Predictive modeling of direct medical costs among frail older adults with diabetes mellitus in a rural outpatient setting

Background: Diabetes mellitus and frailty are associated with increasing healthcare costs, but there is currently a limited number of studies evaluating direct medical costs in elderly people with co-existing diabetes and frailty. Objective: To explore factors associated with total direct medical costs in elderly patients with diabetes and frailty. Methods: A cross-sectional descriptive study was conducted on 166 outpatients at a rural health center in Vietnam. Supervised predictive models, including linear regression and selected machine learning methods, were applied to predict direct medical costs. Results: Marital status, living with family, and duration of diabetes were the three variables with the highest level of importance in contributing to the predictive ability of the model. Unmarried and divorced individuals had lower costs compared to the married group (β = −0.29; 95% CI: −0.51 to −0.07; p = 0.009). Similarly, those living with family were associated with lower costs compared to those not living with family (β = −0.29; 95% CI: −0.50 to −0.08; p = 0.008). Duration of diabetes was positively associated with cost (β = 0.04; 95% CI: 0.00 to 0.09; p = 0.046). Conclusions: Social and health-related factors are associated with total direct medical costs among frail older adults with diabetes mellitus. Healthcare managers and policymakers should consider these factors when planning, designing, and implementing healthcare policies for older adults.