Artificial Intelligence-Assisted Nursing Interventions for Predicting Diabetes Among Homeless Adults

Background: Diabetes mellitus is a global health challenge, especially among the homeless. Nursing plays a critical role in prevention, early detection, and management of diabetes. Aim: to predict diabetes among homeless adults by utilizing artificial intelligence techniques-assisted nursing intervention. Design: A case-control design was utilized to achieve the aim of the current study. Setting: The study was conducted at the Ma'ana Rescue Human Foundation. Sample: A purposive sample of 150 homeless adults was included in this study. Tools for data collection : First tool: a structured interview questionnaire for homeless adults that has four parts. Part I: Demographic data; Part II: Medical history; Part III: Lifestyle factors; Part IV: Knowledge about diabetes. Second tool: Physiological measurements. Results: More than half of diabetic homeless were aged 60 years or more and two-thirds of them had moderate lifestyles and low knowledge about diabetes. The majority had hypertension and central obesity. Various of machine learning models were evaluated. A hybrid meta-learning classifier was subsequently developed, utilizing a stacking ensemble of six base learners, including logistic regression, Support Vector Machine, Random Forest, Decision Tree, and K-Nearest Neighbors, along with an XGBoost Meta-learner. The hybrid model markedly surpassed the individual classifiers, attaining an accuracy of 95.45%, an F1-score of 0.967, and an AUC of 0.979. Conclusion: Artificial intelligence techniques may be able to reliably predict diabetes. Integrating AI into nursing assessment and care planning enhances the ability to identify high-risk individuals early, thereby improving health outcomes. The proposed hybrid stacking model outperformed conventional classifiers in terms of prediction performance, highlighting the benefits of ensemble learning and sophisticated resampling strategies in dealing with imbalanced medical data. Recommendations: It is recommended that healthcare institutions integrate AI-powered diagnostic assistance technology into clinical processes to aid in the early detection and treatment of diabetes.