing religious persecution. On 1 June 1660 its magistrates hanged Mary Dyer, a mother of six surviving children, for returning to Boston as a Quaker after being banished on pain of death. They believed they were protecting the colony.

Eighteenth-century Britain produced parliamentary democracy, the Enlightenment and the world's largest slave trade at the same time. When at least 132 enslaved people were thrown from the Zong in 1781, the case that reached an English court was an insurance claim, argued over the value of lost cargo. Nobody was prosecuted for murder.

And in the 1930s the most scientifically advanced medical profession in the world became the profession most complicit in what followed, a history this document sets out in Volume 6.

The steps are always the same three.

The group, the institution or the doctrine must be preserved. Personal unease is overridden by policy. And human suffering is scrubbed out by language — "cargo", "opportunities to learn and improve", "vague clinical anomalies".

The Chester nursing team did not set out to shelter a killer. They set out to defend a colleague against what they took to be bullying, and the wall they built was made of loyalty, not malice. That is precisely why it worked.

What answers it is proximity

Prejudice and institutional coldness behave like a rock, and you cannot break a rock with more rock.

More argument, more regulation and more structural force meet the same certainty with a different accent. What disarms it is proximity — being close enough to a person that the label stops working.

The inscription at the Kigali Genocide Memorial puts the whole of it in five words: "If you really knew me…"

Every example in this document runs the same way round. The families who would not stop asking were proximity. The consultants trembling in a meeting room were proximity. The executives reading a curated board pack were not.

And proximity can be designed in

This is the part that answers the title: you cannot regulate people to care, but the capacity to act for the good of others can be taught, practised and built into how an organisation runs.

Work-as-imagined and work-as-done. Erik Hollnagel's distinction, published by NHS England itself, is the cleanest statement of the gap that killed the babies at Chester: "Work-As-Imagined is an idealized view of the formal task environment that disregards how task performance must be adjusted to match the constantly changing conditions of work", while "Work-As-Done … describes what actually happens". Boards were governing the first and being harmed by the second.

Psychological safety. Amy Edmondson's work established it as a measurable property of teams, not a mood. A unit reporting no concerns is not a safe unit; it is a unit where nobody speaks.

The patient story first. A number of NHS trusts now open the public part of every board meeting with an account from a patient or family, ahead of the finance and performance reports. It is a practice, not a legal requirement — and it is exactly the design choice that puts a person in the room before a spreadsheet is opened.

The nuclear industry did this deliberately, and it is the closest comparison there is

After its two defining catastrophes, the nuclear industry concluded that distance between the boardroom and the plant floor was itself the hazard, and built institutions to close it.

The Institute of Nuclear Power Operations was created by the US industry in 1979, on the recommendation of the Kemeny Commission after Three Mile Island. The World Association of Nuclear Operators followed in 1989, after Chernobyl.

What INPO does is proximity as a standing obligation rather than an aspiration. A court record describes it plainly: an evaluation team of about fifteen people spends two weeks at a plant and "includes both INPO personnel and peer evaluators from other nuclear plants", and INPO "relies heavily on observations of people at work and on the frank … feedback from working-level employees".

Staff from a competitor stand on your gantry and watch your people work. There is no board pack to hide behind.

And the industry inverts the burden of proof, which is the single change that would have saved those babies.

INPO's Traits of a Healthy Nuclear Safety Culture states that "[o]perators are vested with the authority and understand the expectation, when faced with unexpected or uncertain conditions, to place the plant in a safe condition", and that "[i]ndividuals stop when faced with uncertain conditions".

The governing principle is stated as a reversal: "A proposed action is determined to be safe in order to proceed, rather than unsafe in order to stop."

At Chester the burden ran the other way. The senior nurses and executives demanded proof of deliberate harm before acting, which is an inversion of how precautionary safeguarding is supposed to work. A nuclear operator with the same quality of evidence would have been expected to shut the plant down and be thanked for it.

What Thirlwall recommended, and what it does not reach

The seventeen recommendations are the strongest set any English health inquiry has produced, and they are almost entirely about detection.

Boards must monitor every death of a child or baby, with a predetermined route for escalating concerning trends. A named neonatal lead must report to the board at least every six months. Safeguarding training must cover all staff and board members, including non-executive directors, and must include how to deal with suspicions of deliberate harm by staff. In-cot cameras are to let parents watch their own baby remotely.

Every one of those is worth having. None of them requires anybody to be in the room with a person they have harmed.

There is no recommendation that boards must hear directly from families, and it is worth being honest that this document is asking for something the Inquiry did not ask for.

That is an argument made here, not a finding of the Inquiry. It is made because the record shows what detection alone achieves: the data existed at Chester, in a spreadsheet the unit manager built herself, and it changed nothing, because no one in authority was close enough to a dead baby's parents to be unable to look away.

So: you cannot regulate people to care

Regulation can remove a person. It cannot create the capability that was missing.

A barring scheme takes a chief executive out of the system. It does not give the next one the capacity, under pressure, to hear something unwelcome from someone junior and act on it anyway.

But that capacity is teachable, and the evidence that it is teachable comes from outside healthcare: from atrocity prevention, from active-bystander training, from Safety-II, from the nuclear industry's peer evaluations. It is taught, practised and structured — which is the whole of the proposition in this document.

Care is intrinsic to being human. It has to come from within. And we give from what we have received.

Which is why caring for the carers is not a soft addendum to this argument. It is the supply line.

Annex 14 · Healthcare: why the Art of Enablership is critical · Annex 13 · Radical leadership · Annex 17 · Holding the line · Annex 16 · When the sheriff gets it wrong

A note on sourcing. The Thirlwall material in this annex is quoted from the Inquiry's own report — Volume I, chapters 2, 3, 8, 9, 11 and 12; Volume II, chapters 14, 17, 18, 20, 22, 23 and 27; and Volume III, chapters 37, 39, 43 and 45.

The line of thought came from an AI research conversation, and a number of that conversation's specific claims did not survive checking and are therefore not used here: that a nurse said she was "a little bit mad" at Letby for hovering at a cot (the substance is in the report, the wording is not); that Karen Rees texted Letby "we'll continue to fight for you" or applied for a prison visiting order (not in the report); that Eirian Powell called the case "the most difficult situation I have ever had to deal with" (not in the report); that Sir Duncan Nichol admitted "a big personal failure" and apologised tearfully to a paediatrician (not in the report); that Dr Gibbs said he was "deeply ashamed" (the word does not appear — he said "I regret"); that the NMC apologised and was found to lack "professional curiosity" (the Inquiry's criticism is of its language, not an apology); that Thirlwall described "an escapable culture" (she wrote "it is inescapable that there is a culture"); that any trust is legally required to open board meetings with a patient story (it is a practice, not a law); and a quotation attributed to the World Association of Nuclear Operators, "a safety myth is a precursor to an accident", for which no source exists at all.

Mary Dyer was hanged on 1 June 1660 and is often described as a grandmother; she had six surviving children and no confirmed grandchildren at her death. Counts of the dead from the Zong range from 130 to 142; at least 132 is the safe figure.

Non-Thirlwall sources: Hollnagel, Wears and Braithwaite, From Safety-I to Safety-II: A White Paper (2015), published by NHS England; Edmondson, "Psychological Safety and Learning Behavior in Work Teams", Administrative Science Quarterly (1999); INPO 12-012, Traits of a Healthy Nuclear Safety Culture, as published by the US Nuclear Regulatory Commission; and the description of INPO evaluations from Institute of Nuclear Power Operations v. Cobb County Board of Tax Assessors (Georgia Court of Appeals, 1999).

Glossary

Plain-English definitions of the terms this document uses often — first the ideas, then the British institutions a reader outside the UK would have no reason to know.

Sterling figures throughout the document carry a US dollar equivalent at $1.35 = £1.

The ideas

Humanity-first healthcare. Care in which the humanity is built into the fabric of the organisation rather than left to the character of whoever is on shift. It serves one goal, taken from the NHS Constitution and widened to all care: the patient will be at the heart of everything.

The capacity to act for the good of others. What humanity means in practice in this document. Not a feeling and not a slogan, but the ability, under pressure, to see the person in front of you as a human being and do what is right for them.

Every person counts · Caring for the carers · It's who we are. The three commitments humanity-first healthcare adds: everyone is inside the circle of care; people treat others the way they are treated themselves; and humanity is an identity rather than a programme.

The Champions Pathway. The first of the three structural pillars: training and practising the capacity to act for the good of others, so that humanity is a taught capability rather than a trait some staff happen to have. Built on choice, connection and creation.

Champions Circles. The second pillar: small, self-governing groups of a handful of people who meet regularly and in confidence. The decisive feature is that the circle belongs to a network beyond the employer, so it can do what an internal committee cannot.

Calling it forward, and calling it out. The two jobs of a circle. Calling it forward means naming something being done well that nobody outside has noticed. Calling it out means naming something wrong when the internal route has failed. A group that can only do the second becomes a complaints channel; a group that can only do the first becomes a fan club.

Holding the line. The moment when the cost of doing the right thing lands on you personally — your job, your reference, your standing, your family's security — and you do it anyway. The Chester consultants held it for eleven months; Bevan held it for three years; and in every inquiry in this document the question is whether anybody held it.

Champions need champions. The people who hold the line get targeted, worn down and sidelined, so somebody has to hold it for them: a person outside the chain of command whose responsibility is the human being who spoke up, not the management of their concern.

The Four Stages of Humanity. Champions' description of the journey, after Broadwell's stages of competence: unconscious inhumanity, conscious inhumanity, conscious humanity and, at its best, unconscious humanity — where being human under pressure is simply another day at the office. Everyone moves between the stages, and pressure pulls people back.

Enablership. The third pillar and the name for the whole job: making an organisation able — building excellent care into the rules, the habits and the fabric so that it survives the departure of whoever built it. The phrase is Barrie Watson's, from The Art of Enablership (2026); this document applies it to healthcare.

Clock building and time telling. Jim Collins's distinction, from Built to Last (1994). A gifted person who can tell you the time is useful until they leave the room; a clock goes on telling the time after its maker has gone. A chief executive who personally holds the standard is telling the time. An organisation in which the standard holds without them is a clock.

Radical honesty, accountability, trust and forgiveness. The four commitments that make accountable leadership work as a set: the truth told early, responsibility that is never waived, people trusted to act, and an honest mistake survived. Take away any one and the other three fail.

Obligation, not permission. The design move that separates the places where this works from the places where it does not. Staff are not merely allowed to stop an unsafe process; they are required to. An obligation protects the person who acts, because they are following the rule rather than sticking their neck out.

Core function and form. From implementation research. The function is what a practice is for; the form is how it is done in a particular place. Copy the function, redesign the form. Nearly every failed import in this document copied the form.

Diagnosed, not treated. The shorthand for England's position: an unmatched capacity to investigate its own healthcare failures, and no matching capacity to prevent them.

When the sheriff gets it wrong. The failure of the outside check — the reviewer, auditor or regulator everyone downstream relies on. When their reassurance is wrong, the board, the regulator and the families all inherit it and act on it.

The continuum of destruction. Ervin Staub's finding that serious harm is reached in steps, each small enough to be acceptable after the last one, until behaviour that would once have been unthinkable is unremarkable.

Bright spots. Places that already do this well on ordinary budgets — Tredegar, Mayo, Nuka, Buurtzorg, Esther, Gundersen, East London, Northumbria, Bromley by Bow — read in this document as design examples rather than inspiration.

Ward-to-board. The loop by which something noticed at the bedside reaches the people who can act on it, and an answer comes back. The test of it is whether it works when the chief executive is not in the room.

The revolving door. A manager about whom serious concerns have been raised moving to an equivalent post at another organisation. The Thirlwall Inquiry heard "ample evidence" of it and recommended it be closed.

A prisoner cannot free himself. Rabbi Lord Jonathan Sacks's rendering of a Talmudic maxim (Brachot 5b), used here for the finding that twenty-five years of internally generated reform has produced the same failures in the same words.

Tredegarise. Bevan's own word for what he did in 1948: scaling up to a nation what the Welsh mining town of Tredegar had built for itself from 1890.

British terms, for readers outside the UK

The NHS. The National Health Service, founded on 5 July 1948: funded from general taxation, free at the point of need, and covering almost the whole population. There are four of them — England, Scotland, Wales and Northern Ireland — run separately. Unless stated otherwise, figures in this document are for England.

Trust, and foundation trust. The organisation that runs hospitals or community services in a given area, employing the staff and holding the budget. A foundation trust has more freedom over its money and answers to national regulators rather than to local government.

Integrated care board. Since 2022, the body that plans and buys health services for a population of one to three million.

The CQC. The Care Quality Commission, the independent regulator that inspects health and social care in England and rates it Outstanding, Good, Requires improvement or Inadequate. One of the things it rates is whether a service is well-led — which is why a trust rated well-led while patients were being harmed recurs in this document.

Section 29A warning notice. A formal CQC notice requiring significant improvement within a stated time. The step before enforcement.

NHS Oversight Framework, segment 4. The lowest of four segments, meaning an organisation is in the most intensive national support and scrutiny.

Royal college, and an invited review. A royal college is a professional body for a specialty — paediatrics, surgery, nursing. Its invited review service sells peer review to an organisation that asks for it: colleagues looking at a service and making recommendations. It is not an investigation and has no regulatory power.

GMC, NMC, HCPC, CIPD. The General Medical Council registers doctors and the Nursing and Midwifery Council registers nurses and midwives; both can strike a person off. The Health and Care Professions Council will run the new barring scheme for senior NHS managers. The CIPD is the professional body for human resources, which can expel a member.

Freedom to Speak Up guardian. A member of staff, usually doing the job alongside another one, appointed to receive concerns from colleagues. The National Guardian's Office oversaw them nationally until it closed on 30 June 2026.

Duty of candour. The legal requirement, since November 2014, to tell a patient or family when something has gone wrong, and to apologise. It applies to organisations; the inquiries have repeatedly asked for it to apply to individual managers too.

