A Funding Gap, Not a Science Gap, Is Keeping AI Out of Hospitals

Somewhere in America right now, a patient with a dangerous but treatable heart condition is being misdiagnosed. Not because the technology to catch it doesn't exist. It does. The problem is that most hospitals can't afford to turn it on.

That is the uncomfortable truth behind a gap in how American hospitals adopt artificial intelligence. Clinical AI tools that are ready for widespread use, tested, validated, cleared by regulators, simply aren't reaching most hospitals. The holdup isn't scientific. It's financial. And federal health officials have solved this exact kind of problem before, twice, by letting industry fund adoption under strict guardrails while the government's permanent payment systems caught up.

Consider two conditions side by side. When a stroke patient arrives at the hospital, an AI tool can flag it on a CT scan and coordinate the response across every provider involved in the case. This is now standard practice at many hospitals. A study of 166 hospitals and more than 14,000 patients found it cut time to treatment by nearly 40 minutes, a 44% reduction. That's the difference between a patient walking out of the hospital and one who doesn't.

Now consider hypertrophic cardiomyopathy, a genetic heart condition that can cause sudden cardiac arrest in people who otherwise appear perfectly healthy. AI tools can flag suspect patients on a routine EKG years before symptoms ever show up. Yet roughly 60% of cases are still misdiagnosed, often after four incorrect diagnoses first. The clinical logic behind both tools is identical. What differs is whether Medicare has created a billing code for it. Health systems consistently cite the lack of reimbursement as the primary barrier to adopting more AI tools, and that single administrative fact, not the quality of the science, is deciding who gets caught in time and who doesn't.

The Bridge That Keeps Getting Delayed

To its credit, the Centers for Medicare and Medicaid Services is not standing still. It has proposed a new payment category, Software as a Medical Service, meant to standardize reimbursement for AI tools that read a scan, an EKG, or a chart and return a diagnosis or risk score. It has stood up a dedicated AI office. It is testing outcome-based payment models for technology-enabled chronic care, with the first wave alone targeting 36 new codes.

But rulemaking moves at the pace of rulemaking. Claims-data cycles take years to turn. AI tools keep emerging faster than that. Once a billing code exists, adoption tends to follow fairly quickly. It's the years before that code exists where an already validated tool just sits there, unused, while patients who could benefit from it don't.

CMS actually built a bridge for this kind of gap two decades ago. The New Technology Add-on Payment program, known as NTAP, was designed to provide temporary reimbursement for a handful of new drugs and devices each year while a permanent payment pathway got established. The trouble is that NTAP was never sized for the sheer volume of AI tools now clearing FDA review, and CMS has proposed retiring the program altogether. That would strip away one of the few bridges hospitals currently have during the waiting period, right as the number of tools waiting behind it grows fastest.

Here's where history offers a clearer answer. Since 2006, the Department of Health and Human Services has protected the donation of electronic health record software under a dedicated safe harbor and a Stark Law exception, on the straightforward logic that getting infrastructure into hospitals faster serves patients better. In 2020, facing a near-identical gap in hospital cybersecurity, the Office of Inspector General did it again for donated cybersecurity technology. Hospitals now run donated EHR and cybersecurity infrastructure at scale, and neither arrangement is considered problematic under the Anti-Kickback Statute. It's a tested mechanism, not a theoretical one.

Applying that same tool to AI clinical pathway technology would close the current gap without waiting for CMS rulemaking to catch up. Life sciences companies already fund clinical trials inside hospitals under comparable guardrails: independent clinical judgment, no product promotion, transparent evidence standards. An AI Clinical Pathway Safe Harbor could work the same way, permitting life sciences and technology companies to fund adoption of AI tools that meet three conditions. The tool can't favor one company's drug, device, or service over another's. It has to be grounded in peer-reviewed evidence or established clinical guidelines. And the sponsor can't have any hand in what the tool actually tells the clinician.

Put plainly: a company can pay for the infrastructure, but it cannot buy influence over what a doctor is told to prescribe or order. That distinction matters enormously, and it is exactly the line regulators have already drawn, twice, in the EHR and cybersecurity contexts.

The cost of waiting isn't spread evenly. Rural and critical access hospitals report under 50% adoption of even basic predictive tools, compared with 86% among hospitals affiliated with top-tier health systems. Every year without a bridge widens that gap. And it widens fastest for the hospitals with the least financial room to absorb the delay, the ones serving communities that already face the thinnest margin of medical access.

Why It Matters

HHS has the authority to act today, through the same rulemaking process it used to build the EHR and cybersecurity safe harbors: propose a rule, take public comment, finalize it. Nobody is seriously disputing the clinical evidence behind these tools. What's missing is the financing bridge to get them into the hospitals that need them most, particularly the rural and under-resourced ones falling further behind each year.

This isn't a debate about whether AI belongs in medicine. It already does, in the hospitals that can afford it. The real question is whether financing catches up to the science in time to matter for the patients who are being misdiagnosed right now, while the paperwork slowly works itself out.