The comment period on the CY 2027 Medicare Physician Fee Schedule proposed rule (CMS-1848-P) closed at the end of Monday, September 14, 2026. For the last nine weeks this newsletter has been counting down to that date. That clock is now finished, and a different one has started — and it is the one that actually costs money.
| Clock | Date | Days from today |
|---|---|---|
| Comment period on CMS-1848-P | Closed September 14, 2026 | — |
| FY 2027 ICD-10-CM effective | October 1, 2026 | 13 |
| CY 2027 PFS final rule expected | Early November 2026 | ~45 |
| CPT 2027, final PFS rates, CMS-0057-F APIs | January 1, 2027 | 105 |
The window between a proposed rule closing and a final rule publishing is the single most wasted period in the practice-management calendar. Nothing is certain, so nothing gets done — and then the final rule drops around November 1 and leaves sixty days to reconfigure charge templates, renegotiate contracts, and retrain front-desk staff over the holidays. The proposals below are not guaranteed. They are, however, the best available forecast, and every one of them can be modeled now against your own claims data at zero risk.
CMS calls it “accounting for overlap between stand-alone E/M visits and global periods.” In operation it is simpler than that: when a separately identifiable office or outpatient E/M visit is furnished on the same day as a procedure with a 0-, 10-, or 90-day global period, the most expensive service is paid at 100% and every other service that day is paid at 50%.
CMS’s rationale is that same-day E/M visits reported with modifier 25 frequently share physician work and practice resources with the procedure performed at the same encounter, and that the current fee schedule therefore pays twice for one set of inputs. The argument is coherent. The financial effect on a practice that routinely sees and treats in one visit is severe, and it is not evenly distributed: dermatology, otolaryngology, podiatry, ophthalmology, orthopedics, and several primary care specialties carry the largest exposure.
This is a modeling exercise you can complete in an afternoon. Pull twelve months of claims. Filter to lines where an office or outpatient E/M was billed with modifier 25 alongside a 0-, 10-, or 90-day global procedure. For each encounter, identify the higher-allowed service and halve everything else. The resulting number is your 2027 exposure under the proposal as written — and it belongs in your budget conversation this month, not after the final rule.
CMS will establish a three-tiered, flat-rate payment system for skin substitute products based on FDA regulatory classification. Medicare currently pays for these products as incident-to supplies at a single 2026 rate of $127.28 per square centimeter across all care settings, with no reimbursement for discarded amounts.
Any wound care, podiatry, vascular, or dermatology practice with a skin substitute line needs to map its current product mix to the three FDA classification tiers before contracting for 2027 inventory. A product that sits in the lowest tier under the new methodology is a fundamentally different economic proposition than it was in 2026, and purchasing commitments signed this quarter will outlive the rate that justified them.
The complexity add-on G2211 — the code CMS created to recognize complex, longitudinal primary care — is proposed to convert from a standalone HCPCS code into a modifier appended to the associated E/M base code, and from a flat payment into a percentage add-on: 16% of the base E/M, rising to 32% for practices participating in a qualifying ACO model.
The doubling for ACO participants is the policy signal. CMS is no longer merely paying for continuity; it is paying twice as much for continuity delivered inside an accountable care arrangement. For an independent primary care practice that has been weighing ACO participation as an administrative burden with a speculative return, the 16-point spread is now a line item that can be multiplied by the practice’s own G2211-eligible visit volume.
The mechanical change matters too. A modifier behaves differently from a code in nearly every practice management system: different edits, different claim-scrubber rules, different denial patterns. A practice appending 16% or 32% of its E/M revenue to a modifier it has never used before should be testing that path against its clearinghouse well before January.
On October 1, 2026, UnitedHealthcare eliminates prior authorization requirements on approximately 1,700 CPT codes — by the insurer’s own accounting, roughly 30% of its total prior authorization requirements. The change applies across UnitedHealthcare commercial, Medicare Advantage, Community Plan (Medicaid), Individual Exchange, and Oxford plans.
