The effectiveness of postpartum interventions aimed at improving women’s mental health after medical complications of pregnancy: a systematic review and meta-analysis

Abstract Background Postpartum mental disorders including depression and anxiety are common. Medical complications of pregnancy, such as preeclampsia and gestational diabetes, are thought to increase the risk of mental disorders postpartum. However, it is unclear which interventions may be effective for preventing and/or treating postpartum mental disorders following a medically complicated pregnancy. We aimed to systematically review published literature on the effectiveness of postpartum interventions to improve women’s mental health after medical complications of pregnancy. Methods Systematic review (PROSPERO: CRD42021220030) was performed. Eligibility criteria: (1) randomized controlled trials (RCTs), published 1st Jan 2001-12th August 2021 (2) outcome measures reported on postpartum mental disorders (3) participants had ≥ 1 medical complication during pregnancy (4) intervention entirely postpartum or contained a postpartum component (5) full-text available in English or Chinese. Risk of bias was assessed using the Revised Cochrane Criteria Risk of Bias. Random effects inverse-variance weighted meta-analysis was used to pool the individual standardized mean differences (SMD) in depression or anxiety scores between intervention and control groups. Results Of 5928 studies screened, 9 met inclusion criteria, and were based on non-pharmaceutical, combined lifestyle interventions that began shortly after childbirth, or as part of extended care packages beginning during pregnancy. Of these, 2 were rated as low risk of bias, 1 with some concerns, and 6 were at high risk. Meta-analysis was performed for 8 studies using standardized measures of depression and 7 for anxiety. There were statistically significant reductions in depression (SMD − 1.48; 95%CI: -2.41 to -0.55), and anxiety scores (SMD − 1.98; 95%CI: -3.03 to -0.94) in intervention versus control groups. Considerable heterogeneity was noted for pooled depression (I2 = 97.9%, p