VTE assessment form was never the solution - prescribe and administer

 

Quality improvement · NHS · Venous thromboembolism

England measures VTE form completion. It does not measure whether patients received thromboprophylaxis.

Every year, millions of adults are admitted to NHS acute hospitals. Venous thromboembolism — deep vein thrombosis and pulmonary embolism — remains one of the commonest causes of preventable harm in that setting. The clinical answer has been clear for decades: identify who is at risk, and give pharmacological thromboprophylaxis when the bleeding risk allows. Low molecular weight heparin and related agents work. The point of admission is to stop clots, not to complete paperwork.

Yet what England industrialised after 2010 was not prophylaxis. It was the VTE risk assessment form.

What the form was supposed to do

The national VTE prevention programme required documented risk assessment on admission, aligned with NICE guidance. Trusts adopted the NHS VTE risk assessment tool. Boxes were ticked for reduced mobility, acute illness, age, obesity, active cancer, surgery, and other factors. Bleeding risk was weighed. A score or summary judgment followed: offer mechanical measures, offer pharmacological prophylaxis, or neither.

On paper, this is rational pathway design. In practice, the form became the product. CQUIN payments, standard-contract clauses, and board dashboards converged on one question: was the assessment done? National returns report that roughly nine in ten adult admissions now receive a documented VTE risk assessment. The operational standard is 95%. Compliance charts look like success.

But compliance with assessment is not compliance with prophylaxis. The programme's stated aim was DVT and PE prevention. The mainstay of that prevention, for most adult inpatients in acute hospitals, is chemical thromboprophylaxis. The form is one step on the way to a prescription. It is not the prescription. When the form becomes the national scoreboard, the pathway inverts: documentation is rewarded; delivery is assumed.

Structure without certainty

The VTE assessment form looks objective. It is structured, standardised, and mandated. That appearance is misleading.

Risk stratification through the form is still a clinical judgment wearing a template. Two doctors can admit the same frail, immobile, acutely unwell patient and complete the same form differently. One records high VTE risk and prescribes prophylaxis. Another records lower risk and does not. Inter-rater variation in VTE risk assessment is well described. Different validated risk assessment models applied to the same patients can recommend prophylaxis in as few as one in nine or as many as nine in ten. The Department of Health tool used across the NHS sits within that spectrum of classification behaviour.

The form also encodes a default of withholding treatment until risk is explicitly declared high enough. That is the opposite of how most acute admissions present. A patient occupying an inpatient bed in an acute trust is, by definition, there because ward-level care is required. They are immobile relative to normal life. They have an acute illness, a surgical insult, or an exacerbation of disease. NICE already lists acute illness and reduced mobility as indications to consider pharmacological prophylaxis. If the patient is in hospital, those criteria are usually already met before the first box is ticked.

The form therefore risks redundant work at best and active obstruction at worst. At best, it records what should already be obvious. At worst, it creates a false-negative pathway: a subjective "low risk" classification interrupts the route to prophylaxis, and the national system celebrates the completed assessment anyway.

Voluntary national audit data bear this out. The Thrombosis UK and GIRFT survey of 2019–20 found that among high-risk patients judged eligible for drug prophylaxis, appropriate prescribing aligned with NICE in about 88% of cases — but with enormous inter-hospital variation from 40% to 100%. Missed doses among patients who were prescribed prophylaxis occurred in roughly 8% of cases. A substantial share of hospital-associated VTE episodes were judged potentially preventable because of failures at multiple points in the pathway: missing or incorrect assessment, failure to reassess when the clinical picture changed, delayed or omitted prescription, wrong dose, or doses not given.

None of that is consistent with a reliable, objective filter standing between patients and effective prevention.

The missing national metric

Here is the striking gap. England knows, to the percentage point, how many admitted patients had a VTE risk assessment. It does not publish an equivalent national rate for pharmacological thromboprophylaxis prescribing across the whole admitted population.

That is not a technical limitation. Most acute trusts now run electronic prescribing and medicines administration systems. If a patient received enoxaparin, dalteparin, or fondaparinux for prophylaxis, that order is in the record. A national or regional audit could report:

• the proportion of adult inpatients prescribed pharmacological VTE prophylaxis within 14 hours of admission;
• the proportion in whom it was contraindicated or already anticoagulated;
• the proportion of prescribed doses actually administered.

These are straightforward queries on structured data. They would answer the question patients and clinicians actually care about: was prevention given? Instead, the system measures the proxy — the form — and treats it as if it were the outcome.

