Methods We systematically reviewed the pertinent trial literature until May 2014. The random-effects model was used to compute the pooled relative risk of relapse or recurrence (RR). A distinction was made between two comparator conditions: (1) treatment-as-usual and (2) the use of antidepressants. Other sources of heterogeneity in the data were explored using meta-regression.

Background Major depression is probably best seen as a chronically recurrent disorder, with patients experiencing another depressive episode after remission. Therefore, attention to reduce the risk of relapse or recurrence after remission is warranted. The aim of this review is to meta-analytically examine the effectiveness of psychological interventions to reduce relapse or recurrence rates of depressive disorder.

Results Twenty-five randomised trials met inclusion criteria. Preventive psychological interventions were significantly better than treatment-as-usual in reducing the risk of relapse or recurrence (RR=0.64, 95% CI=0.53-0.76, z=4.89, p<0.001, NNT=5) and also more successful than antidepressants (RR=0.83, 95% CI=0.70-0.97, z=2.40, p=0.017, NNT=13). Meta-regression showed homogeneity in effect size across a range of study, population and intervention characteristics, but the preventive effect of psychological intervention was usually better when the prevention was preceded by treatment in the acute phase (b=-1.94, SEb=0.68, z=-2.84, p=0.005).

Limitations Differences between the primary studies in methodological design, composition of the patient groups and type of intervention may have caused heterogeneity in the data, but could not be evaluated in a meta-regression owing to poor reporting.

Conclusions We conclude that there is supporting evidence that preventive psychological interventions reduce the risk of relapse or recurrence in major depression.