Cost-effectiveness of semaglutide versus dulaglutide for Type 2 Diabetes in China: A Markov Model analysis.

Abstract

From the perspective of China's basic medical insurance, to evaluate the cost-effectiveness of semaglutide versus dulaglutide for type 2 diabetes mellitus (T2DM) in China, informing clinical and health policy decisions. This study employed a Markov model to simulate the disease progression of T2DM over a 25-year time horizon. Transition probabilities between health states were sourced from the UK Prospective Diabetes Study Outcomes Model (UKPDS 82) and the SUSTAIN 7 clinical trial, while health utility values were obtained from published literature using the Chinese EQ-5D-5L value set. Long-term cost-effectiveness was projected through cohort simulation, with both one-way sensitivity analysis and probabilistic sensitivity analysis (PSA) conducted to evaluate the robustness of the model. After 25 years of simulation, the total cost for dulaglutide was ¥193,353.07, with 11.628 quality-adjusted life-years (QALYs) gained. The total cost for semaglutide was ¥240,925.08, yielding 11.900 QALYs. Compared with dulaglutide, semaglutide had an incremental cost-effectiveness ratio (ICER) of ¥174,904 per QALY gained, approximately twice China's 2023 per capita GDP (¥89,358). According to the Chinese Guidelines for Pharmacoeconomic Evaluation (2020), this result falls within the economically acceptable range (1 × GDP < ICER ≤ 3 × GDP). One-way sensitivity analysis showed that although health state utility values were the most influential parameter, the model results remained robust. PSA indicated that at a willingness-to-pay (WTP) threshold of 3 times the per capita GDP (¥268,074), semaglutide had an 89.6% probability of being considered economically acceptable. From the perspective of China's basic medical insurance, semaglutide demonstrates cost-acceptable outcomes compared to dulaglutide in managing type 2 diabetes. This analysis provides crucial evidence for clinical decision-making and health policy formulation.