The absence of this metric is a policy choice, not a data problem.
When a national programme selects one easy-to-collect process measure and ignores the treatment that measure is meant to trigger, boards learn what gets measured. Trusts staff VTE coordinators to chase form completion. Pharmacy and ward teams — who actually prescribe and give heparin — are rarely held to the same national account. We have built a reporting architecture that cannot tell us whether the right patients received the right drug at the right time.

NICE quality standards call for prophylaxis within 14 hours when indicated. Local audit is encouraged. National industrialisation stopped at the assessment.

Inverting the logic

A different approach is to stop asking a variable questionnaire to rediscover what admission already implies.

At Scunthorpe General Hospital, general surgery adopted a departmental protocol: pharmacological thromboprophylaxis for all inpatients unless actively bleeding. That is not a rejection of NICE. It is NICE's acute-illness and immobility logic applied directly, with a narrow exception list instead of a subjective risk score. A quality improvement project led by foundation doctors then checked contemporaneously whether prophylaxis had been prescribed, gave feedback, and prescribed when admitting teams had not. The department already had a 90% rate of appropriate prescribing under the protocol. Reminding staff of guidelines alone changed nothing — the process was already embedded. Active ward-level checking and prescribing raised appropriate pharmacological prophylaxis to 95%, a statistically significant improvement.

The lesson is not that one small DGH is an outlier. The lesson is what had to be done to achieve high prescribing in a setting where assessment compliance was never the barrier. The barrier was execution at the bedside. A simplified rule — inpatient equals indication unless clearly excepted — plus frontline ownership produced rates at or above those seen in national audits of selected high-risk cases, on a broader denominator that included all general surgical inpatients, not a pre-filtered audit sample.

The principles are reproducible. Acute hospital inpatients outside mental health and purely community settings share a common feature: they are sick enough, and immobile enough, to justify considering drug prophylaxis unless an objective exception applies — active bleeding, existing therapeutic anticoagulation, a defined contraindication, palliative intent, or patient refusal. That exception list is shorter, clearer, and less variable than a risk assessment form interpreted differently by every admitting team.

General surgery was the proof of concept because the risk profile is concentrated. The same presumption extends logically across acute medical wards, where immobility and acute illness are the norm rather than the exception. What changes is not the principle but the exception list: medical inpatients carry more bleeding-risk complexity, and a mature protocol must capture that without reverting to a subjective score that lets patients fall through. The form should document why prophylaxis was not given, not gate whether it should be considered.

What should replace the form as the national object

The VTE assessment form may still have a place as a record of exceptions and contraindications, or as a prompt within electronic prescribing. It should not remain the primary national quality measure for VTE prevention.

If the intention is DVT and PE prevention, and chemical thromboprophylaxis is the mainstay, then the national object should be prescribing and administration, measured from electronic prescribing data:

1. Prescribed appropriately within 14 hours of admission, or exception documented.
2. Doses administered as prescribed, with omissions tracked.
3. Reassessment when clinical status changes — particularly before discharge in high-risk groups.

Risk assessment completion can remain in the background. It should not continue to stand in for a prevention pathway that many patients never fully receive because a subjective tick-box said they were low risk.

We spent fifteen years optimising a form.
It is time to measure what we meant to treat.

The epidemiology of hospital-associated thrombosis has not responded as one would expect if that form reliably delivered prophylaxis. The problem is not that prevention fails. It is that we built a structured, subjective gate where a simple presumption and a short exception list would better serve acutely ill, immobile inpatients — and we chose not to count the prescription.

Data sources: NHS England VTE Risk Assessment Collection (2025/26); Thrombosis UK / GIRFT National Thrombosis Survey (2019–20); NICE NG89 and QS201; Scunthorpe General Hospital ASGBI 2025 quality improvement abstract (SP5.12). 

 ©M HEMADRI 

The Price of Understaffing: What UK Healthcare Outcomes Really Tell Us

From workforce gaps and wage suppression to avoidable deaths, delayed cancer diagnoses, record waiting lists and a £60 billion negligence liability — a data-driven examination of consequence.

📊 Operational data | OECD, NHS Resolution, Nuffield Trust    🌍 UK vs Scandinavia vs Western world    📅 2024–2025 figures

This is the third post in a three-part series. If you haven’t read the first two, the argument builds on their foundations:

Part 1 → UK healthcare staff: fewer in number and lower paid than comparable countries

Part 2 → Where does the NHS spend the £60bn it saves on staffing? Non-staff costs examined

In the first two posts in this series we established two uncomfortable truths. First, the UK has significantly fewer doctors and nurses per capita than comparable high-income nations, and pays them less. Second, the money not spent on staff doesn’t disappear — it ends up absorbed by non-staff costs: pharmaceuticals, management consultancies, PFI financing, administration, and IT procurement at a premium. The staffing gap between the NHS and comparable Scandinavian systems is up to £60 billion per year.