Public inquiry, independent review, inquest. A public inquiry is established by a minister under the Inquiries Act 2005 and can compel evidence on oath. An independent review is commissioned, often by the organisation being examined, and cannot compel anything. An inquest is a coroner's hearing into how a person died; a coroner can issue a report to prevent future deaths.

NHS Staff Survey. An annual census of NHS staff in England — 766,285 responses in 2025 — which is why this document can say how people are treated rather than guess.

Clinical negligence provision. The actuarial value of harm already caused and expected to be paid for: £60.3bn ($81.4bn) at 31 March 2026, second only to nuclear decommissioning on the government's balance sheet. It is a liability, not annual spending; the cash paid out in 2025/26 was £3.2bn ($4.3bn).

Adult social care and children's social care. Care at home, in care homes and for children in the care system. It is arranged and largely paid for by elected local councils rather than the NHS, is means-tested, and in England has no professional register for frontline staff.

Hospice. An independent charity, mostly funded by donations rather than the state, caring for people at the end of life. The best-rated sector in England.

Resident doctor, consultant, GP. A resident doctor — until 2024 called a junior doctor — is a qualified doctor still in training. A consultant is a senior specialist. A GP, or general practitioner, is a family doctor and the usual first point of contact.

A&E and the four-hour standard; the elective waiting list. A&E is the emergency department; the standard is that 95% of patients are admitted, transferred or discharged within four hours. The elective waiting list is planned, non-urgent hospital care.

Martha's Rule. Introduced across English acute hospitals from 2024: a patient, family member or member of staff can call for an urgent review by a critical care team independent of the one treating the patient.

Social prescribing, and link workers. Referring a patient to non-medical support in their own community — a walking group, a debt adviser, a housing caseworker — when what is making them unwell is not primarily medical. The link worker is the person attached to a GP practice who makes the connection.

Sources and accuracy

Research synthesis

What. Most of the material this work started from was AI chat transcripts, and several of its best lines were wrong. The three Thirlwall volumes, the Aubrey report and the Lord Mann review have since replaced the largest of those blocks with primary sources, read directly.

So what. Eight claims in circulation do not survive checking. The "61% reduction in cognitive capacity" (the paper reports 15.6%, and Porath's own published figures are 33% and 39% for targets, 22–28% for witnesses); "immune to the sound of pain" attributed to Chris Turner; Riskin's "20% drop" (15–17%) and the "50% reduction in information sharing" (not a finding at all); "30–50% of new nurses leave within two years" (American, and not traceable even there); the Mayos as Quakers; Fry inspiring Nightingale; Northern Ireland's "12–14% overspend"; and the Averroes quotation.

Now what. Read this volume before quoting anything externally. Where a claim is softer than its reputation, the better version is given here — usually a mechanism or a mediation finding rather than a headline percentage, which is stronger evidence anyway.

Source notes and accuracy flags

Unverified material is included in this document where it is useful, but flagged. This volume is that flag — read it before quoting anything externally.

Independently verified. The Thirlwall Inquiry publication date and "profound failure of management, governance and safeguarding" finding, plus all quotations from the three volumes, which were read directly; the Ockenden Nottingham report date; NHS England performance figures; NHS Vacancy Statistics; the NMC register data; the manager barring scheme design, scope and timetable; the NHS College of Leadership and Management launch details; the King's Fund international comparison. Each is linked or cited at the point of use.

Resolved since the first draft.

• Durham. The Aubrey report itself is now the source for the governance material: The Report of the Independent External Review of Governance within the Breast Surgery Services, Mary Aubrey, November 2025, 232 pages, commissioned by the Trust in March 2025. Its own figures are 13 key issues, 37 recommendations (21 high, 15 medium, 1 low priority), and 39 patients identified for full duty of candour disclosure at the time of writing. One attribution point: the harm figures of 514 cases reviewed, 315 harmed, 77 significantly harmed and 20 unnecessary mastectomies are genuine and current — they are the Trust's own look-back figures as reported on 18 September 2026 — but they are not in the Aubrey report, and should be cited to the Trust's review rather than to Aubrey. The one number to drop is the derived "77%" harm rate: 77 of 315 is 24.4%, and the correct headline rate is 315 of 514, around 61%. Two further cautions: the report carries internal date inconsistencies (the index-case MDT appears as both 12 and 14 February 2025), and it states that it "should not be quoted or referred to in whole or in part without prior written consent" — a permissions point to settle before any external use.
• Lord Mann. Published 4 June 2026 (the gov.uk page was last updated 2 July 2026), commissioned October 2025, with 24 recommendations, not 36. The claim that all 36 were "accepted in full" is wrong on both counts: the government response commits to recommendations for the Department and NHS England, supports those aimed at independent regulators only in principle, and makes some conditional on consultation. The briefing's claims about a ban on political symbols on uniforms and mandatory IHRA adoption across all trusts could not be verified, so they are not used.
• Thirlwall. 17 recommendations, confirmed against Chapter 45. The supplied documents gave both 14 and 17; 17 is correct. All three volumes have since been read in full for the families' evidence and the post-2016 conduct in the annex.
• Northern Ireland. The overspend is not 12–14% of the health budget. The Fiscal Council forecasts a health overspend of about £212m in 2025–26 against a budget near £8.9bn; the Department's pre-mitigation gap has been reported at around £800m, closer to 9%. The more telling figure is the baseline falling from £8,898m to £8,481m for 2026–27 while pay costs rise. The waiting list and emergency department figures are from the Department's own quarterly statistics and Pivotal's tracker, and are solid.

Still unverified, and deliberately unused. Named individuals at County Durham and Darlington: dates of office, subsequent appointments, pension and employment status. The Aubrey report refers to post-holders by role rather than name, and this document follows it. CQC oversight-framework rankings and league positions ("45th to 55th of 134", "114th of 134"); the Freedom to Speak Up Guardian numbers (1,200–1,500 guardians, 35,000+ cases a year) and the international whistleblowing-awareness comparisons; the OECD spend-and-beds table and four-nation spend-per-head figures. The King's Fund comparison is used in their place where it covers the same ground.

On the aviation comparison. Use it for the class of system — uncontained failure, system-level scrutiny, mandatory occurrence reporting, findings implemented as a condition of operating. Comparative safety statistics between aviation and healthcare are contested and are not verified here.

On Chester and the convictions. The Inquiry proceeded on the basis of the convictions and was expressly not asked to re-examine the medical and statistical evidence; a Criminal Cases Review Commission application supported by an international expert panel remains live, and the Chair refused an application to pause the Inquiry pending it in her ruling of 19 March 2025. The findings about management, governance and safeguarding rest on the executives' own emails, minutes and evidence, and do not depend on the outcome of that application.

Two corrections on the values lineage

The Mayos were not Quakers. William Worrall Mayo was an Episcopalian, and all three Mayo doctors were Freemasons. The Quaker connection is real but indirect and intellectual: Mayo studied science in Manchester under John Dalton, who was a Quaker, and Mayo Clinic's own UK material credits Dalton with inspiring him with "the Quaker values of tolerance and working together for the common good". Stated as a family affiliation it is simply wrong — and the Fry–Mayo–Aga Khan parallel is stronger without it, since three different faiths arriving at the same structures is a more interesting claim than a shared one.

Elizabeth Fry did not demonstrably inspire Florence Nightingale. The scholarly source normally cited for this — Huntsman, Bruin and Holttum in Medical History (2002) — states the opposite: "written proof of direct contact, although likely, is lacking", and posits only an indirect link. Fry-trained nurses did serve in the Crimea, but in a later party of six (two of them Fry nurses, departing 25 December 1854), not accompanying Nightingale. The paper's own explanation of Fry's obscurity is that "Miss Nightingale's quite extraordinary charisma inadvertently dimmed those who preceded her." The Fry passages here avoid the unsupported claim. Also unverified: the widely circulated Fry line "Punishment is not for revenge, but to lessen crime and reform the criminal" has no traceable primary source — don't quote it.

Two smaller flags. Mayo's requirement that leaders be physicians is confirmed in its bylaws for the President and the Board of Governors, but no rule requiring it of clinical department heads could be found — that appears to be practice, not constitution. And the Mayo physician and scientist headcounts in circulation date from 2022, so only the workforce and revenue figures, which are current, are cited.

Three corrections in the civility material — these matter if you present any of it

The "61% reduction in cognitive capacity" cannot be traced to any source, and the paper it is attributed to reports a 15.6% effect. This has been checked directly against the primary paper. Porath and Erez, "Does Rudeness Really Matter? The Effects of Rudeness on Task Performance and Helpfulness", is in the Academy of Management Journal 2007;50(5):1181–1197 — not the Journal of Applied Psychology, as the attribution often given for the 61% figure states.

The paper does contain a working-memory measure, in Study 3: a paired-associate word task, which the authors describe as "commonly used by cognitive psychologists to test disruption of memory, attention, and other cognitive processes (i.e., working memory capacity)". Across 98 undergraduates, recall was 13.06 in the control condition against 11.02 after rudeness (F = 13.79, p < .01) — a 15.6% reduction. That is the closest thing in the literature to the claim being made, and it is nowhere near 61%. The number 61 appears nowhere in the paper.

The likeliest origin is an inverted denominator. In Study 2, controls solved 5.18 anagrams against 3.21 after rudeness. Controls therefore solved 61.4% more than the rudeness group — which, expressed properly as a reduction, is 38%. Somewhere that 61 was relabelled as a fall in cognitive capacity.

What the paper actually establishes is stronger than the statistic, and should be used instead. Study 3 found that impaired memory-recall significantly mediated the effect of rudeness on anagram performance, creative output and cognitive flexibility. That is an experimentally demonstrated causal chain — rudeness degrades cognitive processing, and the degraded processing is what degrades the work. A mediation finding is better evidence than a headline percentage, and it cannot be dismissed as a misquoted number. Alongside it: rudeness recipients solved 25–38% fewer anagrams, generated 28–37% fewer creative uses for an object, and were 74–84% less likely to help a colleague.

Porath's own published percentages settle it, and they are the ones to use. Georgetown's account of her work ("Civil Actions") states the experimental results as percentages directly: participants belittled before a word puzzle "performed 33 percent worse", came up with "39 percent fewer creative ideas", and those who merely witnessed rudeness fared "22 percent and 28 percent worse on the two tasks". On the positive side, people were "59 percent more willing to share information" and "72 percent more willing to seek advice" with someone civil. Her prevalence figures: 98% have experienced incivility at work, and 62% experience it at least monthly, up from 49% when she began in 1998. Half of those who admit to rudeness say it is because they feel overloaded; only 4% say they do it because it is fun and they can get away with it.

Note what that does to the familiar "61% for the recipient, 20% for onlookers" pairing. The onlooker figure is roughly right — Porath reports 22% and 28%. The recipient figure is roughly double what she reports, which is 33% and 39%. So the honest version of the claim is: rudeness costs its target about a third of their performance and bystanders about a quarter of theirs — which is a devastating finding, needs no inflation, and can be sourced to the researcher's own account.

One quotation to drop: "immune to the sound of pain." The supplied Thirlwall briefing file presents this as Chris Turner's testimony to the Mid Staffordshire inquiry and as "one of the most famous and devastating lines of the entire inquiry". An exact-phrase search returns no trace of it — not in reporting of the inquiry, not on the Civility Saves Lives site, not in any interview with him. A line that famous would leave a trail. It appears to have been invented by the assistant that produced that file.

The underlying recollection is sound, though, and the verified material is better. Turner was a junior doctor in the emergency department at Mid Staffordshire from 2007 — the BMA's own word, and the grade now called a resident doctor; the document no longer says "trainee", which understates a doctor several years into a career — and he did give evidence to the public inquiry, describing a culture of bullying and harassment, staff who were utterly demoralised, nurses threatened "on a near daily basis" with losing their jobs, and "no significant medical leadership". The personal cost is verified from the same source: "First I stopped sleeping and eventually eating"; two and a half stone lost in nine weeks; "after two years I broke under the pressure and resigned". He is a co-founder of Civility Saves Lives, with Joe Farmer — not its sole founder. His verified words on desensitisation are "it's like an emotional muscle — once we start to display rudeness, incivility, it gets easier and easier to accept", and on the cause, "in mid staffs the exec team was measured on finance, and pretty much nothing else". Both are used here. Use those instead; they make the same point and will survive a fact-check.

Two quotations in the continuum material, handled differently. The "ignorance leads to fear, fear leads to hatred, and hatred leads to violence" formulation is attributed to Ibn Rushd (Averroes) without any traceable source. Every instance found is a quotation aggregator or a social media post; none cites a work, a passage or an Arabic original, and "this is the equation" reads as modern phrasing. Treat it as a formulation commonly attributed to him rather than a citation — the chain of reasoning stands on its own merits, and presenting it as a verified twelfth-century quotation to a sceptical audience is an avoidable risk.

The Rwandan inscription is real but the location varies between accounts. A law professor's published account of visiting the memorials places it at the Nyamata Genocide Memorial, the former Nyamata Church, "on plaques throughout the church", worded "If you had known me, and you had really known yourself, you would not have killed me", and describes it as the words of survivors with no individual attribution. Other accounts place a close variant at the Kigali Genocide Memorial, in the Children's Room. Safest usage: quote it as an inscription at Rwanda's genocide memorials, in the words of survivors, without pinning it to one site.

One attribution to be careful with on antisemitism. The sentence "the levels of anti-Jewish racism in the UK constitute a national emergency" is sometimes attributed to Lord Mann. That could not be verified. What is verifiable: his NHS review found that "racism is deeply pervasive and has reached the stage where it has become normalised in the NHS"; his wider commission with Dame Penny Mordaunt for the Board of Deputies found that Jewish people are "tolerated rather than respected" in the UK; and by September 2026 the rise in antisemitism was being characterised as a national emergency in Parliament, in those terms by the Leader of the Opposition. Use "tolerated rather than respected" as the Mann quotation — it is verifiably his commission's, and it is the more useful line, because it describes a posture rather than a crisis level.