This is the largest single prior authorization reduction any national payer has executed, and it is not a pilot. The codes span oncology, cardiology, orthopedic and musculoskeletal procedures, genetic and laboratory testing, chiropractic care, physical, occupational and speech therapy, home health, and durable medical equipment. Named procedures include lesion excisions, fracture treatment, joint injections, arthroscopies, colonoscopies, endoscopies, biopsies, hernia repairs, and soft tissue tumor removals.
Here is where practices will lose money on good news.
A code leaving UnitedHealthcare’s prior authorization list does not mean that code is authorization-free. It means it is authorization-free for the specific UnitedHealthcare plan variants covered by the memo. The same CPT code may still require authorization under a delegated behavioral or musculoskeletal vendor arrangement, under a self-funded employer plan with a carve-out, under a state Medicaid managed care requirement layered on the Community Plan, or under any other payer in your contract portfolio.
The predictable failure mode is a front-desk team that hears “UnitedHealthcare dropped prior auth on arthroscopy” and stops checking arthroscopy authorization entirely — for every payer. The denials from that arrive in November, they are administrative, and they are almost entirely unappealable.
Before October 1, pull the plan-specific code lists from the UnitedHealthcare Prior Authorization and Notification Tool for each of the five plan families you participate in — the lists differ. Update your payer matrix at the code-and-plan level, not the code level. Retrain the front desk on the distinction. Then monitor authorization-related denials weekly through November, segmented by payer, so you can see immediately whether the change is being over-applied.
The 1,700-code cut is the visible part of a broader industry commitment. Cigna expects to standardize more than 70% of its prior authorization volume for medical services by the end of 2026. UnitedHealthcare states it already applies standardized prior authorization criteria to more than half of its reviews. Under the industry pledge signed by UnitedHealthcare, Aetna, Cigna, and Elevance, payers committed to reducing the volume of services requiring authorization, honoring existing approvals for 90 days when patients change plans, and clearly communicating denial rationale and appeal rights — with a common payer-provider framework targeted for January 1, 2027.
That date is not a coincidence. It is the same day the CMS-0057-F Prior Authorization API requirement takes effect, with its 7-calendar-day standard and 72-hour expedited decision clocks. The voluntary pledge and the federal mandate are converging on one deadline 105 days from today.
Not every line moves the same direction. The WISeR model — CMS’s AI- and machine-learning-assisted prior authorization pilot — remains active across Arizona, New Jersey, Ohio, Oklahoma, Texas, and Washington, covering 17 outpatient services including skin and tissue substitutes, electrical nerve stimulators, and knee arthroscopy for osteoarthritis. On July 16, 2026, the Senate voted 46–50 against an effort to overturn it, leaving the model on track through its scheduled end date of December 31, 2031.
Note the overlap: knee arthroscopy is simultaneously losing commercial prior authorization at UnitedHealthcare and carrying an AI-assisted review requirement in traditional Medicare in six states. For an orthopedic practice in Phoenix or Houston, October 1 makes the payer mix — not the procedure — the determinant of whether an authorization is needed.
The AMA’s CPT 2027 release, published this month, added 10 new AI-related codes, bringing the total number of AI codes in CPT to 43. Alongside them, the CPT Editorial Panel refreshed Appendix S — the Taxonomy for Artificial Intelligence in Medical Services and Procedures — effective January 1, 2027, sharpening the boundary between assistive and augmentative services and defining what qualifies as a clinically meaningful output.
The three-tier taxonomy is worth learning precisely, because it is becoming the language payers use to decide what they will pay for:
| Category | Definition | Billing consequence |
|---|---|---|
| Assistive | Identifies clinically relevant data and surfaces it to the physician. No analysis. | Generally not separately reportable |
| Augmentative | Analyzes data to produce a clinically meaningful output the physician can act on | Increasingly the level at which discrete codes exist |
| Autonomous | Interprets data and makes recommendations | Where new code creation is concentrated |
For a practice evaluating an AI vendor, the question to put in writing has changed. It is no longer “is this reimbursable.” It is “which Appendix S category does your product fall into, and which CPT code do you believe supports it.” A vendor that cannot answer that in one sentence is selling an efficiency tool, not a revenue tool — which may still be worth buying, but should be priced as a cost reduction rather than a new revenue line.
