Perceived partner support in pregnancy predicts lower maternal and infant distress

Abstract

Maternal postpartum emotional distress is quite common and can pose significant risk to mothers and infants. The current study investigated mothers' relationships with their partners during pregnancy and tested the hypotheses that perception of prenatal partner support is a significant predictor of changes in maternal emotional distress from midpregnancy to postpartum, and contributes to maternal ratings of infant distress to novelty. Using a prospective longitudinal design, 272 adult pregnant women were interviewed regarding their partner support, relationship satisfaction, and interpersonal security (attachment style and willingness to seek out support), and they completed standardized measures of prenatal symptoms of depression and anxiety (distress). At 6 to 8 weeks' postpartum, mothers reported these symptoms again and completed measures of their infants' temperament. Structural equation modeling (SEM) was used to test direct and indirect contributions of partner support, relationship satisfaction, and interpersonal security to maternal and infant postpartum distress. Mothers who perceived stronger social support from their partners midpregnancy had lower emotional distress postpartum after controlling for their distress in early pregnancy, and their infants were reported to be less distressed in response to novelty. Partner support mediated the effects of mothers' interpersonal security and relationship satisfaction on maternal and infant outcomes. A high-quality, supportive partner relationship during pregnancy may contribute to improved maternal and infant well-being postpartum, indicating a potential role for partner relationships in mental health interventions, with possible benefits for infants as well.

Figure 1

Hypothesized path model.

Figure 2

Final model predicting postpartum maternal emotional distress and infant distress to novelty from prenatal partner relationship variables. Circles denote latent variables; rectangles denote measured variables. One-headed arrows represent regression paths; two-headed arrows represent correlations. Estimated errors are omitted for visual clarity. Standardized coefficients are presented, and all paths and factor loadings are significant at p < .02 (two-tailed). Loadings notated by † were fixed to identify latent factors. Fit statistics: S-B χ²(177, N = 272) = 230.99, p = .004, robust comparative fit index (RCFI) = .098, root-mean-square error of approximation (RMSEA) = 0.034.