The "20% drop in team performance" from Riskin et al. is really 15–17%. The 2015 Pediatrics randomised simulation trial across four Israeli NICUs found diagnostic performance 2.65 versus 3.18 and procedural performance 2.77 versus 3.26 on five-point scales — falls of 16.7% and 15.0%. The paper's own honest headline is that rudeness explained nearly 12% of the variance in performance. Also, the claim of a "50% reduction in information sharing and help-seeking" is not a finding of that paper: those were standardised regression coefficients, not proportional reductions. The cleanest citation for a large clinical effect is Katz et al. (BMJ Quality & Safety, 2019): 91.2% of the control group achieved passing scores in a simulated operative crisis, against 63.6% of the group exposed to rudeness.

The nursing workforce figures, stated properly

"30–50% of new nurses leave within two years" is not a UK figure, and is not traceable to a primary US source either. The actual US primaries are 22.7% of newly hired registered nurses leaving within a year (NSI National Health Care Retention & RN Staffing Report, 2026 edition, 527 US hospitals, 262,405 RNs) and 17.5% leaving their first job within a year (Kovner et al., Policy, Politics, & Nursing Practice, 2014) — both measuring change of employer, not departure from nursing. The "$60,090 replacement cost" and "$295,000 per percentage point" are likewise US figures, usable as illustration only.

No UK cohort attrition rate is published at all. The nearest is the Nuffield Trust's finding (Waste not, want not, September 2023, on NMC data) that 98% of UK-trained nurses and midwives were still on the register three years after joining, alongside roughly 18% having left NHS hospital and community settings within two years. Those are different measures and the difference is the whole point.

The RCN's 67% needs its base. RCN analysis of NMC data (11 November 2024) covers UK-educated nursing staff in England only, and gives estimated counts: within five years of registering, 697 in 2021 rising to 1,166 in 2024; within ten years, 1,607 rising to 2,295. The RCN describes both as estimates, gives no denominator and states no rate. Quote the counts.

Student attrition is a third question again. Nuffield Trust, on HESA data: around one in eight nursing students in England and Wales did not gain their intended degree between 2014 and 2020, about one in six in Scotland. The RCN's 21% is its own calculation on three specific intakes. The NAO's 33% is a completion-on-time measure. Three numbers, three different questions.

A note on the bright spots generally

Several of the best-known claims about them are softer than their reputation, and this document says so in each case: Buurtzorg's "40% system saving" does not survive KPMG's whole-system accounting (49th percentile on case-mix adjusted total cost); the Esther outcome figures have no control group and the county's own evaluators say it was not designed as research; Southcentral Foundation's published percentages are mutually inconsistent across its own documents; the Schwartz Rounds evaluation failed on its primary outcome; Martha's Rule has no completed outcome evaluation; and Learning from Excellence has no published patient-safety effect. The mechanisms are worth copying. The numbers should be quoted with their caveats attached.

A structural caveat about the sources themselves

Most of the documents originally supplied are transcripts of AI search conversations rather than primary research. The three Thirlwall volumes, the Aubrey report and the Lord Mann review have since replaced the largest of those blocks with primary sources, which is why the flags above have shrunk. They are useful as a map of the territory, and much of their content checks out. But they contain internal contradictions — for example, on which executive held which post at Chester, and on the total cost of the Gilby employment case — and the assistant's responses are consistently agreeable, restating the questioner's framing back rather than testing it. Several claims are also dated after the events they describe and presented as settled. Before any of this material goes into a client-facing document, the specific figures should be checked against the inquiry reports and official statistics directly.

Sources consulted

• Thirlwall Inquiry, Report of the Public Inquiry into events at the Countess of Chester Hospital, 2015 to 2018, HC 597-I/II/III, 15 September 2026 — all three volumes read directly (in the project folder), plus GOV.UK, the report, recommendations, Chapter 45 and the publication statement
• Aubrey M. The Report of the Independent External Review of Governance within the Breast Surgery Services, County Durham and Darlington NHS Foundation Trust, November 2025 (in the project folder)
• Lord Mann review of antisemitism and other forms of racism in the NHS and healthcare regulatory system and the government response
• Ockenden Maternity Review — final report, Nottingham
• Infected Blood Inquiry — report publication transcript, 20 May 2024
• NHS key statistics: England — House of Commons Library, BMA backlog analysis, Nuffield Trust performance dashboard
• NHS Vacancy Statistics, England, April 2015 – June 2026, NHS England / NHS Digital, published 27 August 2026
• The NMC Register, 1 April 2025 – 31 March 2026; NMC, Professionals who left the NMC register in 2024/2025, June 2025
• Huge increase in nurses quitting early — Royal College of Nursing, 11 November 2024
• Palmer W, Rolewicz L, Dodsworth E. Waste not, want not: strategies to improve the supply of clinical staff to the NHS, Nuffield Trust, September 2023
• NSI Nursing Solutions, National Health Care Retention & RN Staffing Report, 2026 edition; Kovner CT, Brewer CS, Fatehi F, Jun J. "What Does Nurse Turnover Rate Mean and What Is the Rate?", Policy, Politics, & Nursing Practice 2014;15(3–4):64–71
• Northern Ireland quarterly waiting times, position at 30 June 2026 — Department of Health NI; Pivotal, August 2026; NI Fiscal Council Budget assessment
• Leading the NHS: proposals to regulate NHS managers — consultation response; NHS College of Leadership and Management launch
• How does the NHS compare to the health care systems of other countries? — The King's Fund
• Murray H, Gillies C, Aalamian A. Physician incivility in the health care workplace. CMAJ 2024;196:E295
• Porath CL, Erez A. Does rudeness really matter? Academy of Management Journal 2007;50(5):1181–1197; Civil Actions — Georgetown
• Riskin A, Erez A, Foulk TA, et al. The impact of rudeness on medical team performance: a randomized trial. Pediatrics 2015;136(3):487–495; Katz D, Blasius K, Isaak R, et al. Exposure to incivility hinders clinical performance in a simulated operative crisis. BMJ Qual Saf 2019
• Foulk T, Woolum A, Erez A. Catching rudeness is like catching a cold. Journal of Applied Psychology 2015; Andersson LM, Pearson CM. Tit for tat? The spiraling effect of incivility in the workplace. Academy of Management Review 1999;24(3):452–471
• Czech H, Hildebrandt S, Reis SP, et al. The Lancet Commission on medicine, Nazism, and the Holocaust. The Lancet, 8 November 2023 (in the project folder); Proctor RN. Racial Hygiene: Medicine Under the Nazis, Harvard University Press, 1988
• Staub E. The Roots of Evil, Cambridge University Press, 1989; The Psychology of Good and Evil, Cambridge University Press, 2003, ch. 23; Ervin Staub — ABLE Project, Georgetown Law; Waller JE. Becoming Evil, Oxford University Press, 2002, 2nd edn 2007
• Respect where it's due — British Medical Association (Chris Turner on Mid Staffordshire)
• Mayo Clinic mission and values, bylaws, salary-only compensation study, Little Book of Mayo Clinic Values
• AKUH Nairobi fifth JCI accreditation, Patient Welfare Programme, AKDN ethical framework
• MSF charter; Tackling institutional discrimination and racism within MSF
• Huntsman RG, Bruin M, Holttum D. Twixt candle and lamp: the contribution of Elizabeth Fry and the Institution of Nursing Sisters to nursing reform. Medical History 2002;46(3):351–80
• Southcentral Foundation Baldrige profiles — NIST; Buurtzorg — Commonwealth Fund; Esther — Region Jönköping and the Commonwealth Fund case study
• Martha's Rule — NHS England and first-phase implementation; Martha Mills prevention of future deaths report
• Learning from Excellence; Kelly N, Blake S, Plunkett A. Arch Dis Child 2016;101(9):788–791; Schwartz Rounds origins; Maben J et al. Health Serv Deliv Res 2018;6(37)
• Patient safety aim met 15 months early — Scottish Government; NHS Staff Survey 2025 — UCLH
• Reporting of the Lady Joan Branson inquest, Westminster Coroner's Court, September 2026 (Professor Fiona Wilcox, Senior Coroner for Inner West London)
• CQC inspection history, Countess of Chester Hospital NHS Foundation Trust (provider RJR), inspections of November–December 2018 and February–March 2022, plus the focused maternity inspection of July 2022

Plus the supplied documents: HealthIncivilityOverview, HealthLeadershipOverview, NHS_Leadership_Overview, Thirlwall, Thirlwall_ExtraInfo, and from the project folder — NUH_ChatKeyIssues, NUH_Research-Summary, NUH_LordMannReview, the NUH Trust Board pack of 10 September 2026, the Union Health Service 90-day review pack, the nursing drop-out material, and the incivility research set (Schilpzand, De Pater and Erez 2016; Leiter et al. 2012; the CIPD bullying and incivility evidence review 2022; Milesky et al. 2015; Physician Incivility 2024).

The financial figures in the annex — what is safe and what is not

A note on currency in this volume. Figures here are quoted in sterling exactly as the source publishes them, so that each can be checked against the original document. The dollar equivalents at $1.35 = £1 appear in the volumes themselves, not in these verification notes.

Three corrections to numbers in wide circulation.

1. £60bn is a provision, not a bill. The clinical negligence provision (£60,259.1m at 31 March 2026) is the actuarial present value of all incurred claims, reported and unreported. Annual cash paid was £3,238.5m in 2025/26. Dividing the provision by an annual budget is a category error.
2. 13,329 is claims resolved in 2024/25, not claims received. Claims received were 13,784 (2023/24), 14,428 (2024/25) and 15,236 (2025/26).
3. No average cost per clinical negligence claim is published. "About £230,000" is £3,088.9m of cash paid divided by 13,329 claims resolved — two different flows, since much of the cash relates to claims settled in earlier years. The same method on 2025/26 data gives £273,000. Quote instead: 48% of claims settled with no damages; claims of £1m+ were 2% of volume and 68% of cost; and the one genuine published per-claim average, £11.2m for obstetric cerebral palsy and brain damage claims settled in 2024-25.

Safe to quote: the provision series (£58.5bn / £60.3bn / £60.3bn at 31 March 2024 / 2025 / 2026); the 2025/26 payment split (damages £2,427.9m, claimant legal £622.2m, defence £188.4m); the NAO's 3.7:1 legal-cost ratio on claims of £25,000 or less; the £1,554m obstetric brain-injury cost from 237 claims; the GAD projection of £4.1bn by 2029-30; Thirlwall's £18.58m to 31 March 2026; the Covid-19 Inquiry's £219.8m at 30 June 2026; the median statutory inquiry at £11.5m and £6,600 a day; the Institute for Government's £638.9m across 68 inquiries since 1990; RAND/IPPO's £12.1bn; the NHS sickness absence and Staff Survey figures; the agency spend series; Post Office redress of £1,697m with £182.6m of scheme legal costs; and the coal health figures of £4.1bn compensation against £2.3bn administration including £1.3bn in solicitors' fees.

Not safe to quote: any total lifetime cost of Horizon redress (not published — the £1.7bn paid plus £0.9bn provision add-up mixes two as-at dates); any cumulative total for the Post Office Horizon Inquiry beyond about £48m to 2024; the £57.75m / £10m settlement split (confidential, contemporaneous reporting only); "Post Office spent about £100m defending the group litigation" (untraceable); £9.7m for Operation Hummingbird (a primary document exists but could not be verified directly); any cost at all for Operation Duet, the Ockenden reviews, the Amos investigation, the Kirkup reviews, trust legal representation at Thirlwall, the Aubrey review or the Durham look-back (none published); "£2.4bn a year" for NHS sickness absence (untraceable); any absolute NHS bank staff spend total (NHS England publishes only percentage changes); any national count of NHS employment tribunal claims or NHS spend on settlements and NDAs (unpublished); US per-nurse turnover costs presented as UK figures; press estimates of scandal-specific compensation totals (£101m for Nottingham maternity, "more than £60m" for Chester) — these are FOI-derived journalism, not official series, and the annex does not use them.

On the Durham scaling. The 61.3% harm rate and 15.0% serious-harm rate are the Trust's own look-back figures on 514 cases reviewed to 18 September 2026. Scaling them to 10,000 records is illustrative arithmetic and is labelled as such in the annex, with three caveats attached: the earliest cases reviewed were the highest-risk ones, so the rate is likely to fall; harm is not the same as a successful negligence claim; and the per-claim cost ratio is not a published average. Note also that the Durham harm was avoidable harm from substandard practice and governance, not deliberate harm — the distinction matters, and Chester is the only case in this document involving deliberate harm.

Financial sources. NHS Resolution, Annual Report and Accounts 2025/26 (HC 421) and 2024/25 (HC 1065) · NAO, Costs of clinical negligence (17 October 2025) · Public Accounts Committee, HC 1234 (30 January 2026) · Thirlwall Inquiry financial report FY25/26 (8 September 2026) · House of Commons Library CBP-10981, Public inquiries: costs (8 September 2026) · Lampard Inquiry financial report (19 June 2025) · Institute for Government, Public Inquiries (December 2017) · Government response to the House of Lords Statutory Inquiries Committee (February 2025) · NAO on government compensation schemes (17 April 2026) and IBCA community updates · DBT, Post Office Horizon financial redress and legal costs data (28 August and 30 June 2026) and DBT Annual Report and Accounts 2025–26 · Public Accounts Committee, Coal Health Compensation Schemes (4 March 2008) and NAO (18 July 2007) · RAND Europe / IPPO, NHS staff wellbeing (June 2022) · NHS England Digital, NHS Sickness Absence Rates, May 2026 (27 August 2026) · NHS Staff Survey 2025 (12 March 2026) · NHS England Month 12 financial position 2025/26 and Annual Report and Accounts 2025–26 · Health Foundation on retention (8 January 2026) · Commons Library CBP-10852 on Windrush (16 June 2026).

Note on one supplied link: the Guardian report of 17 October 2025 on liabilities reaching £60bn could not be fetched (the domain is blocked to this session's tools, and no workaround was attempted). It is journalism about the NAO report published the same day, and the annex cites the NAO, NHS Resolution and the PAC directly instead.

