Machine learning algorithm to predict determinants of home delivery after ANC visit among reproductive age women in East Africa: Using SHAP

Abstract Background: Home birth is described as a delivery that takes place at home without the presence of a skilled birth attendant. Home delivery after ANC visit is a major public health concern, and reducing the proportion of home births in East Africa is a key strategy for lowering the maternal death rate. However, no studies on this public health issue. Therefore, this study aimed to assess the machine learning approach to predict determinants of home delivery after ANC visits among reproductive-age women in East Africa. Methods: A community-based, cross-sectional study was conducted using a recent Demographic and Health Survey (DHS) from 2011 to 2021 data set. Nine supervised machine learning algorithms were employed on a total weighted sample of 44,123 women and evaluated using performance metrics using Python version 3.11 statistical software. This study also employed the most popular outlines of Yufeng Guo’s steps of supervised machine learning and SHAP analysis to predict and identify important predictors of home delivery after ANC visits in East Africa. Results: Home delivery after ANC visit was highest in Malawi, Uganda, and Kenya. Among the nine machine learning algorithms random forest was fitted for this study. The Beeswarm plot of SHAP analysis showed that being a rural resident of women and having a second trimester of ANC visit increases the likelihood of home delivery after an ANC visit. Whereas rich household income, secondary educational level of husband, contraceptive use, short birth interval, primary educational level of husband, having no problems of distance to health facility, and having above four ANC visits decreased women’s home delivery after ANC visits in East Africa. Conclusion: The random forest machine learning classification algorithms effectively predict home delivery after the ANC visit. As a result, this study recommends, considering the top ten determinants of home delivery and guaranteeing high-quality health institution services from a qualified practitioner, women should begin antenatal care services early and often throughout their pregnancies. Developing health facilities, promoting media health education, and encouraging women to get adequate information on health care services. Moreover, healthcare policy should give great consideration to women from low-income households.