Will screening individuals at high risk of cardiovascular events deliver large benefits? No
BMJ 2008; 337 doi: https://doi.org/10.1136/bmj.a1395 (Published 28 August 2008) Cite this as: BMJ 2008;337:a1395- Simon Capewell, professor of clinical epidemiology
- capewell{at}liverpool.ac.uk
Evidence supporting the high risk approach for preventing cardiovascular disease is disappointing. The strategy has low effectiveness and is associated with high cost, residual risk, medicalisation, and increasing inequalities. Whole population approaches are cheaper and more effective.
Low effectiveness
The large OXCHECK trial (of a nurse led health check plus health education and follow-up tailored to the level of cardiovascular risk) had only modest effects on cardiovascular events.1 Indeed, more recent literature raises further concerns.2 3 Healthcare professionals’ advice to stop smoking or take more exercise also has frustratingly modest effects.4
All screening programmes are imperfect. Even with generous resources, the call, recall, and follow-up systems require major commitments. Drop outs are substantial. Screening failures will be more common in the more deprived groups, who also experience higher rates of disease, thus increasing inequalities.5 6
All cardiovascular risk scoring systems are inaccurate and potentially confusing. Even the best risk charts have predictive accuracy of only 60%-70% for an individual patient. Furthermore, a bewildering variety of charts exist. Many general practitioners do not understand, interpret, or use the charts well.2 7
Treatment failure will also reduce systematic population benefit. Even in motivated patients, about 5% will be intolerant of statins and 10% intolerant of antihypertensive drugs.8 9 Furthermore, long term adherence with hypertension therapy or statins is consistently less than 50%.10
Residual risk
Drugs do not eliminate the underlying pathology. They merely put a sticking plaster over the problem. The patient remains physiologically hypertensive and vulnerable to hypercholesterolaemia. Thus, stopping treatment will rapidly lead to recurrence of the pathological state with raised cardiovascular risk.
Even with continuing therapy, the risk of subsequent cardiovascular events is reduced by only about 15% with successful control of blood pressure11 and by about 25% with cholesterol reduction.12 The efficacy of statins in primary prevention may be even lower, particularly in women.13 There is also uncertainty about how these benefits might add up. Even optimistically assuming a hypothetical combined risk reduction of 40% (that is, 15%+25%), over half the cardiovascular risk will remain. This conflicts with many patients’ expectations. It is also brave to assume that efficacy in randomised control trials translates into equal effectiveness in the messy realities of clinical practice.14
Cost
The high risk approach would commit the majority of middle aged adults to lifelong drug treatment with huge costs. Over a quarter of UK adults have cholesterol concentrations above 5.2 mmol/l and over half have blood pressure above 140/80 mm Hg. Indeed, over 80% of British men aged 65-74 would be categorised as high risk, needing treatment.15 This partly reflects a progressive reduction in treatment thresholds. Thus, more individuals become eligible for treatment, but the number needed to treat inflates while side effect rates persist and costs spiral.
There are also non-financial costs. The high risk approach medicalises healthy people, turning them into lifelong patients. The implicit message for patients is that “the doctor can fix it.” This takes responsibility away from the individual and may encourage further risk taking behaviour. For instance, continuing consumption of junk foods high in salt and saturated fats.
Furthermore, quality of life often decreases after starting treatment for hypertension or hyperlipidaemia.16 17 A high risk label also has serious implications for personal health insurance. Thus, given the choice, most people would rather opt for behavioural change than lifelong medication.17
The focus on individuals also favours affluent and educated people, thus increasing social inequalities. Disadvantage can occur at every stage in the process, from the person’s health beliefs and health behaviour, through presentation, negotiation, participation, and adherence with treatment.18 19
Cost effectiveness
The high risk approach distracts attention from cheap policy interventions that reduce risk factors across entire populations. Small reductions in population cholesterol concentrations, blood pressure, or smoking can translate into substantial reductions in cardiovascular events and deaths.20 21 Some two decades ago, Geoffrey Rose suggested that: “A large number of people exposed to a small risk may generate many more cases than a small number exposed to a high risk.”22 Empirical evidence to support the Rose hypothesis has progressively emerged from Swedish patients with hypertension, British middle aged men, Dutch and Irish adults, and a pan-European dataset.21 23 24 25 Although Manuel and colleagues recently questioned the Rose hypothesis,26 their brave assumptions were widely criticised.
The medical treatment of high risk people typically costs £5000-£100 000 per quality adjusted life year (QALY).8 27 Population interventions typically cost much less, and can even be cost saving.28 Thus, national policy changes and legislation can be both effective and relatively cheap, whether banning trans-fatty acids (Denmark), halving dietary salt (Finland), or promoting smoke-free public spaces (UK, Ireland, and Italy).29 30
In conclusion, perhaps the greatest harm arising from high risk strategies is misleading professionals, planners, and politicians into thinking they can tick the “mission accomplished” box for preventing cardiovascular disease. In fact, screening for high risk individuals represents a costly and relatively ineffective strategy that distracts from cheaper and more effective policy interventions which benefit entire populations.
Notes
Cite this as: BMJ 2008;337:a1395
Footnotes
Competing interests: None declared.