Fig. 3

Mediation models for breastfeeding at 12 months and (A) postpartum depression and (B) parent-child dysfunction and (C) both mediators combined and total child behaviour problems. Notes: Path A are odds ratios indicating the odds of depression at 1 year (score ≥ 16) or parent-child dysfunction at 2 years (score ≥ 20) if breastfed at 12 months. Path B are beta estimates indicating change in total behaviour problems score if depression score ≥ 16 or parent-child dysfunction ≥ 20. Path C is an odds ratio indicating the odds of parent-child dysfunction score ≥ 20 if depression at 1 year score ≥ 16. Total path are beta estimates indicating the change in total behaviour problems score if breastfed at 12 months. Direct path are beta estimates indicating the change in total behaviour problems score not through depression at 1 year or parent-child dysfunction at 2 years. Indirect path are beta estimates indicating change in total behaviour problems score through depression at 1 year and/or parent-child dysfunction at 2 years. Reference group for the breastfeeding practice is no breastfeeding at 12 months. Adjusted models in all figures and tables include the following confounders: child sex, prenatal stress, prenatal depression, study site, birth mode, birthweight, gestational age, household income, maternal education, maternal race, marital status, number of older siblings, prenatal smoking, attention deficit hyperactivity disorder genetic risk score. ^p ≤ 0.1, *p ≤ 0.05 **p ≤ 0.01, ***p ≤ 0.001