Impact of Socio Demographic Factors on Health Outcomes: An Analytical Study Using Machine Learning Algorithms

Abstract

Socio-demographic factors, including age, gender, education level, socioeconomic status, and lifestyle characteristics, play a significant role in determining health outcomes. This study investigates the influence of socio-demographic variables on disease occurrence and severity among patients in Peshawar, Pakistan, using traditional statistical analysis and machine learning approaches. Four machine learning classifiers, including Decision Tree (DT), Random Forest (RF), Logistic Regression (LR), and Support Vector Machine (SVM), were developed and evaluated for health outcome prediction. The performance of the models was assessed using accuracy, precision, recall, F1-score, and Area Under the Receiver Operating Characteristic Curve (AUC-ROC). The Random Forest model demonstrated the highest predictive capability, achieving an accuracy of 92%, precision of 74%, recall of 68%, F1-score of 71%, and AUC value of 0.78, outperforming SVM (accuracy: 90%, AUC: 0.72), Logistic Regression (accuracy: 89%, AUC: 0.70), and Decision Tree (accuracy: 82%, AUC: 0.62). Feature importance analysis revealed that age was the most influential socio-demographic predictor, followed by lifestyle and occupational factors. Statistical evaluation confirmed significant associations between selected socio-demographic variables and health outcomes (p < 0.05). These findings demonstrate that machine learning models, particularly Random Forest, can effectively capture complex nonlinear relationships between socio-demographic characteristics and disease risk, providing valuable insights for targeted healthcare planning and preventive interventions.