In this third post we ask the most important question of all: what does that gap cost patients?

We examine five major outcome domains — avoidable mortality, clinical outcomes for cancer, heart attack and stroke, elective surgery waiting times, clinical negligence costs, and temporary staffing expenditure — and look at what the operational data (not trial data) tells us about statistical correlation with staffing levels.

---

1. Avoidable Mortality: Deaths the System Should Have Prevented

Avoidable mortality is split into two components. Preventable mortality reflects failures of public health upstream. Treatable mortality — deaths that should not have occurred with timely, effective clinical intervention — is the sharpest and most relevant mirror for health system performance, because it isolates what healthcare itself can and should prevent.

The Nuffield Trust’s analysis of OECD data shows that the UK’s treatable mortality rate was 71 per 100,000 population in 2019 — above that of seven Western European comparator countries for which data was available that year. In this measure, a higher number means more deaths — more people whose lives the healthcare system should have saved but did not.

Country	Treatable deaths / 100,000	Year	vs UK
Switzerland	~39	2021	Better (−45%)
Australia	49	2022	Better (−31%)
Nordic countries (Sweden, Norway, Denmark)	Below UK	2021–22	Consistently better — lowest-mortality quartile across OECD
4 further W. European nations	Below UK	2019	All outperform UK (Nuffield Trust / OECD)
UK	71	2019	Above 7 of its Western European peers
OECD average	79	2021	UK below average only because E. European countries raise it
United States	95	2022	Worst among comparable high-income nations
Sources: Nuffield Trust/OECD (2024); OECD Health Statistics. Age-standardised, deaths under 75. Higher = worse.

The UK at 71 sits below the OECD average of 79 only because that average is pulled upward by Eastern and Southern European countries with significantly weaker healthcare systems. Against its genuine peer group — France, Germany, the Netherlands, Belgium, Switzerland, Australia, and the Nordic nations — the UK performs poorly. The difference is in healthcare capacity and staffing levels.

Statistical Significance Note

A peer-reviewed cross-national panel study using OECD data across 26 countries found that a 1% increase in nurse-staffing density reduces 30-day mortality from heart attack by 0.65%, from haemorrhagic stroke by 0.60%, and from ischaemic stroke by 0.80%. Sweden and Denmark had the highest simulated reductions in overall HCQI mortality from their nursing levels (−3.53 and −3.31 respectively).

Source: Labbé et al. (2018) — 26 OECD countries, 2005–2015.

---

2. Cancer Care: Survival Rates That Lag Behind Our Neighbours

The most recent EUROCARE-6 data analysed across 29 European countries reveals a consistent pattern: Nordic countries dominate the top of survival tables across most major cancers, while the UK is near or below average for its income group.

Cancer Type	Sweden	Norway	Denmark	UK	EU-24 Avg.
Ovarian (5-yr)	46.5%	45%+	–	36.2%	39.2%
Lung (5-yr)	19.5%	19.0%	–	13.3%	~15–16%
Melanoma (5-yr)	87%+	87%+	87%+	~83%	83%
Pancreatic (5-yr)	~10–12%	~10–12%	~10–12%	6.8%	~9%
Sources: EUROCARE-6 (De Angelis et al., 2024); EU Country Cancer Profiles Synthesis Report 2025 (OECD/EU). Five-year relative survival rates.

For ovarian cancer, Sweden records 46.5% five-year survival compared to the UK’s 36.2% — a gap of over ten percentage points that directly translates to lives lost. For lung cancer the UK at 13.3% is significantly below Sweden (19.5%) and Norway (19.0%).

“Survival was persistently higher in Australia, Canada, and Sweden, intermediate in Norway, and lower in Denmark, England, Northern Ireland and Wales, particularly in the first year after diagnosis and for patients aged 65 and older.” — International Cancer Benchmarking Partnership, Lancet, 2011

---

3. Heart Attack and Stroke: Where Every Minute — and Every Nurse — Counts

A landmark study using nationwide registry data — 87 Swedish hospitals (119,786 patients) and 242 UK hospitals (391,077 patients), 2004–2010 — found that 30-day mortality from AMI was lower in Swedish hospitals (8.4%) than UK hospitals (9.7%).

Sweden — AMI 8.4% 30-day case-mix adjusted mortality

UK — AMI 9.7% 30-day case-mix adjusted mortality — and higher variation between hospitals

That 1.3 percentage point difference translates to thousands of preventable deaths annually across 100,000+ AMI admissions per year.

Statistically Significant Correlation: Staffing → Acute Mortality

At the ward level in the English NHS, a retrospective longitudinal study of 66,923 admissions found a statistically significant association between registered nurse fill-rate and in-hospital mortality (OR 0.9883, 95% CI 0.9773–0.9996, p=0.0416). An extra 12-hour shift by an RN was associated with a 9.6% reduction in the odds of a patient death.

