A computational framework for Economic burden estimation and decision support for policy development of mental Health Disorders in Oman

Abstract

This paper presents a computational framework for estimating the economic burden of mental health disorders in Oman using two factors cost-of-Illness (COI) and productivity loss modeling approach. The 10,000 mental health disorder dataset individuals across eleven governorates of Oman were analyzed. Statistical models such as Linear regression, Ridge, ElasticNet and deep learning model multilayer perceptron(MLP) are applied for datasets. Results show that the MLP model achieved higher accuracy for finding direct medical cost prediction with mean absolute error of 1.16, coefficient of determination (R²) of 0.9995. However linear models performed best for productivity loss estimation with higher accuracy. The total economic burden was estimated to be high precision across all models due to the deterministic COI relationship. The estimated average burden per case was 894.74 OMR and productivity loss contributing to an estimate of 54.5% from the current dataset. The MLP also proved effective in extracting complex characteristics, and the classification accuracy was 95.8%. These paper findings significantly advocate the hybrid modeling approach for development policy-driven healthcare planning in Oman.