As in the semaglutide cost-effectiveness analysis for MASH, this study uses a model to project costs and quality-adjusted life years. It follows a hypothetical cohort through simulated health states over decades. Novo Nordisk Pharma funded the study and manuscript, and two authors were company employees [1].
With that established, the interesting part is not the headline ratio. It is what the model had to assume to get there. Japan’s Optimal Use Promotion Guidelines restrict who is eligible for semaglutide 2.4 mg and cap how long they can stay on it. So the authors modeled zero, one or two retreatment courses, with a fixed one-year off-treatment period after each regimen. The two published ratios differ by approximately 1.34-fold. The model estimated a higher ratio in the cohort with type 2 diabetes.
The modeled population was people with a BMI of 35 or more plus at least one obesity-related comorbidity, or a BMI of 27 or more plus at least two obesity-related disorders. The intervention was semaglutide 2.4 mg with diet and exercise, against diet and exercise alone. Our comparison of liraglutide and semaglutide discusses a separate model with its own prices and assumptions. Complication costs in this model came from Japanese sources.
On that point, this desk will not convert the yen figure to dollars. A cost-effectiveness ratio is only meaningful against the willingness-to-pay threshold of the system that produced it, and Japan’s threshold, prices and complication costs are its own. An exchange rate applied to a health-economic ratio produces a number that looks informative and is not.
The authors flag the uncertainties themselves: retreatment efficacy and long-term discontinuation are the weak joints in the model, and both are exactly the places where real data is thin. The article on semaglutide dosing in clinical practice describes what one clinic observed during treatment. Those observations provide context for real-world dosing and do not establish how often patients restart after a treatment cap.