Mediating and moderating effect of postpartum depressive symptoms on maternal psychological resilience and parent-child interaction in early infancy

  Objective  To analyze the association of maternal psychological resilience during pregnancy with parent-child interaction in early infancy and the mediating and moderating effect of postpartum depressive symptoms on the association.

  Methods  The study was conducted during April 2016 – June 2020 among 3 429 mother-infant dyads recruited from the participants of Shanghai Maternal-Child Pairs Cohort. Baseline survey on the mothers was carried out at 12 – 14 weeks′ gestation with a self-administered questionnaire. Maternal psychological resilience during pregnancy was evaluated using the Resilience Scale for Adult (RSA). During follow-ups two months after delivery, postpartum depressive symptoms of the mothers was assessed with the Edinburgh Postnatal Depression Scale (EPDS) and parent-child interaction was estimated by main caregivers of the infants using a self-designed questionnaire under the guidance of community physicians. Unconditional binary logistic regression model was adopted to analyze the effect of maternal higher resilience on insufficient parent-child interaction in early infancy; causal mediation and stratified analysis were used to analyze the mediating and moderating effects of maternal postpartum depressive symptoms on the effect of maternal psychological resilience.

  Results  Of the mothers, 1 738 (50.69%) and 1 691 (49.31%) were assessed as having high and low psychological resilience. Compared to the mothers with low resilience, the mothers with high resilience reported significantly higher proportion of with high socioeconomic status (56.24% vs. 36.94%) but lower proportions of aged < 25 years at the delivery (9.23% vs. 16.46%), tobacco exposure in first trimester (14.84% vs. 19.33%) and postpartum depressive symptoms (17.86% vs. 22.90%) (all P < 0.001). Unconditional binary logistic regression analysis showed that the mothers with high resilience were at a lower risk of insufficient parent-child interaction in early infancy than those with low resilience (adjusted odds ratio aOR = 0.75, 95% confidence interval 95% CI: 0.63 – 0.89) after adjusting main confounders. The results of the causal mediation analysis revealed that high maternal resilience could play a protective role on insufficient parent-child interaction by reducing postpartum depressive symptoms, with a percentage mediated (PM) of 7.19% (OR_{nature indirect effect} = 0.97, 95% CI: 0.95 – 0.99). Compared to the mothers with high resilience and without postpartum depressive symptoms, the mothers with only low resilience or postpartum depressive symptoms were at increased risk of insufficient parent-child interaction, with the aORs (95% CI) of 1.29 (1.06 – 1.56) or 1.45 (1.08 – 1.97); the increased risk was much higher (aOR = 2.20, 95% CI: 1.71 – 2.85) for the mothers with both low resilience and postpartum depressive symptoms.

  Conclusion  High maternal resilience during pregnancy has a protective effect on insufficient parent-child interaction in early infancy, and postpartum depressive symptoms plays both mediating and moderating effect on the effect of maternal resilience.