The last open rulemaking clock of the 2026 cycle expires on Monday, September 14, 2026. Comments on the CY 2027 Medicare Physician Fee Schedule proposed rule (CMS-1848-P), published in the Federal Register on July 16 after CMS released it July 14, close at the end of that day. From this Friday, that is three days.
Most practices that intended to comment have been drafting against the headline numbers. Those numbers are real:
| Measure | CY 2026 | CY 2027 proposed | Change |
|---|---|---|---|
| Conversion factor — qualifying APM participants | $33.57 | $33.17 | −$0.40 / −1.19% |
| Conversion factor — non-qualifying APM participants | $33.40 | $32.84 | −$0.56 / −1.68% |
| Work RVU budget-neutrality adjustment | — | +0.53% | Offsets proposed work RVU changes |
Two conversion factors, both down, in the second year of the statutory split between APM and non-APM participants. A practice that has not yet decided whether an APM pathway is worth the administrative overhead is now looking at a 33-cent per-RVU spread — small in isolation, meaningful at volume, and permanent in direction.
The conversion factor is the number that gets the press release. The proposal that will change more practice economics over the next five years is the practice expense methodology rewrite, and it has drawn a fraction of the attention.
CMS proposes to phase out specialty-specific practice-expense-per-hour data anchored to surveys from 2007 or earlier. That data is the foundation of the indirect practice expense allocation for a large share of specialties — it is, in effect, a twenty-year-old snapshot of what it costs to run an office, still being used to price what it costs to run an office in 2027. Replacing it is defensible. Replacing it reallocates, because practice expense RVUs are budget-neutral: every specialty that gains does so at the expense of one that loses, and the specialties still priced on the oldest survey data are the ones whose numbers move most.
Two related proposals sit alongside it:
Four proposals will land directly in front-line billing workflow if finalized:
The remote therapy monitoring proposal would pay only for services performed by employed clinical staff, not contractors. If your RTM or remote monitoring program runs through a vendor staffing arrangement — and many independent practice programs do — it stops being payable on January 1, 2027 as proposed. This is a contract problem with a three-day comment window and a 112-day remediation window.
Amid the deadlines, one item is settled. The Consolidated Appropriations Act, 2026, signed February 3 after the 43-day partial government shutdown, extended Medicare telehealth flexibilities and the work GPCI floor through December 31, 2027. CMS instructed that claims be paid as they had been before October 1, 2025, and that the flexibilities apply retroactively across the lapse period.
That retroactivity has an unclaimed-revenue tail that most practices have not worked. CMS explicitly asked practitioners to identify beneficiaries who paid out of pocket for telehealth furnished during the lapse, refund those payments, and submit the claims to Medicare instead. For a practice that ran cash-pay telehealth during the shutdown window, that is a defined, finite list of encounters that can still be converted to Medicare revenue.
Prior authorization policy is moving on three tracks at once this month, and they point in opposite directions.
Texas — the state that invented provider gold carding with HB 3459 in 2021 — strengthened its law through HB 3812, effective September 1, 2026. The change is technical and consequential: exemption status is now evaluated against a full year of prior authorization requests rather than six months.
Read that from the practice side. A longer evaluation window means a larger denominator, which means one bad quarter no longer disqualifies a practice from exemption — and equally, that a practice cannot earn exemption on a favorable six-month stretch. Gold carding just became a sustained-performance program instead of a streak. As of 2026, five states have enacted gold carding: Texas, Louisiana, Michigan, Vermont, and West Virginia.
Medicare is testing the same idea. CMMI signalled it intends to pilot a gold-carding feature within the WISeR model by mid-2026, exempting clinicians with consistent approval histories from future prior authorization or pre-payment review. WISeR is running in six states — New Jersey, Ohio, Oklahoma, Texas, Arizona, and Washington — with coverage determinations due within 72 hours, or 48 for expedited cases, and with contractors explicitly not incentivized to deny.
Prior authorization denials are up 31% year over year across commercial and Medicare Advantage payers. That is the number that makes the gold-carding expansion feel smaller than it sounds: an exemption program covering a subset of services for a subset of providers in five states does not offset a nationwide 31% increase in the base rate.
Twenty-nine health systems are dropping Medicare Advantage plans in 2026, and the reasons given are consistent across all of them: prior authorization friction, claim denials, and slow or low reimbursement.
| Health system | Action |
|---|---|
| Lee Health (Fort Myers, FL) | Out of network with UnitedHealthcare MA plans at the end of 2026 |
| Ohio State University Wexner Medical Center | The James Cancer Center and OSU-affiliated physicians go out of network with Humana MA on October 1 |
| Genesis Healthcare System | Notified Aetna, UnitedHealthcare, Medical Mutual and CareSource that MA contracts terminate at the end of 2026 absent “fair new agreements” |
The published count almost certainly understates the trend — smaller and less-publicized non-renewals never make the list. And MA now covers more than half of eligible Medicare beneficiaries, which means each exit is a larger share of a practice’s panel than it would have been five years ago.
Under CMS-0057-F, impacted payers — Medicare Advantage organizations, Medicaid and CHIP managed care entities, state Medicaid agencies, and QHP issuers on the federally facilitated exchanges — must stand up four FHIR APIs by January 1, 2027: Patient Access, Provider Access, Payer-to-Payer, and Prior Authorization.
