CMS has proposed removing 638 services from the Inpatient Only list for CY2027, the second year of a three-year phase-out. For utilization management, the consequential detail is not the site-of-service flexibility CMS describes — it is that 42 CFR 422.101(b)(2)(iii) requires Medicare Advantage plans to cover an inpatient admission when the procedure is on the IPO list. Take the code off the list and that pathway closes.
The number that matters in the CY2027 OPPS proposed rule is 638 — the count of services CMS proposes removing from the Inpatient Only list, across eleven clinical families: auditory, digestive, endocrine, female genital, hemic and lymphatic, integumentary, male genital, maternity care and delivery, mediastinum and diaphragm, respiratory, and urinary. This is year two of the three-year phase-out CMS finalized in the CY2026 rule, which began with 285 mostly musculoskeletal procedures effective January 1, 2026.
CMS frames this as site-of-service flexibility. The agency's language is that the phase-out gives beneficiaries "more choices on where to obtain care with the potential for lower out-of-pocket expenses," and allows these services to be paid in the hospital outpatient setting when clinically appropriate. That framing is accurate as far as it goes, and it is also not the part that will change your Monday.
The comment period on CMS-1850-P closed August 31. What follows is no longer a question of what to put on the record. It is a question of what to build before the final rule lands in November.
Under 42 CFR 422.101(b)(2), a Medicare Advantage plan must provide coverage of an inpatient admission in three circumstances. First, when the admitting physician expects the patient to require hospital care crossing two midnights, based on complex medical factors documented in the medical record. Second, when the admitting physician does not expect two midnights but determines, on complex medical factors, that inpatient care is nonetheless necessary - the case-by-case exception. Third, when the inpatient admission is for a surgical procedure specified by Medicare as inpatient only.
That third clause is a bright line, and it is the only one of the three that does not require a documented clinical judgment to survive a plan's utilization review. It is a list lookup. The procedure is on the IPO list, therefore the admission has a categorical coverage basis — one that does not require documented clinical judgment about expected length of stay or medical necessity to survive the lookup itself, though the admission remains subject to other applicable coverage, coding, and administrative requirements.
Remove the code from the list and the lookup returns nothing. The case does not become outpatient — CMS is explicit that removal makes a service payable in the outpatient setting, not required to be furnished there. What removal does is take away the automatic justification and push the entire determination back onto the first two pathways, both of which turn on documentation of complex medical factors and physician expectation. That is precisely the terrain where MA plans have been most willing to substitute their own review for the admitting physician's judgment.
In CY2026 that shift applied to 285 codes, concentrated in orthopedics, where most UM programs already had mature status-determination workflows because total joints had been through this exact transition in 2020 and 2021. The eleven families in the CY2027 proposal are a different problem. Digestive, respiratory, urinary, and maternity care and delivery are not services most utilization review programs have built IPO-transition muscle around, and they are distributed across service lines that in many organizations do not share a physician advisor.
There is an asymmetry here worth stating plainly, because it changes where your denial risk actually sits.
In the CY2021 OPPS final rule, alongside the first elimination attempt, CMS established that procedures removed from the IPO list beginning January 1, 2021 would be exempted from certain medical review activities related to the two-midnight policy. In the CY2026 final rule, CMS continued that exemption for CY2026 and subsequent years, until the Secretary determines that a given service is more commonly furnished in the outpatient setting than the inpatient setting for the Medicare population.
That is a meaningful protection — in Traditional Medicare. It constrains what contractors can review.
The CMS materials reviewed do not clearly establish that the FFS exemption binds an MA plan's own utilization-management or contract-based review. Assume for planning purposes that it does not. Under that assumption, the phase-out produces a widening gap: for the same procedure on the same patient, a Traditional Medicare inpatient claim carries a medical-review exemption, and a Medicare Advantage inpatient authorization carries no exemption and with the categorical IPO coverage route eliminated, leaving the two-midnight benchmark and case-by-case exception as the relevant inpatient routes under the MA regulation. The clinical facts are identical. The defensibility is not.
Note also what CMS said in the 2024 coverage-criteria rulemaking: care spanning two midnights is not presumed appropriate for inpatient level of care under Medicare Advantage the way the two-midnight presumption operates under Part A. MA plans remain free to audit inpatient claims consistent with their provider contracts. The two-midnight benchmark binds them. The two-midnight presumption does not.
Start by pulling the list and intersecting it with your own volumes. The 638 proposed codes are in the rule's addenda; the operative question is not how many codes moved but how many of your inpatient cases in the last twelve months carried one of them as the principal procedure. In most organizations that number is far smaller than 638 suggests, and it clusters in two or three service lines.
For those cases, the documentation burden shifts from what procedure was performed to why this patient required hospital care crossing two midnights. That is a different sentence in a different part of the note, and it is usually written by someone who has never had to write it for this procedure before, because the code has always been inpatient by definition. Surgeons who have spent their careers not documenting expected length of stay for a given operation are unlikely to change that habit because a rule changed in Baltimore — and that is a planning risk worth naming before January 1.
The second thing to check is contractual. Some MA agreements incorporate the IPO list by reference, or incorporate CMS coverage criteria as they exist at the time of service. Others specify the plan's own medical policy. Those two constructions will diverge on January 1 for 638 codes, and the divergence is worth finding before a denial finds it for you.
CMS is expected to issue the final rule later in 2026, effective January 1, 2027 if finalized. Between now and then, the useful exercise is not modeling revenue. It is identifying which of your surgeons will, on January 2, perform a procedure that has been inpatient-by-definition for their entire career and write a note that no longer justifies the admission. How many of them have you talked to?
Discover how Addie helps health systems, post-acute providers, and payers improve throughput, reduce avoidable days, and deliver better transitions of care.
Get Started