Prevalence figures — two corrections and the verified set

Correction 1: "88% of US healthcare workers experienced intimidating behaviour in a year" is not a finding of the 2013 ISMP survey. That survey (4,884 respondents) reported 77% encountering reluctance or refusal to answer questions, 73% negative comments about colleagues, 69% impatience with questions, 68% condescending language, 46% shaming or humiliation and 7% physical abuse, with 11% aware of a medication error in which intimidation played a role. The 88% appears to be a retrospective restatement of ISMP's 2003 survey and could not be confirmed in contemporaneous reporting; at 23 years old it should not be quoted at all. Use the 2021 survey instead (1,047 respondents, Sept–Nov 2021): 79% personally experienced disrespectful behaviour, 60% witnessed it, 64% had condescending or demeaning comments directed at them, 75% said their organisation does not deal with it effectively, and 27% were aware of a medication error in the previous year in which disrespectful behaviour played a part — up from 11% in 2013. The same applies to the related claim that "nearly half changed how they handled a patient safety issue because of intimidation": that is a 2003 ISMP finding about order clarifications, not a 2013 one, and it is not used here.

Correction 2: the NHS Staff Survey figures in circulation are two years out of date, and the direction has changed. The 2025 survey (published 12 March 2026; 766,285 respondents, 49% response, 238 organisations including all 206 trusts) records 17.01% experiencing bullying, harassment or abuse from colleagues and 9.11% from managers — both the lowest recorded, against 23.2% and 8.7% on the 2023 measures — with discrimination at 8.77%, also the lowest recorded, and 25.25% abused by patients, service users, relatives or the public. Say so plainly: the culture measures are improving from a bad base, while 29.55% often think about leaving (a three-year high) and 16.75% would go as soon as they found another job (an all-time high). Do not carry the older figures forward, and do not claim the numbers never move.

The verified international set.

• Australia, hospital staff: 93.6% experienced at least one unprofessional behaviour; 38.8% experienced incivility or bullying at least weekly; 14.5% experienced extreme unprofessional behaviour; 49.8% said unprofessional behaviour had a moderate or major negative effect on patient care, error rates or quality of service. Westbrook et al., Medical Journal of Australia, 2021 — seven metropolitan tertiary hospitals, 5,178 respondents, 34.0% response rate.
• Australia, doctors in training: 20% experienced and 26% witnessed bullying, harassment, discrimination or racism; 56% among Aboriginal and Torres Strait Islander trainees. Medical Training Survey 2025 (Medical Board of Australia / Ahpra), 17,622 respondents.
• Australasian surgeons: 38.7% had been bullied; 49% had experienced discrimination, bullying, harassment or sexual harassment; 63% among trainees. RACS Expert Advisory Group, 2015, 3,516 respondents — flag as 2015 and Australasian if used. This, and the MJA weekly-incivility figure, are the two real sources of the "38%" in circulation.
• New Zealand: 37.2% of senior medical officers self-reported being bullied; 67.5% had witnessed a colleague being bullied. ASMS, 2017, 1,759 respondents, using the validated NAQ-R. Flag as 2017.
• Canada: no good healthcare-specific national survey found. The nearest is Statistics Canada's General Social Survey 2016, in which 23% of workers in health occupations reported workplace harassment against 19% of all workers — a subgroup of a general sample, and dated.
• International, pooled: bullying prevalence among medical residents 51% (95% CI 36–66), from 13 cross-sectional studies and 44,566 participants across the US, Australia, Australia/NZ, India, Pakistan and Latin America. Journal of Occupational Health 65(1), 2023. For nurses, Lever et al. (Journal of Advanced Nursing, 2019) pooled 26.3%, range 3.9–86.5%.

Always say which of the three things a figure measures — experienced, witnessed, or experienced at a stated frequency. Conflating them is the fastest way to lose an audience that knows the literature.

Founding, values and the College of Leadership and Management

A further supplied document (NHSFoundingEthos) is again an AI chat transcript, so it was treated as a lead list rather than a source. What survived checking:

Verified and used. The NHS was founded on 5 July 1948 under the National Health Service Act 1946, launched by Aneurin Bevan at Park Hospital, Davyhulme (now Trafford General), funded principally from general taxation rather than insurance contributions. The NHS Constitution for England sets out seven principles and six core values — working together for patients; respect and dignity; commitment to quality of care; compassion; improving lives; everyone counts — and the descriptions quoted in Volume 1 and Volume 10 are verbatim from the Constitution.

Corrected. The transcript gives 6 July 2026 as the announcement date for the NHS College of Leadership and Management. The primary record dates the launch correspondence and NHS Employers' notice to 9 July 2026, with the framework overview document dated 7 July 2026; use 9 July 2026, which is what the volumes say. The transcript's claim that the Framework "responds directly to" the Messenger and Kark reviews is plausible in substance but is not stated in NHS England's own launch long read — do not attribute it to NHS England. The rollout dates now in Volume 9 (embedding during 2026/27, manager self-assessment by end of Q4 2026/27, board competency alignment from 2027/28) are from that long read.

Not quoted. The 1948 household-leaflet slogans and Bevan aphorisms in the transcript are widely reproduced but were not checked against the original leaflet or Hansard, so only the best-attested line ("It is not a charity. You are all paying for it.") is used, and it is used as a leaflet line rather than a Bevan quotation. The claim that National Insurance covered "roughly 20%" of the NHS budget at launch was not verified and is not used.

Bevan's words, and Mayo's English origins — what is verified. Volume 10's founding-settlement argument rests on two quotations and one biographical fact, and all three were checked. The collective-principle sentence — "no society can legitimately call itself civilised if a sick person is denied medical aid because of lack of means" — is from Bevan's In Place of Fear, 1952, chapter 5, "A Free Health Service". Three precisions on it: it is a separate sentence beginning "It insists that…", not a continuation of "The collective principle asserts that…", so it is quoted here on its own rather than spliced; no page number is verifiable from the available transcriptions, so cite the chapter and not a page (earlier drafts said p.100); and those transcriptions spell it "civilized" — this document's "civilised" is a spelling normalisation, not a different source. The second quotation, that illness is "neither an indulgence for which people have to pay; nor an offence for which they should be penalised", was previously given here as coming from the same book. It does not. A provenance check traces it to a paraphrase, without quotation marks, in T. H. Marshall's Social Policy in the Twentieth Century (1967), later turned into a Bevan quotation by others; no Bevan source has been found. It has been removed from Volume 10 and should not be attributed to him. The longer version continuing "… but a misfortune the cost of which should be shared by the community" is in wide circulation but was not verified word for word, so neither form is used. "Preventable pain is a blot on any society" has since been verified and the earlier instruction not to use it is withdrawn — it is Bevan's own written words, in the same chapter. "The NHS will last as long as there are folk left with the faith to fight for it" remains untraceable to any primary source — no speech, no date, no Hansard record, and it is not in that chapter; the wording also drifts between versions, which is the usual marker. Do not use it. And the ambition is "universalise the best", Bevan's own words in the Commons on 30 April 1946 (vol 422, col. 49), not the later paraphrase "universalising the best". The full founding record, including what the NHS Constitution does and does not bind, is now in the Annex on the founding vision of 1948. On Mayo: William Worrall Mayo was born in Salford, near Manchester, on 31 May 1819, was a protégé of John Dalton, and emigrated in 1846 — all from Mayo Clinic's own page, titled "English taproot of values", which is also the source for "the Quaker values of tolerance and working together for the common good". The caveat still stands and is still stated: the Mayos themselves were not Quakers. The claim in Volume 10 is therefore that the tradition is British in origin, not that Mayo is a British institution.

Dentistry, hospices, community care and the 2026 policy position

Dentistry. Spend, access and dental-desert figures are from the House of Commons Library (CBP-9597, June/July 2026), NHSBSA (August 2026), the Public Accounts Committee (Fixing NHS Dentistry, 4 April 2025) and a peer-reviewed spatial analysis in the British Dental Journal (14 August 2026). Three cautions. The private market figures (£8.4bn, 69%) are 2023/24 data published in September 2025 and re-cited by the CMA in March 2026 — do not present them as 2026 figures. The widely quoted "97% of new patients cannot access NHS care" is really 96.9% of people who had no dentist and who tried to get NHS care, from ONS 2024 survey waves; use it with that framing or not at all. And the UDA system has not been abolished: the April and June 2026 changes modify its values and add non-UDA payments. There is no published net or gross NHS dental spend for 2025/26, so 2024/25 is the latest usable year. "Golden hellos" date from the 2024 recovery plan and were not re-announced in 2026.

Hospices. Use the National Audit Office (The financial sustainability of England's adult hospice sector, October 2025) for structure and finances, the Public Accounts Committee (18 March 2026) for the 0%–80% ICB variation, Hospice UK for UK reach and the 2026 service-cut survey, Together for Short Lives for children's hospices, and Marie Curie (February 2026) for unmet need. Three figures for the statutory share circulate and they are not compatible: use NAO's 29% (England, adult hospices, ICB share of income, 2023-24), or Hospice UK's "about a third for adult hospices, about a fifth for children's". Do not use "40%" — it appears in a Hospice UK release and could not be reconciled with either of the others. Use NAO's £1.55bn sector spend, not the £1.2bn figure that appears elsewhere. The count of children's hospices in the UK is not published; only "around 40 for children and young people" in England (December 2024) is verifiable, so "54 in the UK" is not used. On palliative need, two defensible figures measure different things — Hospice UK's "90% of deaths could benefit from palliative care" and Marie Curie's population estimate that ~46% need it; the volumes use the unmet-need figure (203,442, 31%) and say whose it is. Marie Curie's 2026 sample was 95% White and the authors state it cannot support conclusions about ethnicity, so it is not cited for that.

Community care and carers. Activity and finance figures are DHSC official statistics for 2024-25; workforce and home-care location counts are Skills for Care (2025 edition, 2024/25 data). The home-care caseload figures from the Capacity Tracker are response-weighted — 505,886 people is from the 84.9% of providers who responded, so the real number is higher, and the response rate is always stated. There is no published 2024-25 headcount of direct payment recipients: the volumes use the ASCOF proportion (24.5%), and the two figures must not be multiplied together because the denominators differ. The direct-payment rule on relatives is quoted from the Care and Support (Direct Payments) Regulations 2014, regulation 3, not from summaries of it. Carer's Allowance rates are DWP's 2026/27 table. Note the distinction between 1.4 million claiming and around 902,000 paid — 26.7% are entitled but not paid because of overlapping benefits, and the two are constantly conflated. The value of unpaid care is £184.3bn for 2021/22 (Centre for Care / Carers UK); the £132bn figure still in circulation is the 2015 estimate. Carers UK's State of Caring is a self-selecting survey, so its percentages are directional, not population prevalence. No authoritative body states that supporting unpaid carers is cheaper than hospital or residential care, so that claim is not made; the costed findings (delayed discharges £1.89bn in 2022/23; the IFS's 3.8% A&E association) are used instead.

The August 2026 mental health expansion. The £343m, the 100 community mental health centres and 59 mental health emergency departments, the £187m/£156m split, the 6-and-23 baseline and the autumn 2026 / March 2027 timetable are all from the DHSC press release of 5 August 2026. Two cautions: the release gives no figure for the number of people the programme will reach — its "around 1 in 5 people" is a prevalence statistic about common mental health conditions, not a count of patients — and nothing published establishes how much of the £343m is new money. It is capital within the 2025 Spending Review settlement, delivering a 10 Year Health Plan commitment, and it exceeds that plan's "up to £120 million" for mental health emergency departments. Say that, and no more.

The change of government, and Andy Burnham. As at the DHSC release of 5 August 2026, Andy Burnham is Prime Minister and Yvette Cooper is Secretary of State for Health and Social Care; James Murray held the health post on 22 June 2026, so the change falls between late June and late July 2026. The exact date was not established, so no volume states it. Burnham's 2010 record is quoted from Hansard (30 March 2010) and the White Paper itself. The "death tax" row is verified as having happened (Health Foundation, July 2026) but its 2010 specifics — the Conservative poster, the collapse of cross-party talks, individual roles — were not verifiable here (the contemporaneous coverage sits on domains blocked to this session), so they are not described. No direct quotation from his 2010 or 2015 leadership campaigns is used; "whole person care" appears as paraphrase only. The 29 July 2026 announcement, the four principles and the Casey acceleration to summer 2027 are from the Downing Street release and the HFMA's report of it. On Greater Manchester devolution, the peer-reviewed evaluation (Britteon et al., Social Science & Medicine, 2024) found life expectancy rose 0.233 years more than in the synthetic control and healthy life expectancy 0.603 years — but the standard error on the second is large relative to the estimate, so it is not quoted as a precise gain.

Social care: what was checked in the supplied research, and what was corrected

A further supplied document (UKSocialAndOtherCare) is an AI chat transcript and was treated as a lead list. Every figure used from it was traced to a primary source; a substantial number did not survive.

Verified and used. The Dash review's wording, published 15 October 2024 ("significant failings in the internal workings of CQC, which have led to a substantial loss of credibility"). CQC's own assessment volumes, from its 2024/25 annual accounts: 10,356 (2022/23) → 6,230 (2023/24) → 3,825 (2024/25). The Homecare Association's 83.5% (June 2026, CQC data to 5 May 2026), split 36.9% never assessed and 46.6% rated four to ten years old. Ofsted's Annual Report 2024-25 on unregistered children's homes (nearly nine in ten councils, almost 900 investigations) and Ofsted's July 2026 figures (900+ homes identified in 2024, ~1 in 10 children in residential care, £353m, some charging up to £30,000 a week); the Children's Commissioner's 669 children at an average £10,500 a week (January 2026); the CMA's 22.6% average operating margin at the fifteen largest providers (2022). ADASS: £623m mid-year projection (Autumn Survey 2025) and £715m outturn, 3.2% (Spring Survey 2026) — the same year, not competing figures — with 80% of councils overspending, 37.9% of net council budgets, £909m of planned savings and the 547 → 712 rise in young adults needing packages over £7,000 a week. CQC State of Care (October 2025): nearly six in ten patients ready for discharge delayed, on any given day in March 2025. The registration gap in England, with SSSC, Social Care Wales and NISCC figures, and Social Work England's 108,345 registrants at 31 August 2026. Shared Lives (9,802 people, 96% Good or Outstanding, LGA's £12,000–£26,000 range). Somerset's 1,250 micro-providers since 2015. The Health and Social Care (Wales) Act 2025 and its 2026–27 timetable.

