The Centers for Medicare and Medicaid Innovation has provisionally approved more than 150 companies and providers to participate in the ACCESS model, a Medicare chronic care pilot that ties payment to measurable health outcomes rather than individual technology services.

The scale of participation caught federal officials off guard. CMS administrators acknowledged that application volume exceeded their expectations, a signal that the program’s modest payment rates and participation restrictions did not deter digital health companies from pursuing Medicare contracts. For an agency accustomed to slow uptake on experimental payment models, that response carries weight.

The ACCESS model, short for Achieving Chronic Care Excellence through Support and Services, was announced late last year by the Center for Medicare and Medicaid Innovation (CMMI). It targets a specific cluster of chronic conditions: diabetes, hypertension, high cholesterol, musculoskeletal pain, anxiety, and depression. Participants receive set payment rates tied to measurable health outcomes rather than billing for individual digital services, a structure designed to hold technology companies accountable for the clinical results they deliver rather than simply the tools they deploy. The ACCESS model program page on the CMMI site outlines the outcomes-based payment framework in detail.

The participant roster is notably heterogeneous. Mental health applications with large consumer followings are represented alongside wearable device manufacturers, at least one life sciences company with organizational ties to Google, and startups that help major health systems track and manage heart failure patients remotely. That combination reflects how broadly digital health has expanded beyond simple consumer wellness into clinically adjacent territory where Medicare reimbursement becomes viable.

Most participants hadn’t previously served Medicare patients.

That fact matters from a translational standpoint. Digital health companies typically build products for commercially insured or self-pay populations, where regulatory requirements are lighter and payment pathways are less complex. Medicare beneficiaries are older, carry higher rates of comorbidity, and rely on technology differently than younger commercial populations. A company that has demonstrated efficacy in a 35-year-old with hypertension has not necessarily demonstrated it in a 72-year-old managing hypertension alongside type 2 diabetes and early-stage chronic kidney disease. The physiological and logistical differences between those populations are not trivial.

The neurological dimension of this expansion deserves particular attention. Two of the conditions targeted by the ACCESS model, anxiety and depression, represent the most neuroscience-dense categories in the program’s portfolio. Effective digital interventions for these conditions can’t simply rest on behavioral nudges. They must engage mechanisms that are reasonably well characterized at the circuit level: dysregulation of the prefrontal-limbic axis in depression, hyperactivation of threat-detection networks centered on the amygdala in anxiety disorders, and disruption of monoaminergic signaling across both conditions. Preclinical data on these pathways are extensive. The translational gap, however, remains wide.

Wearable devices offer one partial bridge. Continuous physiological monitoring can capture surrogates for autonomic nervous system (ANS) activity, including heart rate variability (HRV), galvanic skin response, and respiratory rate, each of which correlates with affective state in ways that have been validated against clinical measures in laboratory settings. Whether those correlations hold with sufficient specificity across the heterogeneous Medicare population is a separate and largely unanswered question. The ACCESS model’s outcomes-based payment structure creates a financial incentive to answer it, which is arguably the most clinically valuable feature of the program’s design.

The initial deadline for the first ACCESS cohort was April 1. CMMI extended it.

Federal officials announced Monday that the deadline would be pushed back to allow additional applicants to join the first cohort. That decision reflects an agency calculating that broader participation now produces better evidence later. A larger, more diverse participant pool generates richer comparative data across condition types, patient demographics, and intervention modalities. For a model explicitly designed to establish whether technology-supported chronic care can work at Medicare scale, sample breadth isn’t just administratively convenient. It’s scientifically necessary.

For providers and researchers who work within Hawaii’s health system, the ACCESS model’s structure carries specific relevance. Hawaii’s Medicare population is ethnically diverse and geographically dispersed, with substantial proportions of Native Hawaiian, Filipino, and Japanese-American beneficiaries. Chronic condition prevalence patterns differ across these groups in ways that affect both disease progression and optimal management strategy. Hypertension prevalence, diabetes complication rates, and depression screening outcomes all vary across Hawaii’s patient communities in patterns documented in state-level public health data maintained by the Hawaii Department of Health. If digital health companies entering the ACCESS model are conducting outcomes measurement without adequately accounting for population heterogeneity, their reported results may not generalize well to Hawaii’s Medicare beneficiaries, regardless of how well those tools perform in aggregate national analyses.

That concern isn’t theoretical. It’s the same concern that has complicated the application of clinical trial results to Hawaii populations for decades. Commercial digital health products have been developed overwhelmingly with mainland, primarily white, commercially insured populations as their design base. The physiological and behavioral assumptions embedded in their algorithms reflect that design history. Entering a Medicare program that demands measurable health outcomes will, for the first time, force many of these companies to confront how their tools perform across a genuinely diverse federal payer population. Researchers tracking that performance data will want to request stratified outcomes reporting from CMMI as ACCESS cohort results accumulate.

The heart failure management companies on the ACCESS participant list represent a different end of the clinical acuity spectrum from mental health apps, but the translational questions are equally pointed. Remote management of heart failure relies on physiological signal capture: daily weights, fluid status indicators, natriuretic peptide proxies in some systems, and symptom questionnaire data. The challenge isn’t sensor technology. It’s integration. Heart failure patients in Medicare have frequently experienced the kind of clinical complexity that makes any single-signal monitoring system inadequate. A 68-year-old with heart failure with reduced ejection fraction (HFrEF), chronic obstructive pulmonary disease, and moderate cognitive impairment generates monitoring signals that require clinical judgment to interpret correctly. Whether the startups managing these patients through the ACCESS model have embedded that judgment into their systems, either through clinical decision support algorithms or human review pathways, will determine whether the outcomes payments they receive are earned or incidental.

As reported by STAT News, the list of provisionally approved participants includes popular mental health apps, wearable device makers, a life sciences company tied to Google, and startups specializing in heart failure management, an unusually broad assembly of digital health stakeholders under a single federal payment model. The diversity of participant types will complicate comparison across the cohort but should, if CMMI designs its evaluation methodology rigorously, produce evidence that is more actionable than what any single-category study could generate.

The CMMI Innovation Center operates under a statutory mandate to test payment and service delivery models that reduce expenditures while maintaining or improving quality of care. ACCESS is structured as an alternative payment model (APM), which means participants who qualify may be eligible for incentive payments under the Medicare Access and CHIP Reauthorization Act (MACRA) framework. That eligibility adds a second layer of financial motivation for digital health companies to demonstrate genuine clinical outcomes rather than engagement metrics that don’t translate into health improvements.

Clinicians considering referring patients to ACCESS-participating providers should prioritize asking what outcomes data those providers are collecting, how they’re defining success for the specific conditions their tools address, and whether their measurement methodology accounts for comorbidity burden. The answers won’t always be satisfying. Many of these companies are entering Medicare for the first time and are still learning what rigorous federal outcomes measurement requires. That learning curve is real, and patients are in the system while it’s happening. The clinical judgment required to manage that reality belongs to the physicians who know their patients, not the algorithms that are tracking them.