The Centers for Medicare & Medicaid Services has selected 150 participants for ACCESS, a ten-year programme designed to test outcome-based payment for chronic care. Formally known as Advancing Chronic Care with Effective, Scalable Solutions, the model began on July 5 and covers conditions including diabetes, hypertension, chronic kidney disease, obesity, depression and anxiety. Its structure could make AI-supported monitoring and coordination economically viable in ways that traditional clinician-time reimbursement does not.
Payment for outcomes rather than activities
ACCESS gives participating organisations predictable payments for managing qualifying conditions, with full payment tied to measurable health goals. Examples include improved blood-pressure control or reduced pain. This differs from paying solely for a defined appointment, check-in or unit of clinician time.
That distinction is the programme’s technological significance. An AI voice agent may contact a patient, coordinate a referral or encourage medication collection without fitting neatly into a conventional billing event. Outcome-based payment can support those activities if they contribute to verified improvements in care.
Pair Team offers an early operating example
Pair Team, founded in 2019, is among the selected organisations. The company combines medical, behavioural and social care for patients whose health may be affected by unstable housing, food insecurity or transport barriers. Its approach recognises that a clinical instruction alone may not solve a practical obstacle.
The company deployed a voice AI agent called Flora for intake, referrals and check-ins between clinical visits. Its experience illustrates how automation might extend contact while human professionals remain responsible for complex judgement. The model is not simply replacing a doctor with a chatbot; it is redesigning the support surrounding clinical care.
Why chronic conditions suit continuous support
Many chronic conditions require repeated small actions over months or years. Blood-pressure monitoring, medication routines, nutrition, appointments and mental-health support cannot be addressed fully during an occasional consultation. Patients may need reminders or help when circumstances change.
AI systems can provide consistent availability and handle routine communication at scale. However, frequency is not itself an outcome. A service must show that its contacts improve understanding, access or health measures without overwhelming patients or creating false reassurance.
Social needs remain part of healthcare
The ACCESS model creates room for services that address housing, food and transport alongside medical treatment. Those factors can determine whether a patient fills a prescription, attends an appointment or follows a care plan. Technology that ignores them may produce polished conversations with little practical effect.
Pair Team’s community-integrated approach is notable because it links automation to a wider human support network. An AI agent can identify a need or coordinate a referral, but local organisations and professionals often provide the actual service. Sustainable outcomes depend on that entire chain.
Sensitive data requires strong safeguards
Chronic-care conversations can reveal diagnoses, mental-health concerns, housing conditions and financial insecurity. When voice systems and connected platforms process this information, privacy and security become central to patient trust. A breach can harm people who already face significant vulnerability.
Participants need clear consent, data-minimisation practices, access controls and procedures for correcting records. Patients should understand when they are interacting with automation and how to reach a person. Technical compliance alone is insufficient if the service is confusing or difficult to challenge.
The financial model carries risks
Outcome-based payment can reward effective care, but it can also create pressure to select easier patients or optimise measured indicators at the expense of unmeasured needs. Programme design must account for patient complexity and ensure that organisations do not benefit from avoiding difficult cases.
The history of federal payment experiments also argues for careful evaluation. CMS innovation programmes have not always produced the expected savings. ACCESS should therefore be judged on patient outcomes, total public spending, equity and operational effects rather than adoption of AI alone.
What patients, providers and investors should watch
For patients, the key questions are whether care becomes easier to access and whether human help remains available when needed. Providers should examine workload, escalation quality and evidence of benefit. Investors should resist treating a reimbursement opening as proof that every AI healthcare startup has a viable model.
The ten-year horizon allows for meaningful evaluation, but early results will still matter. ACCESS is important because it aligns payment with continuous, technology-supported care rather than forcing every service into an appointment. Its success will depend on whether better outcomes follow—and whether vulnerable patients are protected throughout the experiment.
Frequently asked questions about Medicare ACCESS
Does ACCESS pay an AI company simply for using AI?
No. The model pays participating organisations for managing qualifying conditions and links full payment to measurable outcomes. AI is one possible operational tool, not the result being purchased by itself.
Which conditions are included?
The programme covers diabetes, hypertension, chronic kidney disease, obesity, depression and anxiety. These conditions often require continuing support between appointments, making them relevant to coordinated-care and remote-engagement models.
Will patients still have access to human clinicians?
Human oversight and escalation remain essential, particularly for complex or urgent needs. Patients should understand how to reach a person and should not treat routine automated contact as a substitute for emergency care.
Source: TechCrunch.
