Treatment of subjects lost to follow-up in the analysis of mortality studies.

In the analysis of retrospective cohort mortality studies, persons lost to follow-up are either withdrawn at the time of loss or are assumed to be alive at the end of the study. It has been advocated that the former method does not give a biased estimate of expected mortality in this group, and is therefore the preferred method. In this paper, it is argued that this approach is not always the best one, and that in certain situations, depending on the method of follow-up used, assuming that persons lost to follow-up are alive may give more accurate results than withdrawing them at the time of loss. Data from a recent study are used for illustration.