AI Healthcare ROI — Why Proving Cost-Effectiveness Remains Elusive
OVERVIEWContent
Overview
Health systems are increasingly demanding rigorous evidence before acting on AI vendor ROI claims, even as those claims grow bolder. A parallel tension has emerged in which AI is simultaneously described by health system CIOs as the most overhyped and the most exciting trend in healthcare IT.
Vendor Scrutiny at Mass General Brigham
Rebecca Mishuris, MD, Chief Health Information Officer at Mass General Brigham, requires every AI vendor to complete a standardized assessment questionnaire before any procurement conversation can advance. This gatekeeping approach reflects a broader institutional shift toward evidence-based AI procurement, where bold ROI projections must be substantiated with real-world data before pilots are authorized.
CIO Ambivalence
A Becker's Hospital Review poll of CIOs and Chief Digital Officers at the 200 largest U.S. health systems found that for a notable subset of the 25 anonymous respondents, AI was both the most overhyped trend and the trend they were most genuinely excited about. This dual characterization reflects the gap between AI's demonstrated value in narrow use cases and the sweeping transformation claims made by vendors and commentators.
Structural Barriers to ROI Proof
Proving AI cost-effectiveness in healthcare is complicated by attribution challenges (AI-assisted outcomes are difficult to isolate), long payback periods, workflow integration costs, and the absence of standardized ROI measurement frameworks. Until clearer evidence standards emerge, health systems will continue to face pressure from both vendor enthusiasm and internal skepticism.