Estimating depression prevalence from the Beck Depression Inventory: is season of administration a moderator?

The existence of winter seasonal affective disorder (SAD) and its milder population variants implies that depression estimates in a given population may tend to be higher in winter than at other times of the year. The aim of this study was therefore to test whether depression prevalence estimates based on the Beck Depression Inventory (BDI) are systematically moderated by season of administration. Existing information from the screening phase of a multicentre investigation of depression prevalence provided the data for the study. Repeated cross-sectional BDI data from samples in the United Kingdom (n=1299), Finland (n=1352), Norway (n=2711) and Spain (n=1246) were analysed for month- and season-of-administration effects. Whether data were measured continuously or as a dichotomous variable (BDI cutoff >/=13), there was no evidence of a systematic seasonal pattern in depression estimates across the four sites. No seasonal effects reached statistical significance at any single site, and trends in the association between winter and elevated BDI scores were positive in two sites (UK and Norway) and negative in two (Finland and Spain). Although limited by a post hoc analysis of existing data, this study provides the strongest evidence to date that season of administration is not a moderator of depression prevalence as estimated by the BDI.