A systematic review including 7 studies with a total of 494 subjects was abstracted in DARE. Women and ethnic minorities were under-represented in the included studies, so the conclusions cannot be generalised to all depressed individuals who are HIV-positive. When using the random-effects model, the pooled effect size from all of the included RCTs was 0.57 (95% CI: 0.28 to 0.85). Analysis of potential moderators using meta-regression suggested that the only significant moderator was placebo response, which explained 61.6% of the variance (p=0.004). When the studies were stratified according to placebo response, the pooled effect size estimate was 0.80 (95% CI: 0.52, 1.08) for placebo response less than 33% and 0.20 (95% CI: -0.11, 0.52) for placebo response greater than 33%.
Comment: The quality of evidence is downgraded by heterogeneity.