Abstract
We evaluated the cost-effectiveness of sulfonylurea combined with metformin (SU+Met) or pioglitazone (SU+Pio). The simulation model we developed to predict long-term prognosis and medical costs in a case of type 2 diabetes mellitus uses UKPDS regression equations and estimates annual risk for eight complications. Medical costs for individual complications were set based on information from the literature. Long-term prognosis was evaluated using quality-adjusted life years for Japanese subjects by adjusting the difference in incidence for ischemic heart disease and stroke between those observed using JDCS and using UKPDS regression equations. SU+Met had better QALYs at lower cost than SU+Pio. This model should be updated as new epidemiological data in Japan becomes available.
