Study did not take account of confounders

EDITOR,—We welcome the approach taken by P D P Pharoah and W Hollingworth in using life tables to evaluate the cost effectiveness of reducing cholesterol concentration.1 This approach has been widely used elsewhere2 and is a valuable technique for evaluating health service interventions aimed at modifying the risk profiles of a population. Unfortunately, however, their analysis may be flawed as such analyses depend on obtaining an accurate assessment of the overall manner in which population death rates are influenced by the intervention …