Multiple mediation analysis of the peer-delivered Thinking Healthy Programme for perinatal depression: findings from two parallel, randomised controlled trials.

Abstract

Low-intensity psychosocial interventions have been effective in targeting perinatal depression, but relevant mechanisms of change remain unknown. To examine three theoretically informed mediators of the Thinking Healthy Programme Peer-delivered (THPP), an evidence-based psychosocial intervention for perinatal depression, on symptom severity in two parallel, randomised controlled trials in Goa, India and Rawalpindi, Pakistan. Participants included pregnant women aged ≥18 years with moderate to severe depression, as defined by a Patient Health Questionnaire 9 (PHQ-9) score ≥10, and were randomised to either THPP or enhanced usual care. We examine whether three prespecified variables (patient activation, social support and mother-child attachment) at 3 months post-childbirth mediated the effects of THPP interventions of perinatal depressive symptom severity (PHQ-9) at the primary end-point of 6 months post-childbirth. We first examined individual mediation within each trial (n = 280 in India and n = 570 in Pakistan), followed by a pooled analysis across both trials (N = 850). In both site-specific and pooled analyses, patient activation and support at 3 months independently mediated the intervention effects on depressive symptom severity at 6 months, accounting for 23.6 and 18.2% of the total effect of THPP, respectively. The intervention had no effect on mother-child attachment scores, thus there was no evidence that this factor mediated the intervention effect. The effects of the psychosocial intervention on depression outcomes in mothers were mediated by the same two factors in both contexts, suggesting that such interventions seeking to alleviate perinatal depression should target both social support and patient activation levels. None.