Set against a code set that now contains 43 AI codes, the deployment data is sobering. While three out of four healthcare executives report implementing or having used AI somewhere in their organization, 59% say they have not yet implemented AI or automation in the revenue cycle, and just 2% report AI fully or mostly integrated across RCM operations.
Innovaccer’s 2026 State of Revenue Lifecycle in Healthcare report found 63% of organizations running AI in at least one workflow and 52% having expanded implementations across departments, with 45% having established some form of AI governance or ethics structure. The gap between “running AI in at least one workflow” and “fully integrated across RCM” is the entire story of 2026: pilots are nearly universal, production is rare.
Independent corroboration comes from outside healthcare. Gartner expects over 40% of agentic AI projects to be cancelled by the end of 2027, and only one in five organizations has a mature governance model for autonomous agents. The failure mode is consistent across industries — pilots that cannot survive contact with production data.
One category has escaped pilot purgatory. Voice-based ambient documentation was used by 29% of physicians surveyed in January 2026, up from 20% in April 2025. Among Epic hospitals, 62.6% have adopted ambient AI. Roughly one-third of healthcare providers now have access to an ambient scribe, and in May 2026 ambient documentation went generally available for nurses across 250-plus U.S. health systems. Among clinicians who use AI at all, 69% report daily use, including 36% who use it multiple times a day.
Ambient documentation succeeded where denial-management AI has stalled for a structural reason worth internalizing: it sits at the point of data creation, requires no payer cooperation, no integration with an adjudication system, and no governance committee to approve an autonomous financial decision. It improves the note, and a better note is the input to every downstream revenue cycle process. Practices that cannot get an autonomous coding pilot funded should note that the highest-yield AI investment available to them may be upstream of coding entirely.
UF Health announced on September 9 a partnership with R1 to deploy an AI-native revenue cycle operating model on R1’s Phare OS platform, consolidating data management, payer intelligence, and AI-driven automation into a single system. Earlier in 2026, R1 partnered with clinical documentation platform Heidi to embed ambient documentation directly into its revenue operating system — the same upstream-of-coding logic, executed as a corporate strategy.
If three of your vendors have given you three different counts for the October 1 update, none of them is necessarily wrong. They are counting different things, and the difference matters operationally:
| Framing | Count | What it counts |
|---|---|---|
| New billable codes | 190 | Codes you can actually report |
| New entries in the code set | 238 | 190 billable plus 48 non-billable category headers |
| Codes deleted outright | 21 | Removed from the code set entirely |
| Codes no longer reportable | 30 | 21 deletions plus codes demoted to headers and other invalidations |
| Revised code titles | 4 | Descriptor changes only |
The number your charge templates care about is the third and fourth rows, not the first. Deleted and demoted codes deny from day one. A new code you fail to adopt costs you specificity; an invalid code left sitting in a template costs you the claim.
The effective date is governed by date of service or discharge, not claim submission date. Services through September 30 continue to use FY 2026 codes regardless of when you bill them.
The clearest example this year is cardiac. I42.0 (Dilated cardiomyopathy) stops being billable on October 1 and becomes a category header. It splits into:
Any cardiology practice with I42.0 hard-coded in a problem list, a favorites list, or a charge template is generating invalid claims on October 1 unless it is remapped. The same pattern applies to BMI: Z68.1 is deleted and replaced by Z68.18 (BMI 18.4 or less, adult) and Z68.19 (BMI 18.5–19.9, adult) — a change that touches every primary care practice reporting low-BMI diagnoses for nutritional counseling or HCC capture. Among deletions, the S23.42 sternoclavicular sprain series is the most notable removal for orthopedics and urgent care.