Critically, there was no statistically significant association for healthcare support workers or agency nurses — meaning agency staff are not effective substitutes for permanent, ward-familiar RNs.

Source: Propper et al., BMJ Quality & Safety 2023; Dall’Ora et al., JAMA Network Open 2024.

---

4. Elective Surgery Waiting Times: A Crisis Within a Crisis

As of late 2025, 7.3 million elective procedures were on the NHS waiting list in England. Only 62% of patients were waiting less than 18 weeks — far below the 92% constitutional standard. For orthopaedic procedures:

NHS — Hip Replacement 24–28 wks Average wait 2025 (pre-pandemic: 12–13 weeks)	NHS — Knee Replacement 28–29 wks Average wait 2025 (pre-pandemic: ~13 weeks)
Spain, Finland, Italy ~Pre-covid Hip/knee wait recovery broadly on track by 2023	UK vs Peers 50% longer England’s median hip wait still 50% longer than pre-2020

“England has fewer hospital beds, lower numbers of key staff and lower levels of investment in buildings and equipment than many other high-income countries — and this is likely to have affected how quickly the millions of people waiting can have the surgery they need to live comfortably.” — Nuffield Trust, 2024 analysis of OECD Health Statistics

Countries with more staff, more beds, and better-paid permanent workforces recovered more quickly because they had more capacity to absorb the backlog. England entered the pandemic with structural vulnerabilities — the same ones documented in Parts 1 and 2 of this series — and those vulnerabilities have defined the pace of recovery.

---

5. Clinical Negligence: The Hidden Fiscal Iceberg

Metric	Figure	Trend
Annual claims paid (2024/25)	£3.1 billion	+10% year-on-year
Annual claims paid (2023/24)	£2.8 billion	+6.8% on prior year
Annual claims paid (2006/07, real terms)	£1.1 billion	Baseline — 182% real-terms increase since
“Cost of harm” estimate (CNST, 2024/25)	£4.6 billion	Wider measure
Total provision for future liabilities (March 2025)	~£60.0 billion	2nd largest government balance sheet liability
Maternity-related liabilities (since 2019)	£27.4 billion	52% of annual pay-outs relate to obstetrics
New claims filed (2024/25)	14,428	+5% — exceeds pre-pandemic peak

⚠️ KEY FINDING: The £60 billion total negligence liability is not primarily a legal or administrative failure. It is a patient safety failure. Patient safety failures are systematically linked in the operational literature to inadequate staffing, high use of temporary staff, and overworked permanent staff — precisely the conditions documented in Part 1 of this series.

---

6. Agency and Temporary Staff Spend: The Vicious Cycle

The NHS’s reliance on agency staffing is the most direct and visible financial consequence of the workforce shortages described in Part 1. The cycle is self-reinforcing and extremely expensive.

1 Workforce undersupply + below-market pay113,000 NHS vacancies at peak; nurse pay 20–30% below comparable economies

↓

2 Staff leave permanent NHS roles or reduce hoursBurnout, industrial action 2022–24, and active resignation-to-agency arbitrage

↓

3 Trusts pay premium agency rates to fill rotasUp to £2,000 per nursing shift; total spend peaked at £3.5bn (2022/23)

↓

4 Quality of care deteriorates — agency ≠ permanent RNAgency nurses do not reduce mortality risk equivalently to permanent RNs (Propper et al. 2023)

↓

5 Clinical incidents, complaints and negligence claims rise14,428 new clinical negligence claims in 2024/25 — 5% above pre-pandemic peak

↓

6 Budget consumed; less available for permanent staffing investment£3.1bn negligence pay-outs + £2.07bn agency spend = over £5bn diverted from patient care annually

---

7. The Statistical Case: Connecting Staffing Inputs to Outcome Outputs

The six outcome domains above are connected through a common mechanism: the ratio of appropriately trained, well-supported, permanent clinical staff to patients in need. The operational evidence base for this relationship is extensive and consistent in direction.

