THE RELATION BETWEEN TREATMENT BENEFIT AND UNDERLYING RISK IN METAANALYSIS

In meta-analyses of clinical trials comparing a treated group with a c ontrol group it has been common to ask whether the treatment benefit v aries according to the underlying risk of the patients in the differen t trials, with the hope of defining which patients would benefit most and which least from medical interventions. The usual analysis used to investigate this issue, however, which uses the observed proportions of events in the control groups of the trials as a measure of the unde rlying risk, is flawed and produces seriously misleading results. This arises through a bias due to regression to the mean and will be parti cularly acute in meta-analyses which include some small trials or in w hich the variability in the true underlying risks across trials is sma ll. Approaches which previously have been thought to be more appropria te are to substitute the average proportion of events in the control a nd treated groups as the measure of underlying risk or to plot the pro portion of events in the treated group against that in the control gro up (L'Abbe plot). However, these are still subject to bias in most cir cumstances. Because of the potentially seriously flawed conclusions th at can result from such analyses, they should be replaced either by st atistically appropriate (but more complex) approaches or, preferably, by analyses which investigate the dependence of the treatment effect o n measured baseline characteristics of the patients in each trial.