Predictive Model for the Average Cost per Treatment Episode of Type 2 Diabetes Using Multivariate Regression Based on Health Insurance Database in Hanoi, Vietnam

Background: Type 2 diabetes mellitus (T2DM) is a chronic condition requiring lifelong treatment; therefore, the development of a multivariate regression model to support cost prediction is essential. Objective: To develop a multivariate regression model for predicting the average treatment episode cost among patients with T2DM in Hanoi, Vietnam. Methods: A cross-sectional descriptive study was conducted using retrospective data from the Hanoi Social Health Insurance database, covering T2DM patients from January 1, 2018, to December 31, 2022. Results: The study included 1,397,287 patients across 6,940,296 treatment episodes. The male-to-female ratio ranged from 1:1.07 to 1:1.08, with a mean age ranging from 64.59 ± 11.16 to 64.95 ± 11.78 years. The average direct medical cost per treatment episode showed an increasing trend over the years, from 115.93 USD (95% CI: 115.33 – 116.54 USD) in 2018 to 198.10 USD (95% CI: 177.40 – 179.74 USD) in 2022. The most significant increase was observed during 2019–2020 (22.44%). Medication costs accounted for the largest proportion of the total cost and showed a decreasing trend, from 57.92% to 52.01%. Notably, 90% of the treatment cost was reimbursed by health insurance. Conclusion: The average direct medical cost per treatment episode accounted for approximately 8.54% to 14.61% of the average monthly GDP per capita in Vietnam in 2024. The developed model provides a valuable tool for policymakers to forecast costs and implement cost-saving strategies.