Medicare's mandatory specialty model starts Jan. 1 and small practices cannot opt out

On Jan. 1, Medicare starts grading a group of specialists on cost and quality whether they signed up or not, and the score will eventually move every Part B claim they file by as much as 9 percent in either direction. There is no opt-out, no hardship exemption and no small-practice carve-out.

The Ambulatory Specialty Model, finalized in the 2026 Medicare Physician Fee Schedule and run out of the CMS Innovation Center, holds outpatient specialists financially accountable for managing heart failure and low back pain. Five performance years run from Jan. 1, 2027, through Dec. 31, 2031, and the corresponding payment years run from 2029 through 2033. CMS says the payment adjustment will range from minus 9 percent to plus 9 percent in the first two payment years and grow in later years.

What the Ambulatory Specialty Model is testing

CMS picked these two conditions because they are expensive and specialist-driven. The agency puts annual Original Medicare spending on heart failure at $10 billion to $13 billion and on low back pain at $6 billion to $8 billion, and says fragmented care between specialists and primary care leads to delayed diagnoses, avoidable hospitalizations and low-value procedures. The model is a test of whether making specialists carry financial risk for the longitudinal management of a chronic condition changes any of that.

Participation is mandatory by design. CMS argues in its model FAQ that voluntary models suffer from selection bias and provider attrition, and that requiring everyone in a market to participate produces findings that generalize. The model's regulations sit at 42 CFR part 512, subpart G.

Which practices are in, and how to check

Two cohorts are in scope. The heart failure cohort covers cardiologists. The low back pain cohort covers anesthesiology, pain management, interventional pain management, neurosurgery, orthopedic surgery and physical medicine and rehabilitation. Non-physician providers are not eligible at this point.

A physician is pulled in by claims history, not by choice: at least 20 attributed episodes a year under the episode-based cost measure methodology, a specialty type set by the code on the majority of the physician's Part B claims, and a practice location in one of the markets CMS selected, covering roughly a quarter of core-based statistical areas and metropolitan divisions. CMS uses data from two years before the performance year, so 2025 claims determine who is in for 2027. Participants are identified by tax identification number and national provider identifier, which means a two-physician group in a selected market is handled the same way a hospital-employed specialist is.

Practices can check their exposure now. CMS publishes an ASM participant dataset searchable by name and NPI, along with a spreadsheet of the mandatory geographic areas and a participant readiness roadmap. Eligibility is reassessed annually, so a group that misses the episode threshold this year can be pulled in later, and a physician who drops below it gets notified and returns to MIPS reporting.

How ASM scoring works

Scoring uses four categories borrowed from the MIPS Value Pathways framework: quality, cost, improvement activities and interoperability. Quality and cost are assessed at the individual physician level. Improvement activities and Promoting Interoperability are assessed at the group level. Physicians are ranked against peers treating the same condition rather than against a fixed benchmark, which means a practice can hit its numbers and still lose money if the field moves faster.

Reporting runs through the same Quality Payment Program portal practices already use, with data and attestations due March 31 after the close of each performance year. Participants are exempt from MIPS for the years they are in the model and cannot earn a MIPS adjustment for those years. One trap worth flagging to physicians: anyone who is MIPS eligible for 2026 still has to report MIPS for 2026, even if their name already shows up on the preliminary ASM participant list.

Why the model is harder on small and independent groups

In its summary of the 2027 proposed rule, the American Medical Association said the model's redistribution percentage of 85 percent guarantees that most participating physicians take a payment cut no matter how well they score, and that downside risk arrives at the full 9 percent in year one rather than phasing in. The AMA also flagged two proposed scoring bonuses, five points for rural physicians and five points for voluntarily submitting patient-reported outcome and risk adjustment data, warning that better-resourced practices are the ones positioned to collect them. Claims-based quality measures would be calculated at the individual level even when a small practice reports as a group, which the AMA said could penalize physicians who take the most complex patients.

What the 2027 proposed rule would change

The proposed rule offers some relief on the operational side. A specialty group would be allowed a single collaborative care agreement with a primary care practice covering all of its specialists instead of one agreement per physician. Cardiologists could be excused if they notify CMS that their primary specialty is a subspecialty such as electrophysiology or interventional cardiology. The separate security risk analysis attestation would come off the interoperability requirements, and the electronic prior authorization measure would be optional and unscored for 2027. The model itself stays mandatory.

What practices should do before Jan. 1

For an administrator whose physicians are on the list, the work between now and January is concrete: paper the collaborative care agreements with the primary care practices that send the referrals, confirm the EHR is certified and actually exchanging data with those practices, and pull the group's existing MIPS episode-based cost measure feedback to see where the practice already stands. The episode data that will drive the 2027 score is built on the same methodology practices can look at today.

Comments on the proposed refinements are due Sept. 14.