Defining Low Milk Supply: A Data-Driven Diagnostic Framework and Risk Factor Analysis for Breastfeeding Women.

Background: Current low milk supply (LMS) definitions use subjective maternal perceptions or arbitrary thresholds for 24 h milk production (MP), potentially misclassifying cases. This study aimed to re-evaluate the definition of LMS using data-driven approaches and investigate associated maternal risk factors. Methods: Lactating mothers 4-26 weeks postpartum (n = 460) provided demographic, obstetric, and infant data and measured 24 h MP and infant milk intake using the test-weighing method. Infant growth was calculated as their weight-for-age z-score. Latent profile analysis, receiver operating characteristic curve analysis, and multinomial logistic regression were used for classification, diagnostic evaluation, and risk factor assessment for LMS. Results: Four milk supply classes emerged: Class 1 with adequate MP, infant intake and infant growth (n = 254); Class 2 with high MP exceeding infant demand and adequate growth (n = 30); Class 3 with slow infant growth despite moderate MP (n = 120); and Class 4 with extremely low MP and high formula intake (n = 56). Classes 1 and 2 were grouped as the normal milk supply group (61.7%), while Classes 3 and 4 formed the LMS group (38.3%). New thresholds were identified for 24 h MP (708 mL/24 h, area under the curve (AUC) = 0.92) and infant breast milk intake (694 mL/24 h, AUC = 0.94) with high diagnostic accuracy. Moreover, practical alternative thresholds for infant average daily weight gain (26 g, AUC = 0.89), formula intake (122 mL/24 h, AUC = 0.89) and formula-to-growth ratio (4 mL/g, AUC = 0.94) were established for the identification of LMS. Minimal breast growth during pregnancy (Odds ratio (OR) = 4.6, 95% confidence interval (CI): 2.3-9.6), advanced maternal age (OR = 2.1, 95% CI: 1.0-4.5), and gestational diabetes mellitus (OR = 2.1, 95% CI: 1.1-4.0) were significant risk factors related to the LMS subgroups. Co-existence of maternal advanced age and overweight showed greatly amplified risk of LMS (OR = 3.7, 95% CI: 1.3-10.5), and a more pronounced risk was observed for the combination of minimal breast growth and advanced maternal age (OR = 9.2, 95% CI: 3.0-28.3). Conclusions: This data-driven classification of LMS and identified risk factors may enhance the precision of LMS diagnosis and guide targeted interventions for lactating mothers.