Studying new antidepressants: if there were a light at the end of the tunnel, could we see it?

This commentary began with the proposition that the SSRIs have become the standard of comparison for new antidepressants. It is also suggested that the conventional wisdom that all antidepressants are equally effective is no longer true. Rather, it is asserted that the same factors that have compromised the sensitivity of RCTs to detect drug-placebo differences similarly have impaired discriminations between effective and even more effective antidepressants. In order to have the power to make such a distinction, an RCT may need to enroll 300 or more patients per cell. Few studies thus have adequate statistical power. Alternatively, conclusions can be drawn from quantitative methods that combine data from groups of smaller studies. The relative merits and limitations of 2 strategies used to examine the results of comparative studies, meta-analysis and pooled analysis, were discussed. The former method is preferred when there are a large number of relevant RCTs; however, failure to include unpublished data from all relevant studies may inflate results. The latter method, unless biased by selective inclusion of studies or marked heterogeneity of results, is preferred when there are only a handful of comparable studies. Although the task of selecting among a number of good choices remains more clinical art than science, quantitative methods are now available to help determine if there are clinically meaningful differences in antidepressant efficacy. The conventional RCT provides a low power telescope for visualizing differences between effective medications. If there were a light at the end of the tunnel, it would be difficult to see it. Until methods that can improve the ability to discriminate between an active antidepressant and a placebo are identified and implemented, the ability to see modest differences necessitates the use of statistical methods such as meta-analysis and pooled analysis of original data.