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.
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.
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.
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.
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.
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.
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:
- Prescribed appropriately within 14 hours of admission, or exception documented.
- Doses administered as prescribed, with omissions tracked.
- 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.
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.
