A dangerous bias.

A dangerous bias Sir, In their recently published paper (Domar et al., 2013), the authors' conclusion that 'there is little evidence of benefit from the antide-pressants prescribed for the majority of women of childbearing age—and there is ample evidence of risk', is based on a biased reading of the literature. Rather than including a balanced analysis of the studies on the subject, they cite five articles supporting their conclusions, out of thousands. For instance, a recent meta-analysis examining six placebo-controlled studies found a significant treatment effect for patients with non-severe depression (Stewart et al., 2012). Another recent re-analysis of the Fournier article they cite concluded that Fournier's results were based on unreliable data (Isacsson and Adler, 2012). Moreover, a number of the articles cited in the paper support the conclusion that antidepressants are an effective treatment for depression (e.g. DeRubeis et al., 1999; Far-amarzi et al., 2008a,b). Nor is this all. The authors suggest that a lack of data on antide-pressants in pregnant women amounts to proof of their inefficacy. This is wrong on two counts. First, even if there were no data, it would be far more reasonable to assume that, until proven otherwise , drugs shown to work in the general population would also work in specific populations. Secondly, if there is a relative dearth of data, it is only because conducting randomized control trials of medication use in pregnancy is unethical. However, data does exist, in the form of two studies looking at rates of relapse in women who continue or discontinue their antidepressants during pregnancy (Cohen et al., 2006; Yonkers et al., 2011). While Cohen et al. found a significant increase in relapse amongst women who discontinued their medication during pregnancy , Yonkers did not find a difference amongst the two groups. In glossing over the subtleties that could account for these different results, Domar et al. miss Yonker's conclusion that these two studies together likely suggest that relapse is higher for some women, but not all. Amazingly, Domar et al. also claim that 'there is overwhelming evidence that cognitive behavioral therapy is equivalent to antidepressant medication', citing the Canadian Network for Mood and Anxiety Treatments (CANMAT) clinical guidelines (Parikh et al., 2009). In fact, CANMAT concludes that 'combined [my emphasis] antidepres-sant and CBT or IPT are recommended as first-line treatments for acute MDD', because evidence suggests that combined treatment is superior to either medication or …