Background: The reliability of long-term maternal recall of breastfeeding has been assessed previously, but not maternal milk expression (pumping) and child consumption of expressed milk. Objective: To examine the reliability of maternal recall of feeding at the breast, maternal milk expression, and child consumption of expressed milk 6 years after delivery using the Brief Breastfeeding and Milk Expression Recall Survey (BaByMERS). Methods: At 12 months postpartum, women who delivered a singleton, live-born infant at >24 weeks of gestation at a major U.S. academic hospital completed BaByMERS. Five years later, they were recontacted to complete the same questionnaire. Kappa statistics (κ), intraclass correlation coefficient (ICC), and Bland/Altman plots examined agreement. Sociodemographics were examined through stratified comparisons. Results: Of 299 women who completed both questionnaires, 35% had a postgraduate education and 82% identified as white/Caucasian. Kappa statistics showed substantial agreement for ever breastfeeding or feeding breast milk (combined) (κ = 0.71, 95% confidence interval [CI]: 0.44-0.98) and ever feeding at the breast (κ = 0.76, 95% CI: 0.62-0.89). Recall for duration of feeding at the breast was excellent (ICC = 0.96, 95% CI: 0.95-0.97), and of maternal milk expression was slightly less so (ICC = 0.83, 95% CI: 0.80-0.97). Maternal minority race/ethnicity, lower educational attainment, unmarried marital status, public/no health insurance, and smoking were associated with lower reliability; these differences were usually small and not consistent across all feeding practices. Conclusions: Maternal recall of contemporary lactation and infant feeding using BaByMERS was strongly reliable 6 years after delivery. BaByMERS may be useful to collect recall data, with attention to subpopulations that may exhibit lower recall reliability.
Keywords: breastfeeding; infant feeding practices; lactation; milk expression; pumping; recall bias.