Other areas of concentration: expanded specificity for secondary neoplasm sites, a revised exclusion note for breast carcinoma in situ, additional detail in cardiomyopathy and rare cardiac conditions, musculoskeletal specificity, and new codes for plantar fasciitis and sinusitis.
CMS publishes a 2027 Conversion Table identifying every code inactive as of 10/1/2026 alongside its replacement. That file is the single most efficient remediation tool available, and it is free. Run your active diagnosis list against it, not against the new-code list.
The AMA released the CPT 2027 code set this month: 453 editorial changes — 299 new codes, 74 revisions, and 80 deletions — effective January 1, 2027.
The 80 deletions deserve more attention than the 299 additions, because a deleted CPT code is a hard claim rejection and, where it is a bundled code, a contract renegotiation.
Substantive changes by area:
The CAQH Index data — drawn from 600 organizations covering 63% of insured lives — quantifies more than $20 billion in additional annual savings available from shifting remaining manual transactions to fully electronic workflows. Claim status inquiry has reached 81% electronic adoption and claim payment 78%. Prior authorization sits at 40% — and the more revealing figure is that only 35% of medical prior authorizations are conducted fully electronically using the X12 278 transaction.
That gap between “electronic” and “fully electronic X12 278” is where most practices actually live. A payer portal is electronic in the sense that it is not a fax. It is not electronic in the sense that matters: it does not integrate, it does not populate from your practice management system, it cannot be worked in a queue, and it consumes a staff member’s full attention for its duration. Portal work is manual work with a browser.
In 2026, A/R days trended up by 2.4 days at the industry median, driven by Medicare Advantage prior authorization backlogs and pended claims awaiting medical record review. This is the operational counterweight to the prior authorization good news in Section 02 — the volume of authorizations is starting to fall, but the ones that remain are taking longer and dragging cash with them.
Denials remain elevated: 41% of providers now report a claim denial rate above 10%, against an HFMA top-quartile target of under 5% and a published industry average of 9–12%.
| Metric | Industry average | Target | Top quartile |
|---|---|---|---|
| Days in A/R | 40–45 | < 35 | < 30 |
| A/R > 90 days | 18–22% | < 15% | < 10% |
| First-pass clean claim rate | 90–93% | ≥ 95% | 97%+ |
| First-pass denial rate | 9–12% | < 8% | < 5% |
| Denial overturn rate | 40–55% | > 65% | > 85% |
| Net collection rate | 93–95% | ≥ 96% | 98%+ |
| Cost to collect | 3–5% of net revenue | < 3.5% | < 2.5% |
| Electronic prior auth (X12 278) | 35% | ≥ 75% | 90%+ |
| Pre-service patient collection share | 21% | ≥ 30% | 40%+ |
| Patient bad debt | 3–5% of net patient revenue | < 3% | < 2% |
Note the distinction between the third and fourth rows, because practices routinely conflate them. Clean claim rate measures whether a claim passes your edits before submission. First-pass yield measures whether it gets paid on first submission. A claim can be perfectly clean and still deny for eligibility, medical necessity, or a payer-specific rule your scrubber does not know about. A high clean claim rate sitting on top of a low first-pass yield is a diagnostic finding, not a contradiction: it means your edits are checking format and your payers are denying on substance.
Move eligibility verification to scheduling. Practices that run real-time eligibility at scheduling rather than at check-in reduce eligibility-related denials by over 40%. The mechanism is lead time, not technology: verifying 24 to 48 hours ahead creates room to fix a coverage problem before the visit instead of after the claim. Practices on AI-integrated billing platforms report up to a 30% increase in first-pass acceptance versus manual scrubbing.
Collect before the visit. Health systems reported collecting 31% of total patient billings in 2026, up from 24% in 2025, with pre-service collections rising from 16% to 21% of self-pay collections. 91.5% of surveyed organizations now encourage payment, require payment, or capture a payment method on file during the estimate process, up from 81.3% a year earlier. The urgency is arithmetic: patient responsibility is now 30–35% of total revenue, a balance over 90 days has under a 50% chance of collection, and past 120 days that falls to 20% or less.
