J Perinat Neonatal Nurs. 2025 Dec 9. Revista: 10.1097/JPN.0000000000000977. Online ahead of print.

PURPOSE: This study aimed to evaluate the effect of machine learning (ML)-based breastfeeding training (MLBT) on breastfeeding knowledge and continuation among mothers at risk of early cessation, identified by an ML model.

DESIGN AND METHODS: This quasi-experimental study, conducted in 2 phases using a pre-posttest design, included 90 mothers (45 intervention, 45 control). In Phase 1, an ML was developed to identify mothers at risk of early breastfeeding cessation. In Phase 2, MLBT training and data collection tools were developed based on risk profiles and delivered to the intervention group. Data were analyzed using chi-square tests, t-tests, and repeated-measures ANOVA.

RESULTS: The intervention group showed significant improvements in breastfeeding knowledge across all 4 MLBT modules (P < .01) and in total knowledge scores (P < .001). At 2 and 4 months postpartum, full and partial breastfeeding rates were significantly higher in the intervention group than in the control group (P < .005).

CONCLUSION: The MLBT provided to mothers identified by the ML model effectively enhances breastfeeding knowledge and supports its continuation. Integrating ML into nursing practice can improve health care quality through more efficient and personalized solutions.

PRACTICE IMPLICATIONS: In a rapidly changing global context, innovative and evidence-based approaches are essential for breastfeeding support. This study offers guidance for developing and improving programs that promote sustained breastfeeding, contributing to maternal and infant health.

PubMed:41361682 | Revista:10.1097/JPN.0000000000000977