Key operational evidence points (all statistically significant):

OECD: Nurse staffing → AMI/Stroke mortality	1% increase in nurse density → 0.65% reduction in AMI 30-day mortality; 0.80% reduction in ischaemic stroke mortality. Analysis of 26 OECD countries 2005–2015.
NHS England: RN fill-rate → inpatient mortality	Extra 12-hour RN shift: 9.6% reduction in odds of patient death (OR 0.9044; p=0.0416). No equivalent effect for healthcare support workers or agency nurses. Study of 66,923 admissions, 53 wards, 2017.
Multidisciplinary staffing → hospital mortality	Hospitals with lower medical and AHP staff had 4% higher mortality rates (RR 1.04; 95% CI 1.02–1.06). Pooled finding from a systematic review (Dall’Ora et al., 2023).
Temporary staffing → mortality risk not fully mitigated	626,313 patient admissions (JAMA Network Open, 2024): days of low nurse staffing, even when remedied by temporary staff, carried elevated mortality risk compared to adequate permanent staffing.
Nuffield Trust: Structural vulnerabilities → slow elective recovery	Across 10 high-income nations, England’s post-pandemic recovery for hip/knee replacements was slower than Spain, Finland, Italy, Portugal, Sweden, and Norway. “Fewer beds, lower numbers of key staff” explicitly named as causal factors.
NHS RN seniority → mortality reduction (dose-response)	A senior RN (Band 7–8) had 2.2 times the mortality-reducing impact of a Band 5 RN. Pay suppression drives experienced staff out; junior replacements are not equivalent.

---

8. Operational Productivity: Theatre Utilisation, Cases Per List, and the Staffing Paradox

Theatre Utilisation: 38% of Lists Underused Before the Pandemic

An NHS Improvement audit in 2019 found that 38% of theatre lists were underutilised, with unused theatre time estimated to cost the NHS approximately £400 million annually. NHS England’s 2024/25 operational planning guidance set a target of making “significant improvement towards” 85% theatre utilisation — not that the target was being met.

🏥 A note on the metric: The NHS’s “capped theatre utilisation” (CTU) measure in the Model Hospital database is calculated in a mathematically invalid way (Pandit et al., British Journal of Anaesthesia, 2023). The underutilisation problem is real, but the 85% target should be treated as directionally correct rather than a precise comparable benchmark.

Cases Per List: Team Stability Is the Statistically Proven Driver

A study of 255,757 procedures across 38 UK hospitals found that switching between different procedure types on a list increased operative duration by an average of 6.48%. A systematic review of 76 studies concluded that employing specialised and stable teams in dedicated operating rooms showed significant improvements in outcomes; disturbances and communication failures negatively affected operative time and surgical safety.

A case-control study of cataract surgery found an odds ratio of 1.7 (95% CI 1.0–3.1) for complications on lists affected by unplanned staff absence — the direct consequence of thin staffing pools with no experienced cover.

The Waiting List Paradox: More Staff, Barely Any Reduction

NHS Staff Growth 2019–2024 +20% FTE workforce increase	Waiting List Change +73% Mar 2020 → Mar 2024 (4.4m → 7.5m)
Elective Recovery Target 20.9% Activity increase vs 2019/20 set for 2024/25 — significantly missed	Actual YoY Gain 2024/25 2.7% Year-on-year acute productivity growth (NHS England)

Important context: The 2.7% year-on-year rise in 2024/25 is one year’s growth. It does not conflict with NHS England’s position that level productivity remained approximately 8% below 2019/20 — a critical stock-versus-flow distinction often muddled in political debate.

Trust leaders surveyed by NHS Providers (May 2024) identified the top barriers to productivity as: delayed discharges (48%), lack of revenue funding (38%), and patient acuity (37%). The two initiatives most cited as improving productivity: reducing agency spend (55%) and workforce retention initiatives (53%). Not technology. Not management restructuring. Both top levers were about stabilising the permanent workforce.

Productivity Driver	Mechanism	Staffing Link
Theatre underutilisation (38% of lists, £400m/yr)	Beds unavailable; staff absent; lists cancelled	Direct
Switching procedures on list (+6.48% time/case)	Fragmented scheduling; no dedicated specialist lists	Direct
Unplanned staff absence → complications (OR 1.7)	Thin staffing pools; no resilience	Direct
+20% staff, waiting list barely reduced	Younger/less experienced workforce; churn; burnout	Direct — consequence of pay suppression
Delayed discharge blocking beds → cancelled lists	Social care cannot absorb medically fit patients	Indirect — social care staffing crisis
Staff burnout and sickness absence (+18% vs pre-pandemic)	Reduced capacity, unplanned leave, reduced effort	Direct — pay dissatisfaction and overwork

---

9. The Causal Chain: How Pay, Staffing, and Staff Effectiveness Connect to Every Outcome

The eight preceding sections share a single causal root. The chain runs in four documented steps: inadequate pay → attrition and vacancies → depleted, less experienced, less effective workforce → worse clinical outcomes across every domain measured.

Link 1: Pay → Dissatisfaction → Intent to Leave

A 2025 cohort study in The Lancet Regional Health — Europe (UK-REACH cohort) found that pay dissatisfaction is strongly associated with attrition intentions across all staff groups. NHS Employers’ analysis shows a Band 7 ward manager suffered a 13.6% real-terms pay decline between 2013 and 2023. Junior doctor pay satisfaction collapsed from 46% in 2020 to 13.6% in 2023.

