Intended for healthcare professionals

Rapid response to:

Primary Care

Coronary heart disease prevention: insights from modelling incremental cost effectiveness

BMJ 2003; 327 doi: (Published 27 November 2003) Cite this as: BMJ 2003;327:1264

Rapid Response:

Author's response

Correspondence following publication of my article raises a number of
interesting issues. Nevertheless no correspondent disputes either the
approach (incremental cost-effectiveness analysis) or the fundamental
findings: that aspirin and antihypertensive treatment with bendrofluazide
and atenolol are markedly more cost-effective than statins and

Conradi and O’Donnell indicate that the side-effect profile of
statins may be less benign than is generally believed to be the case. If
so, this strengthens the paper’s conclusions.

Mann indicates that I may have been generous in accepting that
clopidogrel is effective in primary prevention. Again this strengthens the
paper’s conclusions. I agree with Mann’s suggestion that individuals need
information on the risk reduction with treatment in order to decide if it
is worthwhile.

Cooke suggests that development of further drugs to prevent coronary
heart disease may not be cost-effective. It is certainly clear that it is
very efficient to prevent coronary disease is using existing, low-cost

If, as Jacobs suggests, treatment effects are not be independent,
second and third treatments may be even less cost-effective than the
analysis suggests. Under present criteria, 80% of those eligible for one
preventive treatment are eligible for at least two and 53% are eligible
for all three.

Das and Vos suggest with oily fish may be more cost-effective than
statins. I agree. Cost per event prevented with sardines three times a
week is similar to that of initial antihypertensive treatment. However
fish oil capsules are much more costly than oily fish and there are wide
confidence intervals around the estimated effectiveness. For this reason
fish oil as an intervention was excluded from the published paper.

Kendrick argues that cost-effectiveness analysis should consider only
overall mortality. This may be problematic when clinical trials are not
powered to detect mortality changes.

D’Avanzo and Mihaylova et al suggest that the analysis should include
cost savings from heart disease prevented. This will not affect the cost-
effectiveness rankings but may indicate that the most efficient treatments
are cost-saving. Mihaylova et al also suggest further refinements to the
model, but acknowledge that these are unlikely to affect the rankings or
overall message. In my view the purpose of economic evaluation is to
improve decision-making. Complex economic evaluation is costly and may not
be understood by clinicians or guideline authors. My analysis is not
intended to assist a decision on allocation of resources between
preventive cardiology and other interventions, only to improve the
efficiency of current preventive cardiology services. It is sufficient for
this purpose. It is an open question as to whether a more complex and
costly evaluation might be justified by even greater service improvements.

Mihaylova et al, Ashcroft and Fells all refer to an immanent fall in
the price of simvastatin. At a pack (28) of £4.45, simvastatin is the
third most cost-effective intervention. Under these circumstances a
rational policy would offer aspirin at 5% coronary risk, bendrofluazide
and atenolol at 7.5%, and simvastatin at 15%. A third antihypertensive
would never be offered, obviating the need to “treat to target”. (Table 1)
The onus is now on pharmaceutical manufacturers to provide simvastatin at
this cost.

A final point not raised by any correspondent concerns the cost of
antihypertensive treatment. With some antihypertensives costing over £200
a year, the cost-effectiveness of antihypertensive treatment is very
sensitive to the treatment regime chosen. Antihypertensive treatment in
the current primary care system is much less cost effective than it could
be because drug costs currently average £150 to £200 per patient year of
treatment. This is hardly surprising given the huge range of services that
primary care teams are expected to provide. Perhaps it is time to look at
new models for provision of prevention services making greater use of
treatment protocols and supplementary prescribers.

Table 1: Cost effectiveness of preventive treatments in patients at
different risk levels if simvastatin’s price falls by 85%

CHD risk level-----------------5%-------7.5%-------10%------15%------


Bendrofluazide & Atenolol-----£24,000---£16,000--£12,000----

Simvastatin (£4.45 for 28)----£34,900---£23,300--£17,500---£11,600---



Competing interests:
None declared

Competing interests: No competing interests

12 December 2003
Tom Marshall
Harkness Fellow in Health Policy
Brigham & Women's Hospital, Boston MA 02120