We read the debate on antidepressants and suicide ^{Reference Isacsson, Rich, Jureidini and Raven1} with interest, especially the issue concerning the importance of controlling for possible confounding variables in ecological studies and the associated difficulties in drawing conclusions from such ecological observations.

One potentially important confounder, which has been overlooked, is the size of the family of origin. Birth cohort studies from Scotland ^{Reference Riordan, Selvaraj, Stark and Gilbert2} and Norway ^{Reference Gravseth, Mehlum, Bjerkedal and Kristensen3} suggest that having elder siblings may be linked with an increased risk of suicide. As the total fertility rates (a reasonable proxy for average family size) fell across most Western countries between the 1950s and 1970s, ⁴ it is plausible that some of the decline in suicide rates observed from the late 1980s onwards may be, in part at least, a resultant cohort effect.