NHS Pay Review Body’s Own Conclusion (2023)

“An investment in NHS pay, by reducing attrition and staff shortages and supporting service reform, should lead to improved public health outcomes, labour market participation, and higher national income.”

NHS Pay Review Body, 36th Report 2023. HMSO Cm 866. — the statutory advisory body to government on NHS remuneration.

Link 2: Dissatisfaction and Attrition → Vacancies → Agency Dependency

In 2022, a record 170,000 NHS workers left hospital and community health services, including 41,000 nurses. By September 2023 there were 121,070 recorded vacancies including 42,300 nursing vacancies. These vacancies are filled by agency staff at a peak annual cost of £3.5 billion, or by internationally recruited nurses with no institutional knowledge of NHS systems.

Staff leavers (2022) 170,000 Record high; including 41,000 nurses	NHS Vacancies (Sept 2023) 121,070 Including 42,300 nursing and 8,850 doctor vacancies
Leavers for health reasons ×4 Nearly quadrupled in a decade	Agency spend peak £3.5bn 2022/23 — financial cost of vacancy-driven attrition

Link 3: Depleted Experienced Workforce → Reduced Clinical Effectiveness

A statistically significant mortality benefit was found only for permanently employed registered nurses — not for healthcare support workers, and not for agency nurses. A senior Band 7–8 RN had 2.2 times the mortality-reducing impact of a Band 5 entry-level RN. Every experienced nurse driven out by inadequate pay and replaced by an agency worker represents a net reduction in the safety of every patient on that ward.

NHS staff sickness absence is 18% higher than pre-pandemic levels, with over a quarter of days lost attributable to anxiety, stress, and mental health — the direct product of overwork, understaffing, and persistent pay suppression.

Link 4: Reduced Effectiveness → Every Outcome Measured

OECD cross-national (26 countries)	1% increase in nurse staffing density → 0.65% reduction in AMI mortality; 0.80% reduction in ischaemic stroke mortality. Sweden and Denmark show the highest system-level benefits.
Cancer outcomes (EUROCARE-6)	Sweden records 46.5% five-year ovarian cancer survival. UK: 36.2%. Lung cancer: Sweden 19.5%, UK 13.3%. Persistent, 30-year gap consistent with structural staffing differences.
AMI 30-day mortality (Sweden vs UK)	Sweden 8.4% vs UK 9.7% — registry data from 391,077 UK and 119,786 Swedish patients, 2004–2010.

The Chain, Summarised

A Pay suppressed in real terms 2010–2023Band 7: −13.6% real terms. Junior doctor pay satisfaction: 46% → 13.6%. 31 studies confirm poor pay → poor retention.

↓

B Record attrition and vacancy accumulation170,000 leavers in 2022; 121,070 vacancies (September 2023); leavers for health reasons ×4 in a decade.

↓

C Experienced workforce depleted; agency and international fillAgency spend peaks at £3.5bn. 42% of new nurses non-UK national (2023 vs 21% in 2018). Band 7–8 RN has 2.2× impact of Band 5. Theatre teams destabilised.

↓

D Sickness absence, burnout, and reduced discretionary effortNHS sickness absence 18% above pre-pandemic. Anxiety/stress = 25% of absence. 60%+ clinicians worn out daily. Only 34% believe teams adequately staffed.

↓

E Clinical effectiveness and throughput reduced+20% NHS headcount; output per head falls. 38% of theatre lists underutilised. Productivity 8% below pre-pandemic despite larger workforce.

↓

F Worse outcomes across every domainTreatable mortality above Western peers. Cancer survival below Nordic comparators. AMI mortality higher than Sweden. Waiting list ~7.4m. Negligence £3.1bn.

This chain is not a hypothesis. Each link is supported by operational data. The NHS Pay Review Body — the statutory body advising government on NHS remuneration — has itself concluded that investment in pay would produce improved patient outcomes. The only question remaining is whether policymakers choose to act on the evidence they already have.

---

10. The False Economy: What the NHS Actually Spends on the Consequences

Cost Category	Annual Figure	Causal Link to Staffing Deficit
Clinical negligence pay-outs (2024/25)	£3.1bn	Strong — patient safety failures linked to staffing levels
Agency staff spend (2024/25)	£2.07bn	Direct — agencies fill gaps from pay-driven vacancies
Bank staffing (NHS flexible workers)	~£2–3bn est.	Direct — same structural cause as agency spend
Projected negligence pay-outs (by 2029/30)	>£4bn p.a.	Trajectory continues without structural reform
Total negligence liability (balance sheet, March 2025)	~£60bn	Accumulated years of system failure — largely avoidable harm

The annual consequence cost — agency spend plus negligence pay-outs — is already running at over £5 billion per year. The entire NHS nursing pay bill could be increased by 10% for approximately £2–3 billion per year — less than the agency spend and negligence claims combined. The NHS is spending more on the consequences of understaffing than it would cost to significantly improve permanent staffing levels.

