Centers for Medicare & Medicaid Services to Launch Landmark ACCESS Program

There’s an audacious new player in the decades-long push toward value-based care in the United States. Launching July 5, the Advancing Chronic Care with Effective, Scalable Solutions (ACCESS) model is a 10-year payment experiment that could unite traditional provider organizations and digital health disrupters under outcome-based reimbursement.

This new Medicare model pays providers for managing beneficiaries’ qualifying conditions with technology-enabled care, and half of that payment depends on providers’ ability to improve patient outcomes at scale. Some 150 providers have already been approved for participation, and they run the gamut of unlikely actors: private technology startups like wearables maker WHOOP, digital coaching apps like Lark, and virtual care companies like Pair Team and WellDoc that combine clinician- and AI-driven interventions.

Policy experts and clinicians say the program is a bold departure from historic initiatives helmed by the Centers for Medicare & Medicaid Services’s (CMS) innovation hub (CMMI).

“It’s not what people may have expected. It really is a lot of tech companies,” says Robert Longyear, CEO of health care policy consultancy Longyear Health in Washington, DC, and cofounder of a remote physiological monitoring company similar to ones participating in ACCESS.

Despite the promise presented by these innovative partnerships, the program’s novelty and ambition raise questions around execution. Reimbursement rates are lower than expected, which could be a sticking point for participating organizations with more expensive products, such as hardware, Longyear says.

‎

Additionally, the rapid uptake of new clinical offerings from nontraditional providers could impact patient safety, says Robert Trestman, MD, PhD, chair of the American Psychiatric Association’s (APA) Council on Healthcare Systems and Financing, which provided input on measures for the behavioral health conditions included in the program’s initial clinical tracks.

“It is a time of experimentation, which is great,” says Trestman, who’s also chair of the Department of Psychiatry and Behavioral Medicine at Carilion Clinic and Virginia Tech Carilion School of Medicine in Roanoke, Virginia. “But this project skips over the validation stage and it makes the patients beta testers—typically unwitting beta testers—for as yet unproven [technology].”

ACCESS Overview

ACCESS is a voluntary model available to Medicare Part B beneficiaries nationwide. The four initial clinical tracks reflect some of the most common—and highest-cost—chronic conditions covered, including hypertension, diabetes, cardiovascular disease, chronic musculoskeletal pain, depression, and anxiety.

According to CMS, each track groups related conditions that are often treated similarly, and participating organizations must manage all qualifying conditions in the track(s) of which they’re part.

CMS plans to issue monthly payments but withhold half of each for reconciliation with clinical outcomes at the end of the 12-month care period. To earn the full payment, at least half of an organization’s aligned beneficiaries must meet all required clinical targets.

ACCESS payment levels are modest compared to existing digital health payment determinants, such as remote patient monitoring (RPM) and chronic care management (CCM) codes. The lower-than-expected rates could force innovation—or “inhibit profitability,” Longyear says, especially when it comes to makers of higher-cost tech, like wearables and other hardware. “I think we’re going to see a lot of exits from the program.”

The Promise

Created as part of the Affordable Care Act, CMMI is charged with delivering better health outcomes at lower cost by shifting payment incentives from the volume of services provided (fee-for-service) to the outcomes patients experience (value-based care).

In its 16 years of operation, CMMI has launched dozens of models with mixed results. A 2023 Congressional Budget Office analysis, for example, concluded that the center had increased federal spending by more than $5 billion over its first decade rather than producing projected savings. Even so, certain models, like accountable care organizations (ACO), “have demonstrated measurable, if modest, capacity to bend the cost curve,” wrote Longyear in a recent article. “The overarching lesson from a decade of ACO evaluations is that meaningful spending reductions are achievable within fee-for-service Medicare, but they require time, organizational learning, appropriate risk calibration, and payment structures that genuinely alter care delivery incentives.”

ACCESS could help build on this lesson through its blending of a capitated rate with payments linked to specific outcomes, Longyear says. He calls that level of granularity a departure from the ACO’s “broad bucket of savings” and complex payment layers, meaning it could offer clearer insight into which interventions move the needle.

‎

Experts also hope that technology-enabled interventions, coupled with stronger protections and human guidance, might help close critical care gaps.

“The treatment of choice for mild to moderate depression and anxiety is psychotherapy. It’s just that there aren’t enough available psychotherapists to do it,” says Trestman. Rather than using medication as a first-line treatment due to resource shortages, clinicians could leverage tools that help patients practice gold-standard modalities like cognitive behavioral therapy, making “the need for psychotherapy more manageable with fewer sessions.”

If ACCESS proves effective, Longyear predicts it’ll be useful across a much wider array of conditions and outcomes. “This is a real attempt at trying to do something clinical and that’s actually scalable,” he says.

A program aspect that could advance this goal is CMS’s plan to randomly assign about 10% of year-one participating beneficiaries to a control group to help sharpen the evaluation of the program’s digital health interventions. “It’s a pretty innovative policy tool,” Longyear says.

The Pitfalls

Though both Longyear and Trestman are impressed by the ACCESS model’s use of a randomized control group, they also point to some inherent weaknesses.

The incorporation of so many different technologies may make it difficult to amass “enough data to drill down for any one of them to determine which ones may be a benefit,” Trestman notes.

There’s also the potential for selection bias. Because the program’s patient participants, like their provider counterparts, are self-selecting, they may already be more comfortable using a health app or more active with their overall health, Longyear says. Additionally, some of the initially covered conditions, such as depression, have the potential to resolve spontaneously for a significant portion of patients within the program’s treatment timeline, Trestman says. “So that is going to be a complicating factor to determine whether...people will be benefiting from this or what will be seen as a massive placebo effect.”

Privacy is another concern.

“As we’ve seen recently, there are profound data breach risks even in well-developed, sophisticated, and protected health information systems,” Trestman says, citing the 2024 Optum cyberattack. “When you have all of these relatively small startup companies, the vendor risk assessments may not be adequate to defend against a hacking event, so this may expose people’s PHI [personal health information] to further risks.”

Aside from the data ambiguities, the model’s emphasis on bringing new technologies to market at scale and speed has yielded what Trestman calls a “very unusual” collaboration between CMS and the US Food and Drug Administration, with the latter loosening requirements around health tech authorization at a time AI disruption is outpacing regulation.

There’s a real risk of “causing harm” unless the tools are “carefully monitored and modified appropriately over time,” Trestman says.

He’s especially worried about the potential for increased errors in diagnosing and treating patients, especially in cases where care is rendered asynchronously. “Most of these companies are not big enough to have that level of investment of oversight.”

Trestman and Longyear are also concerned about the program’s implications for care coordination given the likelihood that it’ll introduce new portals for patients and primary care providers to juggle. “Even today, it’s extremely difficult for us to share information from one health system to another,” Trestman says. “If you’re having innovative technologies, are they really going to be integrated effectively?”

Launch Signals to Watch

The most obvious sign that ACCESS is heading in the right direction will be participants hitting their outcome-aligned payment thresholds, Longyear says, though he notes that selection bias could be a complicating factor.

He’s also keeping a close eye on attraction and retention. “Can you get patients in the door, and can you keep them?”

Health app engagement rates among similar populations have historically been low but intervenable. “The usability really matters,” Trestman says. Part of that, he says, comes down to designing and refining technologies along dimensions like culture and language.

Regardless of how successful the program is, it will stop short of addressing the social determinants of health at the root of poor outcomes and inefficient care, Trestman says. “Unless people have safe places to live, access to good food, a meaningful job that pays the bills, a safe environment in which to live, then we’re fixing something after it breaks rather than preventing it from getting broken in the first place.”