The role of partner stress on the relationship between maternal stress and breastfeeding practices in Quebec : a moderation analysis

Background: The benefits of breastfeeding have been well-established for maternal and infant health, yet exclusive breastfeeding rates in Canada remain below recommended levels. Although poorer maternal mental health has been linked to early cessation and non-exclusive breastfeeding, the specific role of perceived maternal stress during pregnancy and postpartum remains insufficiently understood. Moreover, the influence of partner stress on this relationship has not been examined, despite the interrelated nature of parental mental health. Objective: This study aimed to examine (1) whether prenatal maternal perceived stress is associated with breastfeeding status at three months (none, mixed, or exclusive); (2) whether prenatal and postpartum maternal perceived stress are associated with time to cessation of any breastfeeding; and (3) whether partner stress moderates these associations. Methods: Secondary analyses were conducted with the 3D (Design, Discover, Develop) longitudinal cohort in Quebec, Canada (n=2,366). Outcomes included breastfeeding status at three months (n=1,643) and time to cessation of any breastfeeding (n=1,717). Maternal and partner perceived stress were assessed prenatally and postpartum using the Perceived Stress Scale (PSS-4). Partial proportional odds and Cox proportional hazards models were used, evaluating effect modification by partner stress using interaction terms (α=0.10). Models were adjusted for parental education level, household income, maternal age, maternal ethnicity, parity, and birth stressors. Results: Prenatal and postpartum maternal stress were not significantly associated with time to cessation of any breastfeeding, nor moderated by partner stress. Prenatal maternal stress was not associated with breastfeeding status at three months; however, evidence suggests that partner stress modified this association for both any (mixed/exclusive) versus none (OR=1.12, 95% CI:0.98–1.28, p=0.100) and exclusive versus mixed/none (OR=1.10, 95% CI:0.98–1.24, p=0.091). Marginal predicted probabilities suggested that higher concordant parental stress levels corresponded to higher probabilities of any and exclusive breastfeeding relative to lower or discordant stress levels. Conclusions: Findings suggest that early breastfeeding behaviors may be better understood within a dyadic parental context, where moderate levels of shared parental stress may reflect a coordinated response that supports breastfeeding. Further research is needed to examine whether these patterns are consistent across different population groups and levels of perceived parental stress.