LAUSR

The study by Lacourt reported in a recent issue of Occup Environ Med 1 does not satisfy two fundamental requirements of the design and conduct of a case–control study: these limitations invalidate its results and those of other reports based on the same dataset.2 First, all cases and controls should be sampled from the same population–time experience (‘study-base’).3 An alternative, commonly used way to express this concept is that controls are subjects (or a random sample of theirs) who would have been selected as cases, had they developed the disease of interest. Cases in the study by Lacourt et al were men who were diagnosed with mesothelioma ‘either in five French regions between January 1987 and December 1993 or in 22 French districts between January 1998 and December 2006’.1 The first series of cases was collected in a community-based case–control study,4 and the second series was identified within a mesothelioma surveillance programme. …