Metformin in the Diabetes Prevention Program 3-year trial: The cost-effectiveness that never was.

Abstract

To highlight oversimplified quality-adjusted life year (QALY) calculations and incremental cost-effectiveness ratios (ICERs) regarding lifestyle metformin and placebo as flaws in the trial-based (Diabetes Prevention Program) cost-effectiveness analysis. We revised the QALY calculations to conform to convention and calculated appropriate ICERs using both original and revised QALYs results. We used several additional health economics tools to present results, showing the consistency of each method and the added value of each. We presented net monetary benefits, assessed decision uncertainty and calculated net losses. We also used graphics, making correct interpretation clearer and allowing illustration and calculation of technical and economic inefficiencies of treatments that were not cost-effective. ICERs with either QALY calculation indicated that metformin was technically and economically inefficient and not cost-effective. There was virtually no decision uncertainty. All methods employed pointed to an identical conclusion. In contrast to original claims, although lifestyle was cost-effective in diabetes prevention, metformin was not, and acting otherwise imposes significant monetary and health costs. Various available tools of economic evaluation would probably have prevented the original misinterpretation by the authors had they been used. The varied tools implemented here illustrate with a common example their consistency and value to the field, showing how to depict the results in various ways.