Corrected. The transcript's "inspections fell 68%, from over 10,000 to 3,856" is wrong in both numbers and percentage — the true figures are 10,356 → 3,825, a 63% fall, and they are CQC's, not Dash's. "£20,000–£40,000 a week" for an unregistered placement is not a typical price: the published average is £10,500, with "up to £30,000" in some cases. "Sir Martyn Oliver called it a national scandal" could not be traced to any Ofsted publication or speech and is not used. The public-understanding figures (31%/33% of adults, 52%/44% of 16–24s) are Ipsos for the Health Foundation, May 2025 fieldwork — not King's Fund research, though the King's Fund reported them in March 2026. The Management and Leadership Code for Health and Social Care was launched 14 April 2026, and is a separate thing from the NHS framework and the College of Leadership and Management.

Not used at all. "Around 10% of children's hospice funding comes from the state" — the published figure is 30.1% of charitable expenditure in England (Together for Short Lives, June 2025). "62% of adult hospices in deficit" — use NAO's 64% (2023-24) or Hospice UK's 57% (2024-25). "Palliative care deserts" has no traceable primary source; the real inequality measure is NAO's 19-fold variation in hospice beds per head of the 65+ population between ICB areas. "Total adult social care budget ≈ £20.77bn" is an artefact of dividing an overspend by a percentage. The £77bn and £60bn figures in circulation are gross value added estimates on two different methodologies (Skills for Care 2025, and Alma Economics 2023) — they are economic contribution, not spending, and must never be presented as a budget.

And one correction to this document's own numbers. Adult social care spend "including private spend" was previously given here as ~£44bn. That figure could not be traced. It has been replaced throughout with the LaingBuisson-derived whole-market figure of £35.3bn for England in 2022/23 (local authority £18.5bn, self-funders £11.0bn, NHS £5.9bn), alongside DHSC's £24.7bn net / £29.4bn gross local authority spend for 2024/25. The two are different measures on different dates and should be quoted as such.

On the hospice quality comparison. The 25% Outstanding / 70% Good figures, and the ~6% for NHS acute hospitals and ~2% for care services, all come from a single CQC publication — The state of hospice services in England 2014 to 2017, October 2017. Always attach that date. CQC has published no equivalent cross-sector comparison since, and the 2023 move to the single assessment framework, combined with the inspection backlog above, means most current ratings predate it.

On Buurtzorg in England, which cuts against this document's own enthusiasm. Three English test-and-learn pilots ran between 2016 and 2018 (Guy's and St Thomas', East London in Tower Hamlets, West Suffolk). Independent evaluation found better patient experience but serious implementation failure: the King's Fund's 2019 review of West Suffolk states it "cannot say from this review whether the Buurtzorg model 'works' in the English context", records the team falling from six staff to under two full-time equivalents, and notes that no admissions or cost analysis was completed. The East London evaluation (BMC Health Services Research, 2019) reports mixed success. Do not cite UK Buurtzorg pilots as evidence of reduced admissions or high staff satisfaction — cite the Netherlands original for the model, and use the English pilots as evidence that structure imported without identity and support does not hold.

One scandal to categorise correctly. The Edenfield Centre (Panorama, 28 September 2022) was an NHS medium-secure mental health unit run by Greater Manchester Mental Health NHS Foundation Trust — not social care and not an independent provider. Professor Oliver Shanley's independent review, published 12 January 2024, found unsafe staffing, a closed culture with no psychological safety, incivility between staff and suppression of bad news, and fear-based senior leadership prioritising targets and expansion over care quality. It belongs with Chester and Durham in the argument of this document, not with Winterbourne View and Whorlton Hall in the social care record.

Faith-rooted health systems: what survived checking

A supplied document (Health_FaithBasedSystems) is an AI chat transcript asserting an "absolute, mathematically and culturally verifiable pattern" that faith-rooted systems outperform. Almost none of the performance claims survived; the structural finding did, and Volume 8 now carries it in the corrected form.

Verified and used. The Church of Jesus Christ of Latter-day Saints transferred its fifteen hospitals to an independent non-profit in 1975, retaining no administrative powers — so Intermountain is a secular non-profit with a faith origin, not a faith-based system, and the volume says so. Brent James's clinical process management from 1988, his description of variation work as inquiry rather than verdict ("we couldn't determine who was right or who was wrong, but we could sure show that they were different") and the role of peer normative pressure are from an AHRQ PSNet interview, February 2011. The elective-induction findings — inductions not meeting criteria falling from 28% to under 2%, roughly $50m a year taken out of Utah's cost of care against only $41m of Intermountain's own costs, so about $9m a year of forgone revenue — are from James and Savitz, Health Affairs, June 2011. Fortune/PINC AI ranked Intermountain the top large US health system in 2024. Aga Khan University Hospital's JCI accreditation in 2006, against a Sindh Healthcare Commission that only began licensing in 2022, is verified, as is AKUH Nairobi's fifth successive accreditation in December 2025. St Christopher's Hospice (Cicely Saunders, Sydenham, 1967) originating modern palliative care is verified.

Corrected or dropped. The claim that faith-based providers deliver 30–70% of health care in sub-Saharan Africa does not hold: Olivier et al., Lancet 386(10005), 2015, say the basis "is vague and maybe overstated", household surveys put it below 10% against 30–40% from facility counts, and an independent meta-analysis (Kagawa, Anglemyer and Montagu, PLOS ONE, 2012) puts it at 4.9% outpatient and 6.1% inpatient, tracing the "60% of Nigeria" figure to an unverified personal communication. No WHO or Joint Learning Initiative evidence was found that values-based systems do better on safety, resilience or staff retention — stated in the volume as an evidence gap. The 26% of the world's health facilities figure is a 2010 Vatican press statement counting facilities (including orphanages and pharmacies) with no credible denominator — not used. Ramakrishna Mission does not provide free tertiary care; it publishes a schedule of charges. Intermountain's own savings figures are self-authored, with a generalisability caveat by their author, and no independent peer-reviewed evaluation of its system-wide performance exists — the volume says so. The frequently quoted ARDS survival jump "9.5% to 44%" rests on a 40-patient trial that found no significant difference between arms — not used.

Evidence that runs the other way, and is now in Volume 8. Huft and Guardino, INQUIRY, 2025 (291 Californian hospitals, 2016–2020): religiously affiliated hospitals performed significantly worse than secular ones on several outcome groups at higher revenue levels. Olivier et al. also found faith providers dependent on user fees becoming more expensive for households as public coverage expands. And the counter-case is documented: Ireland's Commission of Investigation (12 January 2021) on ~9,000 child deaths in mother and baby homes 1922–1990s and non-consensual vaccine trials, with Tuam run by the Bon Secours Sisters; the Scottish Child Abuse Inquiry's Smyllum Park findings (11 October 2018), "no love, no compassion, no dignity and no comfort"; Robin Fox's Lancet account of Kalighat (September 1994), care "haphazard", no strong analgesics; the 2022 reporting of Ascension executives boasting of slashing $500m of labour costs; ProPublica's 2019 finding that Methodist Le Bonheur filed 8,300+ medical-debt lawsuits and garnished the wages of 70+ of its own employees; and the Ethical and Religious Directives restricting care in roughly one in six US hospital beds.

The defensible formulation, which is what Volume 8 now argues. An explicit, owner-level commitment to purpose over margin makes better behaviour possible and gives clinicians licence to act on it — but it neither guarantees nor predicts it, and where that commitment substitutes for measurement and independent scrutiny, values-rooted organisations fail at least as badly as anyone else. The transferable feature is constitutional, not devotional, and it is equally available to secular organisations.

The Durham conflict of interest, the £6m and the CQC ratings

Verified against the Aubrey report. The insourced evening and weekend breast clinics were delivered by Durham Surgical Services LLP, established in November 2022, and the Saturday theatre lists by Cathedral Surgical Services Ltd; the Consultant Clinical Lead in breast surgery was a director of both, with a consultant anaesthetist also a director of CSS Ltd. Approximately 50% of all patients referred into the Trust were seen outside standard Trust clinics, either at the independent provider or in the insourced clinics. Payment was per patient, "creating a clear financial incentive", and the review records that "patient selection may have been influenced by throughput and performance metrics ... rather than being driven solely by clinical need". Aubrey's summary finding is "conflicts of interest where consultants held directorial roles in provider LLPs". DSS LLP operated without CQC registration of its own, without a Trust-approved LLP risk assessment and with indemnity cover that was never confirmed; the contract was silent on responsibility for CQC registration and was approved by waiver, bypassing competitive tendering. Contract values in the report: approximately £1.8m ($2.4m) a year for breast care in 2024/25, against actual spend of £1.367m ($1.85m); DSS LLP delivered £1.604m ($2.17m) of activity against £842,460 ($1.14m) of income — a cost-to-income ratio of 133% and a £761,555 ($1.03m) deficit. Capacity: 24 DSS LLP weekend slots at Darlington from September 2022, expanded on 20 January 2023 to 43 weekend slots plus 36 independent-provider slots a week. Contract management was not subject to internal scrutiny until April 2025, and then only because of a patient safety incident.

The £6m is a media figure, not an Aubrey finding. The report contains no cumulative total across the years. "Nearly £6 million had been paid by the trust over six years to breast cancer diagnostic clinics and surgery companies run privately by" the clinical lead is a BBC figure from 2025, relayed by PA (Jane Kirby, 18 September 2026). Attribute it as reported rather than as a review finding, and say "a director of" or "ran privately" rather than "owned": directorship is what the report establishes, and beneficial ownership is not documented in anything checked.

The CQC rating history — the regulator called the Trust well-led throughout. The Trust was last rated "Good" overall in the report published 3 December 2019: Good for effective, caring, responsive and well-led, with safety "Requires improvement", following a well-led review that also concluded Good; the inspection ran 2–4 July 2019. That stood as the standing judgement across most of the period now under review. The CQC then served a section 29A warning notice on 23 December 2025 over the Trust's governance systems, its management of identified risk and its processes for learning from incidents and complaints. Reports published 12 June 2026 downgraded well-led from Good to Inadequate, rated surgical services "Requires improvement" and rated community health services for adults Good; the Trust accepted the findings the same day. One precision: the CQC no longer awards NHS trusts an overall rating, so write "well-led downgraded to Inadequate", never "the Trust was rated Inadequate overall".

Sources for this section: Aubrey, External Review of Governance within the Breast Surgery Service (2025), pp.19, 21, 47, 90–91, 116; CQC provider page for County Durham and Darlington NHS Foundation Trust (RXP); CDDFT statement, 12 June 2026; ITV Tyne Tees, 12 June 2026; PA / The Independent, 18 September 2026.

The "universe of obligation"

Verified, with the name corrected. The term was coined by the sociologist Helen Fein — not "Finn" — in Accounting for Genocide: National Responses and Jewish Victimization During the Holocaust (Free Press, 1979). The definition used in Volumes 3 and 5 is the one quoted from her by Facing History and Ourselves, which uses the concept as a teaching frame: the group within a society "toward whom obligations are owed, to whom rules apply, and whose injuries call for amends". Fein (1934–2022) was executive director of the Institute for the Study of Genocide. 1979 is correct. Describe her as "the sociologist Helen Fein", as the sources do.

One thing to keep in our own words. The gloss that perpetrators exploit the gap left where nobody feels obliged to respond is a fair reading of what the concept describes, and it is the reading Volumes 3 and 5 use — but it is not a Fein quotation, so it is not presented as one. The only words in quotation marks are the definition above.

And one boundary worth stating. Fein was writing about national responses to genocide. Applying her frame to a hospital is an analogy, and the document says so: it is used to name where an organisation draws its circle, not to equate a neonatal unit with a state. The same caution already applies to Staub and Waller in Volume 5 — see the note on the aviation and atrocity analogies above.

Chester after Letby: "Project Countess" and the Gilby tribunal

Verified. Dr Susan Gilby made a whistleblowing complaint in July 2022 about the conduct of the Trust's chair, Ian Haythornthwaite; was offered a secondment in November 2022 conditional on withdrawing it; was suspended on 2 December 2022 and gave notice on 5 December. The Employment Tribunal (Employment Judge Shotter, judgment 13 February 2025) upheld constructive unfair dismissal, automatic unfair dismissal and detriment for protected disclosures, and found a covert campaign called "Project Countess" run through private emails and WhatsApp messages by the chair, the deputy chair Ken Gill, the non-executive director Ros Fallon and the chief people officer Nicola Price. The tribunal found that hundreds of messages and documents had been deleted, that the evidence "points to a deliberate intention to hide documents", and that it was "more likely than not" the chair was involved; it called him an "inaccurate historian" who did not give credible evidence, and Gill a "key player" who was "disingenuous". The chair resigned with immediate effect within days. Thirlwall (27.10) records the judgment as containing "excoriating criticism of the leadership of the Countess at that time". The judgment itself is public on judiciary.uk (Dr Susan Gilby v Countess of Chester Hospital NHS Foundation Trust).

The money: two different figures, used correctly. The damages were £1.4m and the case settled in January 2026; the total cost to the Trust was about £3m once legal costs on both sides are counted (ITV News, 16 January 2026). Do not present £3m as the award, or £1.4m as the total.