KLAS Research published “Revenue Cycle Management Suites 2026: Partnership Is Key in Making the Suite Experience a Success” in July, evaluating how deep adopters — organizations running three or more distinct RCM modules from a single vendor — actually experience their vendors. Waystar led on overall satisfaction. Across the broader 2026 Best in KLAS awards, Abridge, Microsoft, Waystar, and FinThrive took honors across nine revenue cycle technology categories.
The most useful finding in the KLAS report is not the ranking. It is the diagnosis: many RCM suites were assembled rapidly through acquisition, leaving customers with inconsistent product integration, uneven software maturity, and variable support quality across modules sold as one platform. The report’s framing — that partnership, not feature count, determines whether a suite succeeds — is a direct consequence of that assembly history.
For a practice evaluating a suite, that converts into a specific diligence question: for each module you are buying, when did the vendor acquire it, and what is the current state of its integration with the core platform? A five-module suite where three modules were acquired in the last thirty-six months is three separate implementations wearing one contract.
The healthcare RCM market is projected to grow from $189.28 billion in 2026 to $505.8 billion by 2035, an 11.54% CAGR. The outsourcing segment specifically was valued at $61.7 billion in 2025 and is expected to reach $126.5 billion by 2032, a 10.8% CAGR — and 70% of hospitals and health systems plan to expand their RCM outsourcing engagements. Industry estimates put the potential annual savings from AI and automation in the revenue cycle at up to $360 billion.
Treat the $360 billion figure as a market-sizing number, not a budget input. The operational figures in Section 05 — a 40% reduction in eligibility denials from moving verification to scheduling, a 30% lift in first-pass acceptance — are the ones you can actually underwrite.
Becker’s 11th Annual Health IT + Digital Health + RCM Conference ran September 14–17 in Chicago, closing yesterday, with 560-plus hospital and health system speakers and over 90% of sessions led by provider-side leaders. The agenda’s center of gravity is itself a data point: sessions titled “From Pilot to Production: Real-World Lessons on Scaling AI in Revenue Cycle” and “From Reactive to Proactive: How AI Is Redefining Revenue Integrity” reflect an industry that has stopped asking whether to adopt AI and started asking why its pilots will not scale — the same question the 2% integration figure in Section 03 answers from the other direction.
CMS-0057-F requires impacted payers to operate four FHIR APIs by January 1, 2027: Patient Access, Provider Access, Payer-to-Payer, and Prior Authorization. The Prior Authorization API standardizes electronic submission and response, replacing fax, portal, and phone, with standard decisions required within 7 calendar days and expedited decisions within 72 hours.
The technical stack is HL7 FHIR R4, OAuth 2.0 / OpenID Connect, the Da Vinci implementation guides (CRD, DTR, PAS), and SMART on FHIR for EHR-embedded workflows. The rule binds Medicare Advantage organizations, Medicaid and CHIP managed care plans, state Medicaid and CHIP fee-for-service programs, and QHP issuers on the federally facilitated exchanges. Note the staggered compliance dates: MA, Medicaid and CHIP managed care, and QHP issuers were obligated as of January 1, 2026; state Medicaid and CHIP fee-for-service programs have until January 1, 2027.
Your practice has no obligation under this rule, and therefore no deadline — which is exactly why it gets missed. The mandate lands on your payers. The benefit only reaches you if your EHR or practice management system can consume the Prior Authorization API. If it cannot, your payers will have eliminated the fax while you continue to use the portal, and you will have paid for none of the savings and received none of them either.
Combine this with Section 02: UnitedHealthcare’s voluntary 1,700-code reduction on October 1 and the industry pledge’s January 1, 2027 framework target converge on the same date as the federal API mandate. Prior authorization in 2027 will involve fewer transactions, moving faster, over a standardized pipe. A practice still working authorizations through browser tabs in February will be the slowest party in its own revenue cycle.
