Enhancing Mental Health Decision-Making with Artificial Intelligence/Machine Learning: A Prescriptive Analytics Approach for Customised Outcomes

Depression is a complex and widespread mental health condition affecting over 280 million people globally, yet access to timely diagnosis and personalised treatment remains limited. This study explores the application of artificial intelligence and machine learning (AI/ML) to develop a prescriptive analytics framework for supporting mental health decision-making. Using a dataset of 2,556 anonymized datapoints, the study evaluates multiple machine learning classifiers to identify depression risk, benchmarked by accuracy, F1 score, and AUC metrics. Logistic Regression emerged as the most balanced and interpretable model, achieving an accuracy of 97.3% and an AUC of 99.9%, with the lowest false-negative rate—crucial in a clinical support context. The study then introduces an ensemble modeling framework based on stratified Logistic Regression models, enabling consensus-based predictions and enhanced reliability. Beyond predictive analytics, the framework facilitates prescriptive insights by identifying minimal, targeted lifestyle changes likely to shift individuals from high to low depression risk. Feature importance was established using statistical and effect size measures, guiding personalized intervention suggestions with practical interpretability. The study highlights relationships among key features such as academic pressure, work hours, job satisfaction, and financial stress, revealing compound intervention opportunities. Despite limitations related to dataset provenance and generalizability, the framework demonstrates promising utility in non-clinical mental health contexts and forms a tractable foundation for future deployment in professional settings. These findings underscore the potential of AI/ML to augment mental health care by delivering scalable, explainable, and data-driven decision support tools aimed at improving patient outcomes through personalized, prescriptive strategies.