Conclusion: A False Economy With a Human Cost

This three-part series has told a coherent and uncomfortable story. Part 1 showed the UK trains, employs and pays fewer healthcare staff than comparable nations. Part 2 showed the money not spent on staff has been absorbed by non-staff costs, with a combined premium over Scandinavian systems of up to £60 billion per year.

Part 3 shows what happens downstream. Treatable mortality above Western peers. Cancer survival below Nordic countries. Heart attack mortality higher than Sweden. Elective waiting lists of 7.3–7.4 million. Clinical negligence costs of £3.1 billion per year against a £60 billion total liability. Agency spend peaked at £3.5 billion. And an operational productivity crisis: 38% of theatre lists underutilised, £400 million in wasted theatre time annually, a waiting list that barely moved despite a 20% workforce increase.

The statistical correlations between staffing and outcomes are not speculative. A 1% increase in nurse staffing density reduces acute cardiac and stroke mortality by 0.65–0.80%. An extra RN shift reduces ward-level mortality odds by 9.6%. Senior experienced RNs have more than twice the mortality-reducing impact of junior nurses. Section 9 traces the complete causal chain: pay suppression → attrition → vacancies → agency reliance → depleted experienced workforce → worse clinical outcomes. The NHS Pay Review Body itself has concluded that investing in pay would improve patient outcomes.

The question is no longer whether we can afford to invest properly in NHS staffing. The operational data asks a more pointed question: can we afford not to?

Key Sources & Verified References

1. OECD / European Commission (2024). Health at a Glance: Europe 2024. doi.org/10.1787/b3704e14-en
2. OECD (2025). Health at a Glance 2025. doi.org/10.1787/8f9e3f98-en
3. Nuffield Trust (2024). Still waiting: Is it just England that still has a backlog problem? nuffieldtrust.org.uk
4. Nuffield Trust (2024). Mortality rates. nuffieldtrust.org.uk/resource/mortality-rates
5. Propper, C. et al. (2023). BMJ Quality & Safety. PMC10176371
6. Dall’Ora, C. et al. (2024). JAMA Network Open. PMC11333978
7. Dall’Ora, C. et al. (2023). Human Resources for Health. PMC10116759
8. Labbé, V. et al. (2018). ScienceDirect
9. NHS Resolution (2025). resolution.nhs.uk
10. National Audit Office (2025). nao.org.uk
11. House of Commons Public Accounts Committee (2025). publications.parliament.uk
12. House of Commons Library (2025). commonslibrary.parliament.uk
13. EUROCARE-6 / OECD (2024–25). OECD.org (PDF)
14. Coleman, M. et al. (2011). PMID 21183212
15. Jernberg, T. et al. (2015). PMC4528190
16. OECD (2024). oecd.org
17. King’s Fund (2025). kingsfund.org.uk
18. Pandit, J.J. et al. (2023). PMC10308435
19. Institute for Fiscal Studies (2025). ifs.org.uk
20. NHS England (2025). england.nhs.uk
21. ONS (2025). ons.gov.uk
22. Agyemang, C. et al. / UK-REACH (2025). PMC12541634
23. NHS Employers (2023). nhsemployers.org
24. Khamisa, N. et al. (2020). PMC7375434
25. NHS Pay Review Body (2023). assets.publishing.service.gov.uk (PDF)

Multisource Feedback and Systematic Bias

 

Multisource Feedback is currently an established method to obtain information on individuals at workplace. It is an essential component to be able to continue careers for doctors in the UK as it is a requirement for revalidation to remain on the medical register.

On colleague MSF a GMC report states:

‘There was evidence of systematic bias amongst some groups of colleague respondents in respect of assessments provided’

‘Certain doctor characteristics appeared to predict systematic variation in colleague assessments.’

‘In interpreting data arising from such surveys of colleagues, consideration should be given to the possibility of systematic bias in a colleague’s report based on non-clinical aspects of care, as well as the extent of colleagues’ familiarity with the doctor.’

