A Quiet Revolution in Healthcare Reimbursement
Somewhere in the fine print of American healthcare policy, a shift is underway that could define how artificial intelligence gets deployed in medicine for the next decade. It's called ACCESS — and even the engineers building health AI tools have largely missed it.
The core problem ACCESS solves is deceptively simple: there has never been a formal mechanism for Medicare to pay for an AI agent doing clinical work between patient visits. A physician can bill for an office appointment. A nurse can bill for a procedure. But an AI system that calls a patient to check on their recovery, coordinates a housing referral, or nudges someone to pick up a prescription? Under the old model, that work existed in a financial void.
ACCESS changes that. For the first time, it creates a reimbursement pathway specifically designed to compensate for longitudinal, AI-assisted patient management — the kind of continuous, low-intensity care that happens in the gaps between formal medical encounters.
Why the Gaps Matter
Healthcare researchers have long understood that what happens between clinical visits often determines outcomes more than the visits themselves. A patient discharged after a cardiac event who misses a follow-up, stops taking medication, or faces an unresolved social determinant of health — housing instability, food insecurity — is at high risk of readmission.
Human care coordinators have tried to fill these gaps for years, but the scale is unmanageable. There are roughly 67 million Medicare beneficiaries in the United States. Keeping meaningful human contact with even a fraction of high-risk patients strains health systems to their limits.
AI agents — systems capable of initiating outbound calls, parsing patient responses, flagging anomalies, and triggering referrals — are technically capable of operating at that scale today. What they've lacked is a billing code.
The Business Model That Wasn't There
For health AI startups, the absence of a reimbursement model has been a ceiling on ambition. You can build a system that monitors diabetic patients overnight and alerts a clinician to a dangerous trend. But if Medicare won't pay for the monitoring, hospitals won't buy the product, and investors won't fund the company.
ACCESS removes that ceiling. By establishing that AI-assisted between-visit care is a reimbursable service, it signals to the market that health AI isn't just a cost-savings play — it can be a revenue-generating line item.
Analysts expect the model to accelerate consolidation in the health AI space, with larger players better positioned to navigate Medicare compliance moving quickly to establish first-mover advantage in newly billable service categories.
What It Means Beyond the US
For Canadian healthcare observers — including those watching closely from Ottawa, where federal digital health policy is an active file — the ACCESS model is worth studying. Canada's publicly funded system faces similar structural gaps: the space between discharge and follow-up, between a family doctor visit and a specialist referral, where patients fall through.
The political appetite for paying AI systems to do clinical work is far more constrained in Canada than in the US. But as evidence accumulates on the outcomes and economics of AI-assisted longitudinal care, provincial health ministries will face growing pressure to define their own reimbursement frameworks.
For now, ACCESS is an American story. But the underlying logic — that AI doing clinical work should be paid for that work — is a conversation that has only just begun.
Source: TechCrunch
