Ten RCTs (n=1,233) were included, of which 6 were follow-up studies (n=231).
Drop-out rates were significantly greater among patients receiving pharmacotherapy compared with those receiving psychotherapy: 28.4% versus 23.6% (RR: 1.29, 95% CI: 1.07, 1.57, p=0.009). No statistically significant heterogeneity was detected (p=0.73).
There was no statistically significant difference in remission rates at treatment termination between psychotherapy and pharmacotherapy: 37.9% versus 34.8% (RR: 0.91, 95% CI: 0.79, 1.06, p=0.24). No statistically significant heterogeneity was detected (p=0.23). The NNT with psychotherapy to produce one remission from depression that would not have occurred had patients been treated with antidepressants was 32.
There was no statistically significant difference in remission rates at treatment termination between psychotherapy and pharmacotherapy for patients with either chronic depression (36.1% versus 36.6%, p=0.83; based on 3 studies) or non-chronic depression (41.1% versus 32.2%, p=0.12; based on 8 studies). No statistically significant heterogeneity was detected for either chronic (p=0.58) or non-chronic (p=0.14) subgroups. There was no significant difference in remission rates for chronic (p=0.31) and non-chronic (p=0.25) depression for either treatment.
There was no statistically significant difference in remission rates at treatment termination between psychotherapy and pharmacotherapy for patients with either mild depression (46.5% versus 44.4%, p=0.34; based on 5 studies) or moderate depression (33.2% versus 31.9%, p=0.44; based on 5 studies). No statistically significant heterogeneity was detected for the analysis of moderate depression (p=0.55) but statistically significant heterogeneity was detected for the analysis of mild depression (p=0.07; I-squared 54%). Remission rates were significantly greater for mild compared with moderate depression for both psychotherapy (46.5% versus 33.2%, p=0.001) and pharmacotherapy (44.4% versus 31.9%, p=0.003).
The risk of relapse at follow-up was significantly greater among patients receiving pharmacotherapy compared with those receiving psychotherapy: 56.5% versus 26.5% (RR: 0.46, 95% CI: 0.33, 0.65, p<0.0001). No statistically significant heterogeneity was detected (p=0.68).