To test whether perception of insufficient milk (PIM) supply in the breast-feeding relationship of one child predicts how long mothers breast-feed subsequent children, and whether this association differs for first-time mothers v. mothers with previous children.

Secondary analysis of Infant Feeding Practices Study II (ordinary least squares regression) and Year 6 follow-up.

Mailed, self-report survey of US mother–infant dyads, 2005–2012.

Women pregnant with a singleton were recruited from a consumer opinion panel. Exclusion criteria included: mother age <18; infant born <5 lbs, born before 35 weeks or with extended NICU stay, and mother or infant diagnosed with condition that impacts feeding. A subsample with PIM data (n 1460) was analysed.

We found that women who weaned because of PIM with the index child stopped breast-feeding 5·7 weeks earlier than those who weaned due to other reasons (4·9 weeks earlier for multiparas, P < 0·001; 7·1 weeks earlier for primiparas, P < 0·001). Using Year 6 follow-up data (n 350), we found subsequent child 1 weaned 9·2 weeks earlier if the mother experiences PIM as a multipara (P = 0·020) and 10·6 weeks earlier if the mother experiences PIM as a primipara (P = 0·019). For subsequent child 2 (n 78), the magnitude of association was even larger, although insignificant due to low power.

These findings indicate that PIM may carry forward in the reproductive life course, especially for first-time mothers. Perceptions of breast milk insufficiency and contributors to actual inadequate milk supply with the first child should be targeted, rather than intervening later in the reproductive life course.