Modelling economic benefits after such long term treatment is inappropriate

EDITOR—The West of Scotland Coronary Prevention Study Group used an epidemiological model to estimate the cost effectiveness ratio for cholesterol lowering with pravastatin.1 Although the title suggests otherwise, this is not a direct health economic analysis of original trial data, and why such an analysis has not been published is unclear. Original trial data would provide direct estimates of the number of life years of survival in treated and untreated groups, over the five years of the study. Similarly, a direct estimate of the drug costs would have been available; these would have been offset by an estimate of savings from the reduction in non-fatal adverse events.

A close approximation to this can be carried out by using published data. The all cause death rate in the treated group over five years in the trial was 3.2%, compared with 4.1% in the untreated group.2 With a starting population of 10 000, this corresponds to a total number of life years remaining of 49 200 compared with 48 975—a saving of 225 life years. The net cost of treating this population for five years is estimated at £22 811 769,1 or £101 386 per life year saved during treatment. This is similar to our estimate of the average cost effectiveness in a lower risk population of £136 000.3

The cost effectiveness ratio can be dramatically reduced by modelling survival after treatment has stopped. The treated group continues to benefit in terms of life years gained, …