CMS Proposes New Payment Framework for Clinical AI Software

Medicare moves to standardize AI reimbursement

The Centers for Medicare and Medicaid Services has proposed a new approach to paying for clinical artificial intelligence and software tools, acknowledging that its current payment methodology struggles to account for algorithmic diagnostics and digital clinical applications.

In proposed rules released this month covering hospital outpatient payments and physician fees for 2027, CMS outlined plans to create a more systematic payment structure that considers how software-based tools affect patient outcomes. The agency characterized the 2027 changes as an interim step focused on revising how it categorizes and reimburses several clinical software and AI services.

Why it matters

Medicare's payment policies directly shape which technologies hospitals and physicians adopt. Without clear reimbursement pathways, even clinically validated AI tools face adoption barriers. A standardized framework could accelerate deployment of algorithms that improve diagnostic accuracy or treatment planning while establishing precedent for how payers evaluate digital health technologies' value. The outcome-based approach also signals a shift from paying for tools based on their cost to paying based on their clinical impact.

The reimbursement challenge

CMS has historically excelled at calculating costs for tangible medical items and equipment depreciation. A cotton swab has a clear unit cost. A CT scanner's wear and tear can be amortized over its useful life. But algorithmic tools present a different challenge.

An algorithm that analyzes CT scans to predict cardiac risk or an AI system that maps prostate cancer progression lacks the physical cost structure Medicare's payment models were designed to handle. Development costs, ongoing maintenance, validation studies, and computational infrastructure don't fit neatly into existing reimbursement categories.

The proposed framework represents CMS's attempt to build payment logic that accounts for software's unique economics while incorporating clinical value. By factoring patient outcomes into the payment structure, the agency is moving beyond cost-plus reimbursement toward value-based payment for digital tools.

Implications for 2027 and beyond

The agency explicitly described the 2027 proposals as transitional. The interim labeling and payment changes for select clinical software and AI services suggest CMS is testing approaches before implementing a broader policy framework.

This measured rollout gives the agency time to gather data on how outcome-based payments for software perform in practice, while giving developers and providers visibility into Medicare's evolving reimbursement philosophy for AI-enabled care.

The details were first reported by STAT News as part of its "Paying for AI" series examining how clinical artificial intelligence tools affect healthcare costs and patient outcomes.