More attention on the quality of the economic analyses is needed
Marshall(1) analysed a number of interventions to prevent coronary
heart disease and presented analyses of their incremental cost-
effectiveness in terms of incremental drug treatment cost per incremental
first major coronary event prevented. We welcome the author’s reiteration
of the principles of incremental cost-effectiveness analysis and agree
that such an approach should form the basis of guideline development in
this area. However, the analysis as presented has a number of major
limitations, and does not even comply with the BMJ’s guidelines for
• The impact of the interventions on subsequent major coronary events
and on other major vascular events within the timeframe of the analysis is
ignored. For example, the Heart Protection Study reports on the benefit
of statin intervention on first and subsequent major vascular events (3).
• The savings to the health service due to the avoidance of coronary (and
other vascular events) are not included in the analysis.
• The evidence for the effectiveness of interventions is largely based on
intention-to-treat analysis of naturalistic RCTs (or their meta-analysis),
while the costs of the drug interventions are estimated according to a
treatment protocol; in consequence, incremental effectiveness and
incremental cost are not calculated on the same basis. For example, the
economic analysis of the LIPID study showed that patients in the placebo
arm incurred significantly more non-study cholesterol-lowering drug costs
• The author has ignored the impact of uncertainty in the resource use and
cost data on the estimated cost-effectiveness. Furthermore, the cost-
effectiveness of statin treatment is very sensitive to the cost of statin,
which is rapidly changing in Britain at present due to the recent
expiration of the patent for simvastatin. In Sweden, where simvastatin
has recently come off patent, the price has decreased by 85%.
Although it might be argued that these limitations affect all
interventions such that any rank ordering of interventions would be
unaffected, this would be to miss the aim of the cost-effectiveness
analysis entirely. The point is that, as the accompanying editorial
suggested, treatment thresholds are points on a continuum. In order to
provide reliable bases for issuing guidance, what is required are credible
analyses that capture all important concerns to decision makers.
Unfortunately, the present ‘back of the envelope’ model does not provide a
credible basis for choosing the most appropriate treatment thresholds.
(1) Marshall T. Coronary heart disease prevention: insights from
modelling incremental cost effectiveness. BMJ 2003; 327(7426):1264-0.
(2) Drummond MF, Jefferson TO. Guidelines for authors and peer
reviewers of economic submissions to the BMJ. Br Med J 1996; 313:275-283.
(3) Heart Protection Study Collaborative Group. MRC/BHF Heart
Protection Study of cholesterol-lowering with simvastatin in 5963 people
with diabetes: a randomised placebo-controlled trial. The Lancet 2003;
(4) Glasziou PP, Eckermann SD, Mulray SE, Simes RJ, Martin AJ, Kirby
AC et al. Cholesterol-lowering therapy with pravastatin in patients with
average cholesterol levels and established ischaemic heart diesease: is it
cost-effective? MJA 2002; 177:420-426.
We are currently involved in the economic analysis of the Heart Protection Study.
Competing interests: No competing interests