On Nicola Price, three precisions. She moved to Mid Cheshire Hospitals NHS Foundation Trust as chief people officer in January 2024 and resigned from that post following fit-and-proper-person concerns. The CIPD imposed its most severe sanction — a Class 6 "permanent lifetime exclusion from CIPD membership" — for failing to investigate a grievance properly, participating in the effort to remove an employee and delete documentation, and dishonesty in the tribunal proceedings; the case is listed on the CIPD's published breaches-of-code page. But: CIPD membership is not a licence to practise, so write "permanently expelled from her professional body", not "struck off" — the stronger claim is also the inaccurate one. And retirement is not established: she resigned, and nothing found says she was forced to retire, so the document does not say it.

Still live, and separate. As at June 2026 Ken Gill held a non-executive post at the Legal Aid Agency, with Dr Gilby's lawyers calling for his removal on Nolan-principles grounds; that is a request, not an outcome. Keep all of this apart from the criminal investigation — arrests have been made, nobody has been charged, and no individual is linked here to any arrest.

Sources for this section: judiciary.uk (judgment); Personnel Today, 14 February 2025 and 16 January 2026; ITV News Granada, 18 February 2025 and 16 January 2026; BMA media statement on the judgment; CIPD breaches-of-code cases page; Roger Kline, A whiff of accountability for unacceptable behaviour; Legal Futures, 29 June 2026.

Johnson & Johnson: the Credo, Burke and Tylenol

Verified. Our Credo was written in 1943 by General Robert Wood Johnson and presented to the board that December, a year before the company went public; it appeared in his 1944 book But, General Johnson… as "An Industrial Credo" and was renamed "Our Credo" in 1948. Order of responsibilities: patients, doctors, nurses and all others who use the products; employees; communities; stockholders last — "Our final responsibility is to our stockholders. Business must make a sound profit." Credo Challenge meetings: 1975, organised by James E. Burke so that management could discuss whether the values in the Credo were still relevant; repeated on the Credo's 70th anniversary in 2013. Credo Survey: 1985, asking employees worldwide to rate how the company was living up to each responsibility. Burke was chief executive 1976–1989. Tylenol: autumn 1982, seven deaths in the Chicago area from cyanide placed in Extra-Strength capsules; 31 million bottles recalled at a cost of about $100m, ahead of any regulatory requirement; relaunched two months later in tamper-evident packaging. Tylenol held about 35% of a $1.2bn analgesic market, fell, and was back to 30% by mid-1983 and 35% by the end of it. Burke's line — the Credo "gave me the ammunition I needed to persuade shareholders…to spend the $100 million on the recall" — is from Wharton's account of his career.

One quotation deliberately not used. The widely repeated Burke line about either committing to the Credo or tearing it off the wall could not be traced to a primary source. J&J's own heritage material describes the 1975 meetings as management discussing "whether the values in Our Credo were still relevant", and that is what the document says.

One date to get right. Wharton's piece says the credo was "established in 1887", which conflates it with the company's founding era; J&J's own heritage pages and the 1943 text itself give 1943. Use 1943.

And the counter-case is current, which is the point of including it. In February 2022 J&J agreed to pay up to $5bn towards US opioid settlements, on terms it states are "not an admission of any liability or wrongdoing". In July 2026 it proposed $5.5bn to settle roughly 76,000 remaining ovarian-cancer claims over talc, after a bankruptcy court rejected an earlier $9bn plan; the company maintains it would have prevailed with further litigation. Present these as settlements and litigation positions, not as established causation — the argument they support is about what happens to values that stop being maintained, and it does not need any claim about the science.

Sources for this section: jnj.com, Our Credo and 8 Fun Facts About the Johnson & Johnson Credo; Knowledge at Wharton, Tylenol and the Legacy of J&J's James Burke; J&J press release, 25 February 2022; Fortune, 29 July 2026.

Founding ethos, the care emergency, and the workforce (21 September 2026)

Three further supplied transcripts — an expanded NHSFoundingEthos, UKSocialChildrensCareScandals and BMA_And_ResidentDoctors — were treated as leads and checked independently. The care transcript proved more accurate on 2025–26 events than earlier ones; the founding transcript contained several invented quotations.

Invented or unsupported, and not used. The Sylvia Diggory lines "There was a tremendous buzz… You could feel the relief in the room" and a direct-speech "Do you understand what is happening today?" (her genuine account is in reported speech: Bevan told her it was "the most civilised step any country had ever taken"). A quotation attributed to Dr Julian Tudor Hart about 5 July 1948 — he was a student in 1948 and began practice in 1961. An unnamed "East End GP" on five shillings. Yvonne English on "huts" and "grass skirts" and "we became like sisters" (her verified words are "unity is strength" and "we need to leave something behind to show that we made our contribution", Yorkshire Post, 2019). A nurse named "Rachel Serendo", who cannot be found. Bevan calling limits on care for foreigners "un-Christian". And "Illness is neither an indulgence…", traced to a 1967 paraphrase by T. H. Marshall rather than to Bevan — now removed from Volume 10.

Verified, with precisions. Spectacles: "about 4½ million pairs" supplied in the first year and three to four million on order (Hansard, 28 July 1949) — not "5.25 million in nine months"; no reliable denture count. "It's a piece of real socialism. It's a piece of real Christianity too, you know" exists on undated film; nothing supports "1959". "Tredegarise" is attributed, with no primary source. The BMA plebiscite of early 1948 went against the scheme by roughly nine to one (the exact count needs the BMJ primary); "Medical Fuehrer" was Alfred Cox. Windrush arrived at Tilbury on 21 June 1948 and passengers disembarked on 22 June; that nurses were aboard is not established. The 54,000 nursing vacancies figure is widely cited by historians but not from a primary government source. Recruitment, the 3,000–5,000 Jamaican nurses by 1965 and the SEN channelling are from Stephanie Snow and Emma Jones (History & Policy). Daphne Steele: British Guiana, arrived 1951, matron of St Winifred's in Ilkley in 1964, "the first Black matron in the National Health Service" (Historic England). May Parsons, from the Philippines, gave Margaret Keenan the first vaccination on 8 December 2020 (RCN; Science Museum). About 21% of NHS England staff report a non-British nationality (Commons Library, June 2025). Infant mortality 12 per 1,000 in 1980 and 3.7 in 2021 (Commons Library); no verified 1948 figure is used.

Care: corrections to the transcript. Lord Mann made 24 recommendations, not "36 structural changes". The IICSA Nottinghamshire report was 2019, not 2022. At Orchid View neglect contributed to five of nineteen deaths. There is no CQC "revenge eviction tracker"; the evidence is ITV's 2023 investigation. The claim that Innovative Start "wiped" twelve incidents is not supported (the CQC found failures to notify, rated it Inadequate in 2025 and Good in June 2026); it is not used. "Up to £40,000 a week" is Public First's upper figure; Ofsted says up to £30,000 — both are named. Prospero is a staffing agency, not a home operator. Blue Ribbon Healthcare runs supported living and learning-disability services, not ordinary care homes, and nothing has been determined. Edenfield staff are charged, not convicted. The Casey commission has not yet reported.

Care: verified as stated. Southport Phase 1 (13 April 2026, 67 recommendations accepted 3 July 2026); Catalyst Care (Ofsted's first successful prosecution, fined £92,400 in August 2026); the Children's Wellbeing and Schools Act 2026 (Royal Assent 30 April 2026; profit powers are a backstop, not a cap); TEWV inquiry chair appointed 17 September 2026; Common Wealth's 11 of 20 (3 September 2026); TBIJ's £255m to 480+ companies (8 September 2026). Anthony May: Nottinghamshire Director of Children's Services by 2009; inadequate safeguarding 2010, notice lifted 2011, Good 2015; Chief Executive from April 2015; NUH from September 2022; NUH leadership Inadequate (2021) to Requires Improvement (2023); OBE June 2023. The IICSA finding that the Councils did not learn from their mistakes is stated alongside. Greater Manchester: the measured gains are modest and mixed, and are presented that way.

Workforce: verified, with the contested points flagged. Renaming 18 September 2024, 91% support. Settlement 29 June 2026: 52.9% accept on 57.3% turnout. The pay fall depends on the index — 26% (BMA, RPI, 2008/09–2021/22), 4.7% CPI or 17.9% RPI (Nuffield Trust), 6–7% HCI (Full Fact) — and all are given. Student debt "over £100,000" is an upper-range BMA figure, not an official average. Strike costs are estimates; no audited total exists. The smoking–lung cancer link is not credited to the BMA. The BMA's June 2026 IHRA motion is as reported by the Jewish Chronicle; the motion text and vote are not public.

David Bevan's death, 1925. Verified: he died of pneumoconiosis in February 1925, and no compensation was paid because it was not then classed as an industrial disease under the Workmen's Compensation Act (Wikipedia, citing Michael Foot's biography, 1962/1966, pp. 11–14). That he died in Aneurin's arms, and was asking for the Daily Herald and news of John Wheatley, is Bevan's own later account as reported by Spartacus Educational and others — hence "as his son later recalled" in the founding annex; Foot's text itself was not read. Aneurin was born on 15 November 1897, so he was twenty-seven, not twenty-eight, in February 1925.

Bevan's last year. Verified: elected Labour deputy leader unopposed in 1959; admitted to the Royal Free on 27 December 1959 for expected ulcer surgery, when stomach cancer was found; home on 14 February 1960; died at Asheridge Farm, Chesham, on 6 July 1960 (Wikipedia, citing standard biographies). "We had to wait a long time for it" and the "bankrupt nation" line are in the same undated film transcript as "real Christianity" (Speakola, which asks for the date and place) — not dated to 1959. Not used: that he was never told his diagnosis; that he spoke of his care as a patient or praised his nurses, or that Jennie Lee recorded this — none traced to a source. "Faith to fight for it" remains untraceable; the supplying transcript itself calls it oral history, and its citations are social media and campaign sites. The founding annex now says plainly that it is not his words, while agreeing with it.

Bevan the man. Verified (Dictionary of Welsh Biography; Gwent Archives; Nuffield Trust; Hansard; Open University Jennie Lee archive; Chesham Museum; Blaenau Gwent council): born 15 November 1897, sixth of ten children (sources differ on how many survived); left school about 13 and underground in 1911; severe childhood stammer, managed by vocabulary, recitation, elocution and public speaking; Central Labour College 1919–21 on a miners' union scholarship; Tredegar council 1922, county council 1928, MP for Ebbw Vale 1929–60. Tredegar Medical Aid Society: formed 1890, about 3d in the pound, around 20,000 covered by 1925; Bevan on its hospital committee from 1924, chair 1929–30. Married Jennie Lee 1934; the OU archive records that she "did not tell Nye or their friends of the seriousness of his condition" — whether he ever knew is unresolved. Churchill's "squalid nuisance" is Hansard, 6 December 1945. Memorial service, Westminster Abbey, 26 July 1960; Waun-y-Pound stones unveiled by Michael Foot, 1972. Softened: he headed the library's book-selection committee (not "chairman of the library"); the Query Club is described as a group he was a leading member of, since the council credits others as founders; the "Methodism than Marx" line is Morgan Phillips's 1950 point in later, neater words; "Tredegarise" is printed in the BJGP (2018) but no primary record has been found. Corrected and not used: "2,751 hospitals" — the sourced figure is 2,688 in England and Wales (1,143 voluntary, 1,545 municipal; Nuffield Trust); "the duchess and the dustman" is a paraphrase, not Bevan's words (his are "universalise the best" and "the rich and the poor are treated alike"); a telegram reading "sad hearts in the valleys tonight" and Churchill sitting alone in the chamber after the tributes, neither traced to a primary source; any permanent Bevan memorial in the Abbey (there is none — the Abbey memorial is to Ernest Bevin).

The Champions Pathway and the 3.5% rule. The Choice → Connection → Creation reading of Bevan's life is the author's lens, and the founding annex says so. The 3.5% figure is Erica Chenoweth's: "no government has withstood a challenge of 3.5% of their population mobilized against it during a peak event" (Harvard Carr Center, April 2020), from the dataset behind Chenoweth and Stephan, Why Civil Resistance Works (2011). Chenoweth calls it "a descriptive statistic… not necessarily a prescriptive one", notes exceptions (Bahrain 2011–14), and stresses organisation and momentum; the annex reports those caveats. It is a finding about national political campaigns, applied here by analogy to organisational culture. "Never doubt that a small group of thoughtful, committed citizens can change the world" is widely attributed to Margaret Mead, but the Institute for Intercultural Studies has found no documented source, so it is not used.

Why being human is good for health. The OECD's 2013 Guidelines on Measuring Subjective Well-being and its December 2024 working paper on eudaimonia were read at source; the health claim in the latter is the literature's, which the OECD cites rather than tests, and the annex says so. The purpose-and-mortality finding is presented with its challenge: Sias and Turtle (PLOS ONE, May 2026) re-analysed the US Health and Retirement Study sample and found the effect largely disappears when baseline health is measured well. The social-connection evidence (Holt-Lunstad 2010 and 2015; WHO Commission on Social Connection) is the most robust, but it is observational. Continuity of care (Sandvik, BJGP 2022), the Kaptchuk warmth trial (BMJ 2008) and the relationship meta-analyses (Kelley 2014; Howick 2018) are cited at their published effect sizes, which are modest. Not used: "40 seconds of compassion" as a clinical effect; Hojat's empathy-and-diabetes findings as established (not independently replicated); Rakel's cold-duration subgroup results as a general effect.

