810-P: The Diminishing Cost-Effectiveness of the Newer Glucose-Lowering Drugs in the United States, 2010-2018

Background: The difference between the costs of the newer and older glucose-lowering drugs has been steadily increasing since 2010. In 2018, newer drugs cost 8-12 times more than older drugs (except for insulin). This study aimed to understand how this change in drug cost influenced the cost-effectiveness (CE) of newer drugs compared with older drugs. Methods: We conducted a systematic literature review to identify US-based CE studies comparing newer (i.e., dipeptidyl peptidase-4 inhibitors (DPP4), glucagon-like peptide 1 receptor agonists (GLP1-RA), and sodium-glucose transport protein 2 inhibitors) with older glucose-lowering drugs. The identified 12 studies all reported the CE of newer drugs based on drug costs estimated before 2010. We have updated the corresponding CE of each study by replacing the old cost estimates with 2018 estimates from the 2018 IBM® MarketScan® Commercial Claims Databases. The time window and willingness to pay threshold were consistent with each original study. Results: Only 8% of the original studies suggested that the older drugs were cost-effective. However, 58% of studies were in favor of the older drugs after the cost update. Among the 4 studies comparing newer drugs with thiazolidinediones, all studies suggested thiazolidinedione to be cost-effective in the updated analysis, while the original results all favored newer drugs. For the 4 studies comparing newer drugs with sulfonylureas, 2 studies suggested the sulfonylureas to be cost-effective after the cost update. All 4 studies suggested newer drugs to be cost-effective when compared with insulin in the original study. Only 1 flipped its conclusion when 2018 costs were used. Our sensitivity analysis shows that our results are robust up to a 30% rebate. Conclusion: Significant changes in the cost of glucose-lowering drugs have impacted the economic value of different glucose-lowering agents substantially. More CE analyses are warranted to support the drug choice in T2DM management.