In their multiple-treatment meta-analysis, Andrea Cipriani and colleagues1 conclude that clinically important differences favour escitalopram and sertraline over other second-generation antidepressants. We published a similar study,2 with similar but not identical methods, and drew a different conclusion: benefits do not differ materially across the drugs. We believe that Cipriani and colleagues have overstated the evidence and ignored imprecision inherent in this analysis, thereby conveying overly precise findings from an approach fraught with uncertainties.