Enforcement of the updated Hospital Price Transparency requirements finalized in the CY 2026 OPPS/ASC final rule began April 1, 2026 — roughly five and a half months ago. The regulations became effective January 1, 2026, with CMS delaying enforcement one quarter to allow hospitals to update machine-readable files and related compliance processes.
The updates represent the most comprehensive change to price transparency since the rule’s inception, moving it from a basic disclosure obligation toward a standardized, data-driven pricing framework. The direct obligation sits with hospitals, but physician practices with any hospital-affiliated, provider-based, or joint-venture ASC arrangement should confirm which entity owns the machine-readable file for services rendered under that arrangement — and that the rates published in it match the rates in your own contracts. A mismatch between a published MRF and a negotiated rate is a discoverable fact.
| Item | Date | Who it binds |
|---|---|---|
| FY 2027 ICD-10-CM effective | October 1, 2026 | All providers |
| UnitedHealthcare prior auth reductions | October 1, 2026 | UHC-contracted practices |
| CY 2027 PFS final rule expected | Early November 2026 | All Medicare providers |
| CPT 2027 and Appendix S effective | January 1, 2027 | All providers |
| CMS-0057-F four FHIR APIs | January 1, 2027 | Impacted payers |
| Industry prior auth framework target | January 1, 2027 | Pledge signatory payers |
| WISeR model runs through | December 31, 2031 | AZ, NJ, OH, OK, TX, WA |
The structural arithmetic underneath every consolidation headline has not changed, and this year it is unusually clean. Between 2001 and 2026, Medicare physician payment rose approximately 10% while the cost of running a medical practice rose approximately 63%, according to AMA analysis. Adjusted for practice cost inflation, Medicare physician payment has declined roughly 33% since 2001.
In athenahealth’s 2026 Physician Sentiment Survey, 89% said staying independent has become harder — including 88% of physicians in practices with five or fewer physicians. A separate survey found 80% of independent primary care physicians worried about money, yet only 2% considering leaving independence. Those two findings are not in tension; they describe the same population under strain and unwilling to sell. The financial pressure is nearly universal. The willingness to consolidate is not.
The modifier 25 proposal is regressive by structure. A large multispecialty group can absorb a 50% haircut on same-day secondary services across a diversified service mix. A three-physician dermatology or podiatry practice — where see-and-treat in one visit is the business model — cannot diversify away from it. The specialties CMS’s own impact analysis shows taking the largest reductions are disproportionately small-practice specialties.
The UnitedHealthcare prior auth reduction is progressive, and only if you can operationalize it. The 1,700-code change is worth real staff hours to a small practice — proportionally more than to a health system, because authorization work is a larger share of a small practice’s total administrative load. But capturing it requires pulling five plan-specific code lists, updating a payer matrix at the code-and-plan level, and monitoring denials weekly. A health system has a payer-relations department that does this as routine work. A five-physician practice has an office manager who also handles payroll. The benefit is larger and the capacity to capture it is smaller.
The CMS-0057-F API mandate has the same shape. A large group can require its EHR vendor to deliver FHIR integration on a contractual timeline. An independent practice on a small-vendor platform can ask, and wait.
The G2211 ACO doubling is the one clear opportunity. A 16% add-on becoming 32% inside a qualifying ACO is the largest single upside available to independent primary care in the 2027 proposal — and it is the one lever a small practice can actually pull without vendor cooperation, capital, or payer consent.
| Metric | Where most independents are | Where you should be |
|---|---|---|
| Days in A/R | 42–50 | < 35 |
| First-pass clean claim rate | 88–92% | ≥ 95% |
| Denial rate | 10–14% | < 8% |
| Denial overturn rate | 35–50% | > 65% |
| Electronic prior auth (X12 278) | < 35% | ≥ 75% |
| Pre-service collection share | 12–18% | ≥ 30% |
| Cost to collect | 5–8% of net revenue | < 4% |
| Administrative FTEs per physician | 2.5–3.0 | < 2.0 |
| Payer matrix review cadence | Annual or ad hoc | Monthly |
What to prioritize next, in order: (1) run the modifier 25 exposure model against twelve months of your own claims — it is free and it is the largest 2027 variable; (2) remediate deleted and demoted ICD-10 codes before October 1; (3) build the UnitedHealthcare code-and-plan matrix; (4) if you are primary care and not in an ACO, price the 32% G2211 add-on against your eligible visit volume.
