We welcome Riordan & Stark's suggestion that declining family size may have contributed to declining suicide rates. I agree. As we noted, there are many potential confounding variables in the relationship between antidepressants and suicide.

Riordan & Stark cite two birth cohort studies using data linkage that both found that birth order was independently associated with suicide risk. In Riordan et al's 2006 analysis of linked data from the Scottish Morbidity Record and Scottish death records, higher maternal parity, younger maternal age, non-professional parental occupations and low birth weight were all independently associated with higher suicide risk of offspring in young adulthood. ^{Reference Riordan, Selvaraj, Stark and Gilbert1} In Gravseth et al's study using several Norwegian national registers, suicide risk factors included higher maternal parity, instability of maternal marital status during childhood, low education levels and severe mental illness. ^{Reference Gravseth, Mehlum, Bjerkedal and Kristensen2} Notably, maternal parity remained a significant risk factor even after adjustment for mental illness.

Data linkage studies such as Riordan et al's and Gravseth et al's are particularly important in suicide research because of the need for adequately large sample sizes and sufficient statistical power to investigate suicide, a relatively rare event. They are methodologically superior to ecological studies, which are the mainstay of claims that antidepressants reduce suicide. The inherently weak methodology of ecological studies is often compounded by failure to control for potential confounding factors and by biased interpretation. ^{Reference Safer and Zito3,Reference De Leo and Cerin4}

Data linkage studies generally reveal multiple significant contributory factors, many of which are linked to social adversity. Many such factors also contribute to other forms of premature mortality such as accidental death and natural death due to preventable conditions. For example, Riordan et al also found an association between higher maternal parity and increased risk of offspring death from causes other than suicide. The aetiological overlap means that primary prevention focusing on shared determinants has greater potential to reduce overall mortality. ^{Reference Neeleman5}

Data linkage studies provide valuable evidence that challenges the simplistic belief that depression is the cause of suicide and antidepressants are the solution. As emphasised by De Leo & Cerin, ^{Reference De Leo and Cerin4} suicide is not simply a function of depression, and suicide prevention is far more than a psychiatric enterprise.