Meta-analysis of antidepressant prescribing

Drop out rates were presented in a misleading manner

EDITOR,--The meta-analysis by I M Anderson and B M Tomenson is an interesting development in the literature addressing the effectiveness and cost effectiveness of prescribing of antidepressants.1 The authors estimate that the risk of patients stopping treatment is 10% lower with selective serotonin reuptake inhibitors than tricyclic antidepressants. They note, however, that this corresponds to an absolute reduction in the total drop out rate of only 2.5% (their table IV) and is of borderline significance. Also, as they suggest, biases such as “patient selection, previous treatment, methodology, and study setting” may exist and make …