Two things happen to most specialties in the next 105 days, and they point in opposite directions: a proposed CY 2027 payment impact, and a prior authorization reduction on October 1. Read them together — an authorization burden lifted does not offset a rate cut, but it does change where your staff hours go.
| Specialty | CY 2027 proposed impact | What else changes | Priority action |
|---|---|---|---|
| Primary Care | Net positive via G2211 | G2211 converts to a modifier paying 16% of base E/M, 32% in a qualifying ACO; Z68.1 deleted → Z68.18 / Z68.19 for low BMI | Price the 32% ACO add-on against eligible visit volume; remap BMI codes before October 1 |
| Cardiology | ~+1% overall | I42.0 becomes non-billable → I42.00 / I42.01 / I42.09; UHC drops PA on cardiology codes Oct 1; 3 new LVAD codes in CPT 2027 | Remap I42.0 everywhere it is hard-coded; pull the UHC cardiology code list |
| Orthopedics | −7% (hand surgery −5%) | UHC drops PA on arthroscopies, fracture care, joint injections, soft tissue tumor removal; S23.42 sternoclavicular sprain series deleted; knee arthroscopy still under WISeR in 6 states | Model modifier 25 exposure on same-day injection + E/M; separate UHC from WISeR logic by payer |
| Oncology | Hem/onc −1.5% | Radiation oncology +1.5% to +3% depending on which read of the impact table you use; UHC drops PA on oncology codes Oct 1; new ICD-10 specificity for secondary neoplasm sites | Update secondary neoplasm coding; verify which oncology codes left the UHC list |
| Dermatology | −9% — among the largest proposed reductions | Highest modifier 25 exposure of any specialty; skin substitutes move to 3 flat FDA-based tiers | Run the modifier 25 model first; re-tier the skin substitute product mix before 2027 purchasing |
| Otolaryngology | −9% | Significant modifier 25 exposure; new ICD-10 sinusitis codes effective October 1 | Model same-day procedure + E/M encounters; update sinusitis coding |
| Radiology | Interventional radiology +3% | CPT 2027 adds radiology codes and a new head and neck MRA table; UHC drops PA on imaging codes Oct 1 | Load the new MRA table; confirm which imaging codes left the UHC PA list |
| Gastroenterology | Carries the 2.5% efficiency adjustment to work RVUs on endoscopy and other non-time-based codes | UHC drops PA on colonoscopy and endoscopy Oct 1 | Recalculate endoscopy margin post-adjustment; update scheduling scripts referencing authorization |
| Neurology | Smaller proposed decrease | Denial rates average 18% — roughly double the 9–12% industry average | Segment the 18% by root cause before buying any denial tool |
| Mental / Behavioral Health | Clinical psychologists and social workers among the largest proposed increases | Digital prescription therapeutics billable incident-to; in-person requirements for tele-mental health enforced since January 31, 2026 | Verify in-person compliance documentation; evaluate digital therapeutic billing |
| Physical / Occupational Therapy | Smaller proposed increases | UHC drops PA on PT, OT, and speech therapy codes Oct 1 | Re-scope authorization staffing after October 1 |
| Obstetrics | Framework change rather than rate change | CPT 2027 replaces the global maternity bundle with a team-based framework | Reprice every global maternity contract before January 1 |
The single cross-cutting observation: the specialties facing the deepest proposed cuts — dermatology at −9%, ENT at −9%, orthopedics at −7% — are also the specialties with the heaviest modifier 25 exposure. The impact tables and the same-day E/M proposal are not independent risks. They compound, and CMS’s published impact percentage may understate the effect on a practice whose visit pattern is procedure-plus-E/M by default.
Ten moves for the week of September 18, 2026 — each tied directly to something in this issue.