Baseline Risk Heterogeneity in Cost-Effectiveness Analysis of the Diabetes Prevention Program
Author(s)
Sultan Alolayan1, Tewodros Eguale, PhD, MD2, Alissa R. Segal, PharmD2, Brian Rittenhouse, PhD3.
1MA, USA, 2Massachusetts College of Pharmacy and Health Sciences, Boston, MA, USA, 3Professor, MCPHS University, Winchester, MA, USA.
1MA, USA, 2Massachusetts College of Pharmacy and Health Sciences, Boston, MA, USA, 3Professor, MCPHS University, Winchester, MA, USA.
OBJECTIVES: Several cost-effectiveness analyses (CEAs) have examined the entire sample from the Diabetes Prevention Program (DPP) trial, concluding that both metformin and individual lifestyle modification were cost-effective. Critiques and revisions to these CEAs have shown that only lifestyle interventions are cost-effective when Incremental Cost-Effectiveness Ratios (ICERs) are appropriately calculated. There has been speculation that cost-effectiveness may differ by baseline risk of diabetes. We sought to examine whether metformin might be cost-effective in a high baseline risk group.
METHODS: Three-year cumulative incidence of diabetes data were available based on a diabetes progression model applied to the DPP, separating the DPP sample into 4 baseline risk groups. We estimated cumulative incidences in year 1 and 2 and applied those data to data also available on utilities and costs conditional on diabetes status to obtain average total costs and QALYs for three years by treatment. We estimated two QALY values - one as the original authors had done and one more conventional. As in the original DPP, CEAs we also formulated a group lifestyle option (with assumed lower costs but equal effect). We then calculated appropriate ICERs to determine cost-effectiveness.
RESULTS: Quartiles 1-3 results were qualitatively identical to those of the overall sample (lifestyle was cost-effective, not metformin). In the high-risk group it is possible that there is a place for metformin as a cost-effective therapy. The group alternative was cost-effective with conventional QALYs and a willingness-to-pay (WTP) of more than $155,000 (metformin otherwise). Using original CEA, unconventional QALYs, group was cost-effective if WTP was more than $61,840. Individual lifestyle was not cost-effective with either QALY calculation.
CONCLUSIONS: Metformin or group lifestyle treatments are each possibly cost-effective in the high-risk subgroup. WTP thresholds and type of QALY calculations will determine which. It is likely that all previous DPP CEAs erred by not considering effects of baseline risk.
METHODS: Three-year cumulative incidence of diabetes data were available based on a diabetes progression model applied to the DPP, separating the DPP sample into 4 baseline risk groups. We estimated cumulative incidences in year 1 and 2 and applied those data to data also available on utilities and costs conditional on diabetes status to obtain average total costs and QALYs for three years by treatment. We estimated two QALY values - one as the original authors had done and one more conventional. As in the original DPP, CEAs we also formulated a group lifestyle option (with assumed lower costs but equal effect). We then calculated appropriate ICERs to determine cost-effectiveness.
RESULTS: Quartiles 1-3 results were qualitatively identical to those of the overall sample (lifestyle was cost-effective, not metformin). In the high-risk group it is possible that there is a place for metformin as a cost-effective therapy. The group alternative was cost-effective with conventional QALYs and a willingness-to-pay (WTP) of more than $155,000 (metformin otherwise). Using original CEA, unconventional QALYs, group was cost-effective if WTP was more than $61,840. Individual lifestyle was not cost-effective with either QALY calculation.
CONCLUSIONS: Metformin or group lifestyle treatments are each possibly cost-effective in the high-risk subgroup. WTP thresholds and type of QALY calculations will determine which. It is likely that all previous DPP CEAs erred by not considering effects of baseline risk.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
EE71
Topic
Economic Evaluation, Health Technology Assessment, Study Approaches
Topic Subcategory
Trial-Based Economic Evaluation
Disease
Diabetes/Endocrine/Metabolic Disorders (including obesity)