When Health Behavior Meets Technology: Artificial Neural Network-Based Prediction of Breast Self-Examination...

Early detection of breast cancer is crucial, particularly in low-resource settings. Emerging technologies like artificial neural networks (ANNs) offer innovative approaches to predict health behavior intentions. Hence, this study used an artificial neural network-multilayer perceptron (ANN-MLP) model to determine the predictors of breast self-examination (BSE) intention among female college students. With ethical approval, an analytical cross-sectional study was conducted among 382 female college students from a private higher education institution. Data on sociodemographic factors, knowledge, practices, perceived barriers, and health beliefs were collected using validated research questionnaires. Pearson's correlation, chi-square, and ANN-MLP modeling were performed to identify significant predictors of BSE intention and actual breast screening behavior. Knowledge adequacy, frequent BSE practice, perceived susceptibility, and self-efficacy were significant predictors of intention to perform BSE. The ANN-MLP model achieved a testing accuracy of 74.1% and an AUC of 0.698 in predicting BSE intention. For actual breast screening behavior, significant predictors included degree program, knowledge, practice, self-benefit, and self-efficacy. The corresponding ANN-MLP model achieved 88.7% testing accuracy with an AUC of 0.779. The interplay of cognitive and behavioral factors significantly influences BSE intentions. The use of ANN-MLP modeling can effectively determine key predictors of BSE intentions, offering opportunities for robust, data-driven, and targeted health promotion interventions. The findings may inform patient, public, and professional cancer education by identifying cognitive and behavioral predictors of BSE intention, enabling health professionals to design tailored education and training programs that enhance early screening behaviors and advocacy for breast cancer prevention.