Why being human — older people and children (section 6). Checked against abstracts: Inouye 1999 (9.9% vs 15.0%; matched, not randomised); Hshieh 2015 and 2018 (delirium OR 0.47; savings $1,600–3,800 per patient); WHELD, Ballard 2018 (d≈0.23–0.30; no reduction in already-low antipsychotic use); CADRES, Chenoweth 2009 (agitation down; falls up with person-centred care); Fossey 2006 (antipsychotics 42.1% → 23.0%, agitation unchanged); Steptoe 2013 (isolation HR 1.26; loneliness HR 0.92, not significant); Temel 2010 (16% vs 38% depression; 33% vs 54% aggressive care; 11.6 vs 8.9 months, a secondary outcome); Haun, Cochrane 2017 (QoL SMD 0.27; survival not confirmed); Rosa 2019 (delirium 18.9% vs 20.1%, null); Coleman 2002 (Eden: no benefit, more falls); Mead 2010 (befriending SMD −0.27 from 9 studies); WHO iKMC 2021 (12.0% vs 15.7%, RR 0.75); Conde-Agudelo, Cochrane 2016 (RR 0.60; sepsis 0.35; hypothermia 0.28); O'Brien 2018 FICare (breastmilk 70% vs 63%); Rubin 2007 (22% → 36%, a 63% increase — not "36–63%"); Bethell 2019 (aOR 0.28); Birnie, Cochrane 2018 (SMD −0.56, low certainty); Robling 2016 (FNP null on all four primary outcomes). Not confirmed from the abstract and so not quoted as numbers: the Bucharest IQ figures (commonly given as ~81 vs ~73) and the "before 24 months" threshold — the annex says only that gains were largest for the youngest placed. Karver 2018 (r≈0.19) and the Cochrane review of parental presence at induction were located but not re-checked. Not used: "loneliness is as deadly as 15 cigarettes a day"; palliative care "extends survival" as established; "positive ageing adds 7.5 years" as cause; Spitz and Harlow as evidence rather than history.

The care emergency — money, and the join with the NHS. Adult social care overspend £715m in 2025/26, 3.2% of £22.6bn; £774m in 2024/25; 80% of councils overspent both years; adult care 35.5% (2025/26) and 37.9% (2026/27) of whole council net budgets (ADASS Spring Surveys 2025 and 2026). Note: the "about 3%" is a share of adult social care budgets, not of total council spending, and it covers adult care only; children's social care overspent by £2.3bn (21%) in 2023/24 (Institute for Government) — no verified national figure for later years was found. Adult and children's care about 65% of budgets (County Councils Network, March 2024; IFS 2024). Balanced-budget duty: Local Government Finance Act 1992 s.31A (billing authorities); s.114 notices under the 1988 Act. Exceptional financial support 2026/27: 35 councils, ~£1.5bn (February 2026); gov.uk list updated 18 August 2026 shows 36. Delayed discharge: "nearly 13,000" a day, £2.7bn, £562 a bed day (King's Fund, 7 July 2026); 11.0% of bed days (Health Foundation); responsibility last attributed in February 2020 (NHS 60%, social care 30%, both 10%). Northern Ireland: integrated since 1973 (HPSS (NI) Order 1972); NIAO January 2024 (50.9% vs 5.4% over 52 weeks, December 2022); DoH NI 31 March 2026; Nuffield Trust 2019 and 2021 quotations verified; COPNI February 2026 (26.17 → 20.60 packages per 1,000 aged 60+). The 40% care-home share of lost bed days is from departmental figures reported by the Irish News (secondary). Bengoa 2016 is cited by title only; no verbatim quotation was verified. Audit Scotland briefing, January 2026; NAO, February 2017.

Volume 8 — Leeds and Manchester. From Ofsted reports and DfE evaluations. Leeds: safeguarding inadequate January 2010, improvement notice March 2010 to December 2011; good 2015 (leadership outstanding; restorative practice "having a transformational impact on culture and professional practice"); outstanding 2018, 2022 and 2025; Nigel Richardson DCS for about six years to September 2016; Child Friendly Leeds launched 2012; child protection plans down nearly 50% 2011–17 (DfE, April 2019); CP rate 31.8 vs 55.8 per 10,000 in 2018–19 (Partners in Practice evaluation, April 2020, which notes the lack of a comparison group); Family Valued evaluation (Mason et al., 2017) — significant falls in looked-after children and CP plans, "too early to tell" on family outcomes. Manchester: inadequate 2014 (486 cases awaiting assessment; 1,406 children in care); Paul Marshall director from March 2016; requires improvement with leadership good 2017 (40% more frontline social workers; Signs of Safety); good 2022; outstanding, published 11 November 2025 ("Social workers talk about children with warmth and kindness"). Both asked to improve their response to neglect. Strengthening Families, Protecting Children: £84m over five years (DfE, April 2019), spreading Family Safeguarding, No Wrong Door and Family Valued. Not verified: Leeds looked-after numbers before 2014; Manchester's current children-in-care and turnover figures.

Volume 5 — ripples and signals (opening). Heinrich, Industrial Accident Prevention (1931): 1 : 29 : 300 from 75,000+ insurance reports; original data lost, ratios vary by industry, severe and minor events often differ in cause. Bird (conventionally 1969): 1 : 10 : 30 : 600 from ~1.75 million reports, 297 companies. Baker Panel (2007) on Texas City, verbatim as quoted. Hickson et al., JAMA 2002 (645 physicians; complaints predicted risk-management files and lawsuits). Cooper et al., JAMA Surgery 2019 (13,653 patients, 202 surgeons; ~18% higher complications with 1–3 coworker reports, ~32% with 4+) — figures from Vanderbilt's and Fierce Healthcare's accounts, full text not read. Vanderbilt pyramid: most physicians respond to the informal conversation (~97% by Vanderbilt's account). Hutchinson et al., QSHC 2009: higher-reporting trusts had better safety culture and risk-management ratings, no link to mortality ratios. Caveat: no healthcare study was found showing that near-miss counts predict serious harm in Heinrich's sense; the healthcare evidence that small signals predict harm is at the level of the individual clinician (complaints and coworker reports).

Annex 8 · Loneliness. IPPR, The Loneliest Generation (1 September 2026): no close friends 1 in 100 (2014) → 1 in 20 (2026); psychological distress 47% of young women, 36% of young men aged 16–24; Morjaria quote verified. The release does not name the data source; press reports say Understanding Society (run by ISER, Essex — not NatCen as one report said) and give the age band as 16–21. Confirm before external use. Community and Engagement Survey 2025/26 (DCMS, 14 August 2026): 7% often/always lonely; 16–24 9%, 65–74 4%. Health Survey for England 2024 (27 January 2026) runs the other way on chronic loneliness (16–24 lowest, 4%) — stated in the annex. WHO Commission on Social Connection (30 June 2025): 1 in 6; 17–21% of 13–29s; 871,000 deaths a year is a modelled estimate. Meta–Gallup (2023): 27% of 19–29 vs 17% of 65+. US Surgeon General (2023): 150 → 40 minutes a day with friends, 2003–2020 (pandemic year). Wang et al., Nature Human Behaviour 2023 (HR 1.32 isolation, 1.14 loneliness); Valtorta et al., Heart 2016 (mostly isolation measures); Luchetti et al. 2024 dementia (secondary); Loades et al., JAACAP 2020. Sirois & Owens 2021 (GP visits); Wilding et al., BJGP 2025 (social prescribing, no effect on loneliness); GMC, Caring for doctors, caring for patients. UK policy: minister January 2018; A connected society October 2018; no published update since 2023. Not used: "as bad as 15 cigarettes a day" (Holt-Lunstad says it aggregates all social-connection measures); "9 million lonely Britons" (polling, not official statistics); Campaign to End Loneliness GP-poll figures (unverified); INTERACT age percentages (describe the sample, not prevalence); any causal claim about social media. The BBC article supplied as a source could not be accessed and is not relied on.

Nicola Price, after Chester. Appointed Chief People Officer at Mid Cheshire Hospitals from January 2024 (the trust's own announcement, 31 October 2023). She no longer appears on that trust's board, where the Chief People Officer is now Alfredo Thompson, and her public professional profile lists her as retired. That she resigned or left because of the February 2025 tribunal, and the date she left, were not confirmed from a primary source — hence "following the tribunal's findings she is no longer there". "Forced into early retirement" and "struck off" are not used; the CIPD sanction is a permanent expulsion.

Annex 9 · Joining up health and social care. Nuffield Trust, Change or collapse (July 2019): "more than 3,000 times as likely" (NI vs England, over-a-year waits); NIAO, Tackling Waiting Lists (January 2024): 26.3% vs 12.4% of population waiting, 50.9% vs 5.4% over 52 weeks; NISRA, 31 March 2026: 510,680 waiting, 54.1% over 52 weeks; RCEM (23 April 2026): 30.5% within four hours, 23.5% over twelve hours in January–March 2026 (no four-nation comparison in that release); Full Fact (July 2024) on comparability and the Encompass data gap; Nuffield Trust, Integrating health and social care (December 2021). DoH NI (3 June 2026) waiting-time progress is the department's own account. Torbay: King's Fund (March 2011), with its causation caveat. Greater Manchester: Britteon et al., Lancet Public Health 2022 (0.196 years, 95% CI 0.182–0.210) and Social Science & Medicine 2024 — multiple causes, not integration alone. Audit Scotland (November 2018; July 2024). NAO (February 2017). Health Foundation (October 2021; February 2023). Baxter et al. 2018 (NIHR report); Liljas et al. 2019. Burnham: gov.uk speech, 29 July 2026; Hansard written statement, 30 March 2010; King's Fund speech, 24 January 2013 (transcript via the Socialist Health Association). Not used: any claim that NI is worst on life expectancy, avoidable mortality or cancer targets (not sourced); precise four-nation league tables for late 2023–24; any claim that Burnham proposed free personal care or an NI-style merger in 2026; Canterbury (NZ) figures, which come from programme insiders.

Added 21 September 2026. Loneliness annex: Professor Peter Fonagy's comments on the forthcoming independent review of mental health, ADHD and autism are taken from a BBC report of 21 September 2026 (read from a copy supplied by the author); the review was unpublished at the time of writing, so its figures should be checked against the published report. Care annex, Oxford: the 370-plus figure and the quotations on disbelief are from the Oxfordshire Safeguarding Children Board's 2015 serious case review press release; the inquiry's chair, start date and scope are from the Independent Inquiry into Grooming Gangs' own website; survivors' and the former detective's words are from BBC Panorama, 21 September 2026; the council's response is from Oxfordshire County Council's statement of the same day. Vol 10, Chicago pilot: the evaluator's words ("tangible performance lift", "directional") are from page 4 of the Union Health Service 90-day review pack, which carries the attribution to Lauren Easterday, EastLight; the title Humanity as a Performance Multiplier is the pack's own, and the pack notes it includes additional observations by Champions.

Added 22 September 2026. Executive summary, Chester: "a complete failure to protect babies on the neonatal unit" and "dysfunctional management and governance; a gulf between hospital leadership and clinicians" are from the Thirlwall Inquiry's statement on publication of its report, 15 September 2026. The "revolving door" quotation is from the Inquiry's summary report, Part Two, paragraph 20; the recommendation on senior managers moving between trusts is Recommendation 13 (Volume III, Chapter 45).

Care annex, Lincolnshire: the £12.6m forecast, the £18,000 a week and the contingency figures are from council committee reporting in Lincolnshire Online (28 August 2026) and Hits Radio Lincolnshire (21 September 2026); the £9m children's social care figure, the £51m schools overspend and Cllr Natalie Oliver's words from Lincolnshire World (4 September 2026); the 61 per 10,000 rate from Lincolnshire Online (7 September 2026). The source note supplied with it was an AI search transcript; its unsourced claims (named providers, "112 children", a 1,000% rise to 2026) were not used. National figures: DfE Children looked after in England 2025; NAO Managing children's residential care, 12 September 2025; LGA high-cost placements survey, 29 November 2023; Become, 20 November 2025; Institute for Government Performance Tracker 2025; Children's Charities Coalition / Pro Bono Economics, 3 September 2024; YMCA, 10 February 2026; County Councils Network, 9 February 2026.

Care annex, Teesside: Middlesbrough's 2023 rate of children in care (150 per 10,000; 513 children) and the North East's 2024 rate (114) are from Blue Cabin's summaries of DfE data; Middlesbrough and Thornaby East child poverty from End Child Poverty, 2025; Ofsted's 2020 findings as reported by Community Care, 27 January 2020, and the 2026 judgement as reported by Teesside Live; violence figures from the Cleveland Police and Crime Commissioner's Strategic Needs Assessment (data October 2017–October 2022); drug deaths from ONS, 2024 registrations; the A66 crash from ITV News Tyne Tees, 23 August 2026, and the occupants' previous offences from ITV News Tyne Tees, 24 August 2026, citing court records and Cleveland Police (the count of "more than thirty" is our tally of the offences listed there; UnHerd gives 37 convictions); the extra officers from Teesside Live; residents' fear and the police numbers from UnHerd, 31 August 2026 (a feature, and the police figures should be checked against Home Office workforce statistics before wider use); care cliff figures from Become. The estates UnHerd describes (TS6) lie mostly in Redcar and Cleveland borough, not Middlesbrough borough; the text therefore speaks of Teesside.

Annex on applying lessons from the past, the Northern Care Alliance: the Section 29A warning notice of 13 August 2026 and its wording are from the CQC's own provider record; the June 2026 enforcement wording, the Good Governance Institute review and its 43 recommendations from HSJ (23 June 2026) and The Independent (19 June 2026); segment 4 from the trust's own statement of 12 December 2025; the consultant survey, the corridor and incident-downgrading allegations and the Good Employment Charter membership from Channel 4 News (19 August 2026), which describes a survey of 400 consultants taken in November 2025 and leaked to it and to The Independent — The Independent's earlier account (21 June 2026) of a related poll of about 426 consultants gives different proportions, and neither identifies who commissioned it; the no-confidence threat from The Independent (21 June 2026); the gynaecology harm from The Independent (17 May 2026) and the further review from The Independent (19 August 2026); the strike, the critical incident and the suspended walkout from UNISON, ITV News Granada (24 August 2026) and the Oldham Evening Chronicle; the 2021 transfer from NHS England's own publication; the ratings from CQC reports; income and deficit from the 2024/25 provider accounts; the emails to the mayor's office from Channel 4 News. The source note supplied for this annex was an AI search transcript: its unverified claims — a named recovery director, specific deficit figures, a reported rise in stillbirths at one hospital, and a quotation attributed to a former medical director at Chester — were checked and left out. Dalton material is from the CQC reports of March 2015 and August 2018, Salford Royal's annual reports, the Dalton Review (5 December 2014) and Building a culture of candour (March 2014); walkround evidence from Sexton and colleagues, BMJ Quality & Safety 2018, and Martin and colleagues, Joint Commission Journal on Quality and Patient Safety 2014; merger evidence from Gaynor, Laudicella and Propper 2012, the Health Foundation 2014 and the Health Foundation's Leading NHS groups, August 2025. The Visitation material is from the Catholic Encyclopedia and the order's own federation; the claim, in circulation, that its constitutions required correction to be private could not be verified and is not used.

