The use of statins: a case of misleading priorities?BMJ 1997; 315 doi: https://doi.org/10.1136/bmj.315.7112.826 (Published 04 October 1997) Cite this as: BMJ 1997;315:826
National guidance that does not link costs and benefits is worthless
- Nick Freemantle (), Senior research fellowa,
- Ruth Barbour, Medical adviserb,
- Richard Johnson, Pharmaceutical adviserb,
- Mike Marchment, Chief executiveb,
- Andrew Kennedy, General practitionerc
- a Centre for Health Economics, University of York, York YO1 5DD
- b Warwickshire Health Authority, Warwick CV34 4DE
- c Croft Medical Centre, Royal Leamington Spa, Warwickshire CV31 1SA
Last month the NHS Executive distributed to health authorities and general practitioners a statement from the Standing Medical Advisory Committee on the use of lipid lowering drugs.1 This argues for a strategy of treatment on the basis of underlying risk, but it adopts a level that is probably unachievable, fails to present the evidence, and ignores cost effectiveness. In so doing it jeopardises the objective of targeting high risk patients, but it also raises questions about the worth of guidance that does not link benefits and costs.
The statement recommends that, having considered other methods of reducing the risk of coronary heart disease, clinicians should give statins to the following three groups of patients. The first priority are patients who have had a myocardial infarction and have low density lipoprotein values of 3.2 mmol/l or more; second are those with angina or other clinically overt atherosclerotic disease with low density lipoprotein values of 3.7 mmol/l or more; and third come those with a high risk of developing coronary heart disease according to the revised Sheffield tables2 and a low density lipoprotein value of 3.7 mmol/l or more. The statement estimates that meeting all three priorities will mean treating 8.2% of the population aged 35-69. No recommendations are made for patients aged over 70, although there is no evidence that benefits are limited to younger patients. The statement was sent to all general practitioners and health authorities and trusts with a commendation from the chief medical officer3 or the NHS Executive.4
In Warwickshire Health Authority full implementation of these recommendations will involve treating about 17 000 patients, some 10 000 for secondary prevention and 7000 for primary prevention. The cost of the statins (using the cheaper evaluated agent) with doses reflecting those in major trials will be about £8m for the authority, representing 20% of the drugs budget, at an average cost of about £100 000 per practice. These costs do not include diagnostic tests and the time required to identify, counsel, and treat patients. The statement gives no indication of the likely benefits if the recommendations are followed, although three large randomised trials,5 6 7 and an epidemiological study2 8 are cited. No additional resources have been made available, and the statement gives no indication of where savings should be made to pay for the proposed changes in practice.
How should health authorities and clinicians respond to the statement? The first issue is to identify the number of patients likely to benefit. The numbers randomised in the major trials and reductions in deaths or myocardial infarctions are described in the 1. In the secondary prevention trials, which included mainly patients with previous myocardial infarctions and a few with angina, the cholesterol concentration required for inclusion differs. A total plasma cholesterol value of less than 6.2 mmol/l and low density lipoprotein of 3-4.5 mmol/l was required in the CARE trial6 and a total serum cholesterol value of 5.5-8.0 mmol/l in the 4S trial (the Scandinavian Simvastatin Survival Study)5 (average low density lipoprotein at baseline was 4.87 mmol/l). In both trials patients were treated for about five years. The 4S trial included higher risk patients, which probably explains the larger estimated benefit, but was also stopped early on the basis of a data driven criterion and thus may overestimate the effects of treatment. The Sheffield tables are derived from epidemiological data from the Framingham population2 8 and may not adequately take into account benefits from alternative recently introduced treatments.
The absolute risk reduction in the 4S trial was 3.3% over the five years of treatment (see 1). In other words, about 30 patients will require treatment for five years to avoid a death (or 150 for one year). If, in line with the medical advisory committee statement, treatment is broadened to patients in the lower risk band evaluated by the CARE study,6 128 patients will require treatment for five years to avoid a death (or 640 for one year). The authors of the revised Sheffield tables argue that targeting treatment at patients at a 3% annual risk of a major coronary event,8 twice that in the WOSCOPS study,7 will lead to twice the benefits. This implies that about 55 patients will require treatment for five years to avoid a death (or 275 for one year). Even using “time to event data”9 in the original trial reports, only the 4S trial achieved narrow confidence intervals in the estimated reduction in overall mortality.5 The benefits in primary prevention depend on strong assumptions and are somewhat speculative.
Achieving desired change in the NHS is a difficult process that requires planning resources. The Faculty of Public Health Medicine, which also issued advice to directors of public health in August,10 recommends an approach based on locally developed guidelines that target high risk patients and consider issues of cost effectiveness. However, the faculty does not provide examples of the likely benefits for a range of risks for secondary and primary prevention patients. A first step in planning local implementation may be to ensure that patients meeting the inclusion criteria in the 4S trial are targeted for treatment as the benefits for these patients are substantial.
Neither the faculty advice nor the Standing Medical Advisory Committee statement refer to cost effectiveness, although the statement notes that such studies are being commissioned. If NHS resources are to be used efficiently policies have to be based on information on costs as well as benefits. Failure to do this can lead to the sort of inefficient practices apparently advocated by the Standing Medical Advisory Committee and supported by the Department of Health. There are also competing demands such as ensuring that patients with heart failure get angiotensin converting enzyme inhibitors.11 12
Cynics may see in the release of the statement and instruction from the NHS Executive in August an element of “passing the buck.” Enthusiasts for cholesterol lowering drugs may adopt the priorities but unless they curtail prescribing in other areas they will soon run into financial difficulties. Given the absence of information on cost effectiveness, they may do more harm than good through making savings in more cost effective areas to pay for statins. Most practitioners will probably be overwhelmed by the size of the proposed changes and do little, which may mean that high risk patients likely to benefit substantially from treatment will remain untreated.
National evidence based guidelines in which recommendations are linked explicitly with evidence on benefits and costs and which provide decision makers with the information they need to determine local priorities are necessary to support strategic change. The four page statement from the Standing Medical Advisory Committee, which describes neither the likely benefits and associated costs nor a realistic range of options, is simply inadequate.