Evaluating the effectiveness of breastfeeding education provided to mothers predicted by machine learning to cease breastfeeding early

Evaluating the Effectiveness of Breastfeeding Education Provided to Mothers Predicted by Machine Learning to Cease Breastfeeding Early This research was conducted with the aim of developing a machine learning (ML) model to identify mothers who are at risk of early discontinuation of breastfeeding, and to evaluate the effectiveness of breastfeeding education provided to these mothers. The study employed an experimental design, including pre-test and post-test with intervention and control groups. The sample of the study consisted of a total of 90 mothers (45 intervention, 45 control group) who attended postnatal check-ups at an education and research hospital and seven family health centers in the city center of Ankara, and who met the identified risk factors. The research was conducted between February 4, 2022, and October 30, 2023. In the first phase of the research, an ML prediction model was developed using data from the TNSA 2013-2018 to identify mothers at risk of early breastfeeding cessation. In the second phase of the research, breastfeeding education programs were developed for the risk group. Data were collected between May 11 and October 30, 2023, using the Mother-Baby Information Form and the Breastfeeding Education Evaluation Form (BEEF). The data were analyzed using various statistical methods including frequency, percentage, mean, Pearson-χ2, independent and paired samples t-tests, Cramer's V, and Cohen's d. In the study, it was found that the scores from the BEEF significantly increased in the intervention group after the education (p<0.05); however, the difference between the intervention group and the control group was not statistically significant (p>0.05). Although there was no significant difference in breastfeeding rates between the groups before the education, the full breastfeeding rates increased, and partial breastfeeding and cessation rates decreased in the intervention group after the education. These differences were statistically significant at 2 and 4 months postpartum (p<0.05). The breastfeeding education provided to mothers identified by the ML prediction model was found to be effective in sustaining breastfeeding. It is recommended to use ML algorithms to identify mothers at risk of discontinuing breastfeeding before six months and to offer AI-based solutions in breastfeeding counseling practices. Keywords: Artificial intelligence, breastfeeding prediction model, innovative nursing solution, machine learning, risk of discontinuing breastfeeding