Annex on radical leadership: aviation from FAA Advisory Circulars 00-46F (reporting protection and the ten-day rule), 120-66C (the Aviation Safety Action Program) and 120-51E (crew resource management), 14 CFR 91.3, and NASA's ASRS programme briefing of December 2025; TeamSTEPPS and the CUS wording from AHRQ; the simulation study from Pian-Smith and colleagues, Simulation in Healthcare 2009; incident reporting evidence from Stavropoulou and colleagues, Milbank Quarterly 2015; DuPont from the company's own history, Hagley Museum and Library, Klein's history of DuPont process safety (2009) and the US Chemical Safety Board's investigations at Belle (2010) and La Porte (final report, 2019); Toyota's andon description from Toyota's own account of the Toyota Production System; Virginia Mason from Furman and Caplan, Joint Commission Journal 2007, with the Lean caveat from Moraros and colleagues, International Journal for Quality in Health Care 2016; Martha's Rule and its figures from NHS England, including the eighteen-month data published in May 2026; PSIRF, the withdrawal of the 2018 just culture guide and the Being fair tool from NHS England; the incident decision tree from the National Patient Safety Agency via AHRQ; the "no blame" findings from the Thirlwall Inquiry summary report, Part Two, and Volume III, Chapter 39; manager barring from the DHSC consultation response of 21 July 2025 and the government's response to Thirlwall of 15 September 2026. Where a popular version of a story is wrong — "immunity" for a self-report, a legal duty to take the controls, an andon cord stopping a factory — the annex says so.

Added 22 September 2026 — Annex 10 · Why we all love the Welsh. Built from an AI transcript supplied by the author (Research/BevanWasAChampionofHumanity.docx) and checked claim by claim. Verified and used: Griffith Jones (born 1683, Penboyr; shepherd; Llanddowror 1716; circulating schools from 1731; 3,495 schools and over 158,000 scholars by 1761 — Dictionary of Welsh Biography; "over 200,000" of about 480,000 and "one of the most literate countries in the modern world" — E. Wyn James, Cardiff University; report by a commissioner for Catherine II in 1764 — DWB; five pamphlets against him 1745–52 — DWB; "morning star" — Dictionary of Methodism in Britain and Ireland); Bridget Bevan (patron, continued the schools to 1779, £10,000 bequest — DWB); Whitefield and Harris, Cardiff, 7 March 1739; Watford 1743; 1851 census, 80% of worshippers Nonconformist; 1847 Blue Books quotation (National Library of Wales); wealh (Bosworth–Toller) and Cymry (etymonline); Maes-yr-Haf 1927; Tredegar Workmen's Medical Aid Society 1890, 3d in the pound, cottage hospital 1904, central surgery 1911, Bevan on the hospital management committee from 1924 and chair 1929–30 (Gwent Archives); institutes funded by 1d–2d in the pound and Bevan heading Tredegar's book-selection committee (Jonathan Rose, The Intellectual Life of the British Working Classes); Bevan's birth, family and parents (DWB; Spartacus Educational); Evan Roberts and the 1904–06 revival (DWB; Christian History Institute); Hansard 2 July 1942 and 6 December 1945; Nicolson diary (via Lord Lexden); "lower than vermin", 4 July 1948, and "Minister of Disease", 10 July 1948 (History Today; International Churchill Society); "Do not undereshtimate your husband" (Jennie Lee, My Life with Nye, 1980, via Finest Hour 193); Jennie Lee's "great humanist" line (as quoted by Humanists UK — original publication not located). Corrected or not used: Bridget Bevan was not Jones's sister-in-law; "a burning and shining light" was Whitefield on Howell Harris, not Jones; Jones's smallpox, asthma and father's death before his baptism (no source); a Whitefield–Jones meeting in Bath (no source); "the highest literacy in Europe"; Bevan's "purpose of power is to give it away" (no primary source); 7 Charles Street and its bathroom; the father "loving poetry"; the family living on a sister's wages; the Medical Aid Society paying for his father's care; grandparents' names; Evan Roberts in Tredegar in January 1905 and the Tredegar recording; the chapel eisteddfod prize; Churchill sitting alone in the chamber after Bevan's death; Manic Street Preachers lyric.

Added 22 September 2026 — Annex 11 · Why a top-down model fails. Built from a second AI transcript supplied by the author (Research/WhyWeLoveAndNeedTheWelsh.docx) and checked claim by claim. Verified and used: William Williams MP's motion, 10 March 1846, citing "the Newport and Rebecca riots" (Hansard); the commissioners — Lingen (b. 1819), Symons (b. 1809), Vaughan Johnson — non-Welsh-speaking Anglican laymen, Welsh-speaking assistants, reliance on gentry and Anglican clergy (National Library of Wales; Pembrokeshire Historical Society; Genuki); Lingen Permanent Secretary to the Treasury 1869–85 and a peer; Y Gymraes, 1850, Ieuan Gwynedd, Elen Egryn's introduction; the Welsh Not predating 1847, a local practice, not state policy (Martin Johnes, Welsh Not, 2024); Wilberforce on trade unions, the 1799 Combination Act, 1795, 1817 and 1819 (History of Parliament); Darby, Coalbrookdale 1709; Bournville; Macaulay's Minute, 2 February 1835; Knox, The Races of Men, 1850 (quotations from the text); Great Famine deaths and emigration and Trevelyan's letter (general reference sources — an academic source would be better for print); Rothschild 1847, Lords' rejection 1848, Jews Relief Act 1858; Aliens Act 1905 (National Archives); Galton 1883; Eugenics Education Society 1907 (English Heritage); Churchill's December 1910 letter (International Churchill Society); Mental Deficiency Act 1913 and the 10,000 to 60,000 estimate (UK Disability History Month — a secondary figure); Ely Hospital and the Howe report, March 1969, and the 1971 white paper; Winterbourne View 2011 and Whorlton Hall 2019/2023; Chadwick 1842, Public Health Act 1848, the Great Stink 1858, Bazalgette 1865–75 (UK Parliament; London Museum); Beaufort Research/Golley Slater survey, March 2022 (used in the Welsh annex). Corrected or not used: the claim that the Blue Books led to the Welsh Not; "three young barristers" (only two are documented as barristers); "Christianity, Commerce and Civilisation" as a Clapham Sect slogan (it was Livingstone's, 1850s); any suggestion that the commissioners were anti-Jewish (no evidence); Churchill as Home Secretary when the 1913 Act passed (he was not); the claims that limited liability "disconnected conscience" from business and that paternalism alone caused the fall of the Empire (interpretations, not facts); Ely as "the first" NHS scandal ("one of the first").

Annex on enablership: the framing is from Barrie Watson, The Art of Enablership: Rethinking Leadership for a Complex World (12 April 2026), quoted from the book's published description; Madison, Hamilton and Washington from the Federalist Papers (Avalon Project), the Farewell Address and Washington's letter of 9 January 1790, with his resignation words as recorded by Mount Vernon; Collins from his own published concepts and Harvard Business Review, January 2001; NHS chief executive tenure from the King's Fund, Leading in today's NHS (2018); the chief-executive effect from Quigley and Hambrick, Strategic Management Journal (2015); programme sustainability from Zurynski and colleagues, Implementation Science (2023). The line "no man is indispensable" is not in the Farewell Address and is not used.

Annex of examples: spread and adaptation from Greenhalgh and Papoutsi, BMJ (2019), Perez Jolles and colleagues, Journal of General Internal Medicine (2019), and the Health Foundation's The Spread Challenge (2018); Tredegar from Gwent Archives, the British Journal of General Practice (2018) and Steve Thompson (Aberystwyth, 2018); Elizabeth Fry from Huntsman, Bruin and Holttum, Medical History (2002), quoting the Institution's own report of 1848 — her institution was founded in 1840 as the Protestant Sisters of Charity, Guy's took its first two trainees in September 1841, and no documented acknowledgement by Florence Nightingale could be found, so none is claimed; Mayo from the clinic's own heritage pages (born Salford, 31 May 1819; the primary value is Mayo's distillation of the 1910 Rush address, not a verbatim quotation); Esther from the Commonwealth Fund case study and working paper, BMC Health Services Research (2023) and the International Journal of Integrated Care (2019) for Kent, where no quantified English outcome has been published; Buurtzorg from the Commonwealth Fund (2015), Buurtzorg's 2024 quality report, Lalani and colleagues (2019), the King's Fund's West Suffolk review (2019) and Hegedüs and colleagues (2022) — the widely quoted "40% saving" is a misreading of Ernst & Young's finding that Buurtzorg used about 40% of the authorised care hours, and KPMG found 108 hours against 168 with total case-mix-adjusted cost close to the national average; Virginia Mason from Furman and Caplan (2007) and, for the counterweight, the Washington State Nurses Association (2023); Nuka from the King's Fund (2015) and the Baldrige profiles (2011 and 2017), whose emergency-department figures differ (more than 45% and 50% respectively); Gundersen from its own releases and the Respecting Choices programme history — the often-quoted 96% is the share of existing advance directives found in the record, not the share of residents holding one, which was 85% and later 90%; ELFT and Northumbria from CQC reports, NHS Staff Survey benchmark reports, the trusts' own publications, the Institute for Healthcare Improvement and Price and colleagues, Emergency Medicine Journal (2020); Bromley by Bow from the Centre's annual review (self-reported outcomes) and the King's Fund; social prescribing from Wilding and colleagues, British Journal of General Practice (2025), Bu and colleagues (2025, 2026) and O'Sullivan and colleagues, Clinical Rehabilitation (2024).

The Harvey board remark. The words "There is no doubt this team has been tested. It's all Tony's fault", with the exchange that follows, are quoted from the Daily Mail's profile of the Countess of Chester's leadership (article 12287421, 2023), which reports that Ian Harvey addressed the hospital board shortly before his departure. The author's note gives the source as the trust's board minutes of 4 October 2018. Searches of the Thirlwall Inquiry's evidence index, transcripts and report for these words returned nothing, so the minutes themselves have not been seen here: the quotation is used on the Daily Mail's authority and attributed to it in the text, and the date is not stated. If the minutes are obtained, the citation should be changed to them.

Annex on external review ("When the sheriff gets it wrong"): the Chester account is taken directly from the Thirlwall Inquiry report — Volume I, Chapter 12 (the child death panel and the April 2017 meeting with Cheshire Police) and Volume II, Chapters 19, 20, 23, 25 and 26 (the Speak Out Safely policy, the RCPCH review, the consultants' treatment, NHS England's involvement and the escalation to the police) — with all quotations as the report gives them, including "The RCPCH took on a task for which it was ill-equipped", "worse than inappropriate, it was egregious" and "The families were failed by the Countess and by the RCPCH". The review team is named as the report names it: Dr Milligan (lead), Dr Graham Stewart (second reviewer, who died during the Inquiry), Claire McLaughlan (lay reviewer) and Alexandra Mancini (nursing reviewer, nominated by the RCN), with Sue Eardley as Head of Invited Reviews and Dr David Shortland as clinical lead of the service. Lists of reviewers circulating in AI-generated summaries — naming Dr Susan Edees and a nurse called Tuohy, and placing Dr Shortland on the visit — are wrong and are not used. The wording of the consultants' letter of apology, reported elsewhere, could not be verified at source and is not quoted. The Andersen material comes from Arthur Andersen LLP v. United States (2005) and United States v. Arthur Young & Co. (1984), SEC press releases 2001-62 and 2002-89, the Senate Governmental Affairs Committee's Financial Oversight of Enron (October 2002) for the $52m/$25m fee split and the internal-audit point, the Senate Permanent Subcommittee's report on Enron's board (July 2002), the Sarbanes-Oxley Act 2002 and its SEC implementing release, and contemporaneous reporting for the "2X" strategy. The founder's "not enough money in the city of Chicago" reply is firm lore recorded in Inside Arthur Andersen (2003) rather than documented fact, and is presented as such; the versions that name a railway are not supported. The claim that Andersen "surrendered its licence" is avoided: the firm told the SEC it would cease practising before the Commission by 31 August 2002.

Fry and Nightingale. The social link — Lady Inglis, president of the ladies' committee of Fry's Institution and "a friend of the Nightingale family" — and the conclusion that "written proof of direct contact, although likely, is lacking" are both from Huntsman, Bruin and Holttum, "'Twixt candle and lamp", Medical History 46 (2002), pages 351–380. The girlhood meeting is from the Dictionary of National Biography 1912 supplement, which itself hedges it: Nightingale's horizon was widened "possibly by a chance meeting in girlhood with Mrs. Elizabeth Fry". The same entry lists the thirty-eight nurses who sailed for Scutari in October 1854 as ten Roman Catholic sisters, eight Anglican sisters of mercy, six from St John's Institute and fourteen from various hospitals — so the common claim that Fry's nurses sailed with Nightingale is not supported and is not made here. The Florence Nightingale Museum's register of nurses sent to the East does contain two women trained at the "Training Institution Devonshire Square" — the address of Fry's Institution — Sister Laura Crofts, sent to Renkioi in November 1855, and Mrs Christina Montague, at Kulali and then Scutari; both went out after the first party. The Fliedner link is from Sarah Tooley's biography of Nightingale, which records that his most important London friendship was with Elizabeth Fry; Tooley also asserts that Nightingale and Fry met, while admitting "we have no detailed record of what words were spoken". Three further claims in circulation could not be verified and are not used: that Lady Maria Forester funded three Fry nurses and recruited Nightingale to lead them, that Fry was a distant relative of Nightingale, and that Nightingale's aunt Patty accompanied Fry to Newgate.