 

On Patient MSF a GMC report states: 

'There was evidence of systematic bias in reporting amongst some groups of patient
respondents. '

'In interpreting data arising from such surveys of patients, consideration should
be given to the possibility of systematic bias in a patient’s report based on nonclinical aspects of care, as well as the socio-demographic profile of the patient
sample'

Software systems that are used for MSF have begun to use benchmarking – it is not clear if these benchmarks are statistically adjusted or standardized to consider and recalibrate for these biases. It is also not known whether the software systems identify these biases and inform the doctor that the responses could be impeded by biases.

It is not clear if doctors undergoing revalidation for GMC licence to practice are explicitly informed of these biases and offered guidance and support on how to choose MSF respondents to negate the existing biases or counselling on how to recognise, challenge and deal with any biases that becomes apparent.

Appraiser training does have an EDI component, but it is unsure if this training includes methods on how to consider these biases including any undue positive biases?

There are publications that suggest that the MSF could be one of the triggers that may be used when considering performance. This makes any issues regarding bias paramount to the topic.

It is very important to prevent issues – unfortunately it may be the case that the MSF methodology as currently implemented for doctors (and perhaps other health professionals) has a built-in flaw in the structure when it does not consider and adjust for biases.

Doctors should be aware of this and be prepared to address this issue should it affect them. Doctors’ organisations should resolve this at the earliest opportunity.

 

 

 

Ref:

  https://pubmed.ncbi.nlm.nih.gov/23095930/ 

 https://www.gmc-uk.org/-/media/gmc-site/about/rtguidanceresearchandpilotingofgmcquestionnairesdc8593pdf64915124.pdf

UK postgraduate medical examination pass rates – what are your chances if you an IMG/BME?

The GMC annual survey report for 2018 is now published (https://www.gmc-uk.org/about/what-we-do-and-why/data-and-research/national-training-surveys-reports#). While the GMC’s focus seems to be on the training environment, it is important to delve into the online reporting tool (https://webcache.gmc-uk.org/analyticsrep/saw.dll?Dashboard ) where the detail resides.

I have a long interest in poor outcomes for discriminated groups and looked at the pass rates for postgraduate examinations. The results are still not that great or even for IMGs and BMEs. Overall, if you are a UK graduate with a white ethnicity – then the pass rate is 75.1% and if you are an IMG-BME the pass rate is 42.2% and if you are an EEA-BME then it falls to 37.2%. Basically if you are white UK graduate you have a 100% higher pass rate than some BME doctors. We can argue about the unfairness of this and have done so in the past.

However, let us focus on UK graduates alone for now. Overall there a white UK graduate has a 12 percentage point higher pass rate than a BME UK graduate. That is pretty gross. When you point this out, the classic false arguments based on sub-classification starts (schools in UK, candidates from abroad joining UK medical schools, etc) which essentially devalues the many years of UK medical school teaching and training.

Then if we start looking at how individual specialties are faring in their faculty and royal college pass rates, we find that most have an approximately 12 point difference biased towards white UK graduate doctors and biased against BME UK graduate doctors.

There are one postgraduate examination where UK white doctors and UK BME doctors have more or less even pass rates (with only a 3 percentage point lower for UK BME doctors) – that would be the Faculty of Public Health examination.

Then…

There is one postgraduate examination where UK BME graduates fare better than the UK white graduates by 10 percentage points. That would be the Faculty of Occupational Health Examination.

The biggest difference between UK white doctors and UK BME doctors is the RCGP examination where there is a 14.5 percentage point difference in favour of UK white doctors.

It is a sad fact of UK healthcare that such differences exist.

Overall if you are an IMG-BME you have a 57.8% failure rate in UK postgraduate examinations according to the GMC’s latest report.

Due to the medical staff vacancy crisis in the UK, the powers that be have increased the ‘MTI’ duration from two to three years with a view to attracting more IMGs to UK. The IMGs would hope to obtain some UK qualifications – though the reasons they are recruited is because they are already qualified specialists in the first place.

Make wise decisions.

Would you put your money in a scheme where you have a 57.8% chance of losing it? That is the downside. On the upside, you do have a 45.2% chance of not losing your money and in the bargain getting your qualification – what happens then? You need a more detailed analysis of what your chances of getting into a training post, chances of returning to the country of origin without a CCT/CESR, getting back to your country of origin with a CESR, etc.

As a doctor with a scientific qualification data based decision making would be recommended. Think deeply.

Also think deeply of the impact of the process, – knowing that you have a high chance of failing examinations, knowing about the realities of career progression – even if you were eventually successful in your examination or career progression.

Think deeply, very deeply. Make wise decisions. It is not easy but it does affect your life.

©M HEMADRI

Follow me on Twitter @HemadriTweets 

If you are a doctor wanting interview skills training for job interviews you may want to check out my website:

https://successatmedicalinterviews.co.uk/

We also offer highly bespoke  Leadership and Management training and unique QI skills