RESUMEN
Introduction: Women below the poverty threshold have lower representation and retention in breastfeeding studies. Methods: A secondary analysis of a longitudinal randomized controlled self-management for breast and nipple pain during breastfeeding study. Participants completed online surveys at discharge, weeks 1, 2, 3, 6, 9, 12, 18, and 24, with face-to-face interviews at 6 and 24 weeks. Text messages were sent to participants when modules and surveys were due. Retention was assessed in R with descriptive statistics, Mann-Whitney, Pearson's chi-square, and Cox Proportional Hazard Regression. Results: Two hundred and forty-four women (89 ≤$50,000 and 155 >$50,000) were recruited. Retention rates at 1 (93%), 2 (87%), 6 (82%), 9 (77%) and 24 (72%) weeks. For women of low income compared to those of high income there was a hazard ratio (HR) of 2.5 (p=0.0001) for retention. For non-Hispanic Black and Hispanic women compared to the combined non-Hispanic White and Other group, HRs for retention were 3.3 and 2.6 respectively (p=0.0001). Adjustment for age in the final hazard regression model of income, age, race and ethnicity decreased the HR for women of low income to 1.6 and HRs for non-Hispanic Black and Hispanic women to 2.1 and 1.9, respectively (p=.0001). However, none of the individual factors in the model achieved statistical significance. Discussion: Retention in breastfeeding studies impacts breastfeeding duration, a key lifelong preventative health behavior. Despite accessible study design, retention of women desiring to breastfeed was adversely affected by the intersection of income, race and ethnicity, and age.