The decision-timeframe half of that rule has already been in force since January 1, 2026: 72 hours expedited, 7 days standard, with specific denial reasons required. The first public prior authorization metrics reports, covering calendar year 2025, were due March 31, 2026. Separately, Cigna has committed to standardizing more than 70% of its prior authorization volume for medical services by the end of 2026, part of the industry-wide commitment UnitedHealthcare, Aetna, Elevance and Cigna signed onto with a framework target of January 1, 2027.
Build your gold-card dossier now, whether or not your state has a program. Pull twelve months of prior authorization requests by payer and by service line, and calculate your approval rate for each. Texas has just made twelve months the standard of proof, WISeR is piloting the same concept in Medicare, and every payer running a voluntary exemption program asks for the same data. A practice that cannot produce that table on request is not eligible for any of it — and the table is also the single best input to a contract negotiation.
Three product announcements landed in the last week, and read together they describe a single shift: RCM AI is moving from point tools that do one task to models that hold the whole clinical-to-cash path in one representation.
Maverick Medical AI released its Clinical-to-Revenue Foundation Model on September 8, 2026, positioning it as a single intelligence layer that begins at the point of care and runs through autonomous, billing-ready coding — spanning physician documentation, coding, compliance, medical necessity, and denial prevention. The framing the company chose is “revenue certainty,” which is a deliberate step away from the accuracy-percentage marketing that has defined autonomous coding for three years.
Xsolis announced general availability of its GenAI Peer-to-Peer Clinical Synopsis. This one deserves more attention than it will get. The peer-to-peer review is the last fully manual, physician-time-consuming step in the denial workflow — a scheduled phone call in which a treating physician argues medical necessity to a payer medical director, prepared for by hand, at a cost measured in clinical hours rather than staff hours. Automating the synopsis a physician walks into that call with is a narrow intervention with an unusually favorable ratio: it saves the most expensive minutes in the revenue cycle. In KLAS survey data, 89% of Xsolis users say they rely on its AI to minimize preventable denials.
Abridge announced a partnership with NEJM and JAMA to surface evidence-based guidance inside ambient clinical conversations, explicitly to ease downstream coding burden. That is documentation-side AI acknowledging that its output is a revenue cycle input.
The most useful survey of the quarter is Innovaccer’s State of Revenue Lifecycle in US Healthcare 2026 — 150 healthcare professionals across 103 organizations, peer-reviewed by Frost & Sullivan.
| Finding | Share |
|---|---|
| AI integrated into at least one workflow | 63% |
| Expanded implementations across departments | 52% |
| AI governance or ethics structure in place | 45% |
| Self-classify as early-to-mid maturity | 70% |
| Operating at enterprise scale | 8% |
| Cite fragmented data as the top barrier | 62% |
| Where the AI is: workflow automation / documentation / scheduling and access / revenue cycle | 52% / 46% / 41% / 38% |
The 62% figure is the finding to act on. For three years the stated barrier to RCM AI was trust — would it code correctly, would it hold up in an audit. That is no longer what practitioners report. The barrier is now plumbing: the model cannot see enough of the encounter because the encounter lives in four systems. That reframes the buying decision. If fragmented data is the constraint, then buying a fifth point solution makes the constraint worse, and the highest-ROI project may be an integration project that ships no AI at all.
The 2025 CAQH Index, released February 19, 2026, found that more than 50% of health plans use AI tools in administrative workflows, against 25% of provider organizations. Payers are automating the adjudication side at twice the rate providers are automating the submission side — and the party with the faster automation writes the rules of the exchange.
The narrowest and best-documented gap remains denials. Only about 14% of providers use AI specifically to reduce denials, yet 69% of those who do report fewer denials and better resubmissions. It is the highest-yield, least-adopted application in the revenue cycle, for the third consecutive quarter of survey data.
The three product announcements this week all attack the same target — the handoff. Maverick’s foundation model spans documentation-to-coding, Xsolis automates the clinical-to-payer handoff, Abridge wires evidence into the note so coding inherits it. That is a market responding to the 62% finding. The gains left in the revenue cycle are not inside any single step; they are in the seams between steps, which is exactly where a practice running four disconnected systems is losing them.
October 1, 2026 is twenty days away. The FY 2027 ICD-10-CM update — 190 new codes, 30 deletions, 4 revisions — applies to encounters from that date through September 30, 2027. The clinical themes are consistent: greater specificity, concentrated in a handful of chapters.
That work is well understood and most practices have it scheduled. The larger event is 92 days later.
CPT 2027 restructures maternity care services for the first time in roughly three decades. Maternity care becomes reportable at the service level across four separately identified phases — antepartum, labor management, delivery, and postpartum — and the familiar bundled global maternity codes are deleted.
For an OB practice this is not a coding update. It is a contracting event. Commercial obstetric contracts are overwhelmingly priced against the global codes: a single negotiated rate for a global obstetric package. When the codes those rates are attached to cease to exist on January 1, every one of those contract line items needs a replacement rate structure — negotiated, loaded, and tested — and the party that arrives at that negotiation without a modeled expectation of the four-phase total will accept whatever the payer’s crosswalk produces.
Two further CPT 2027 items:
While the code sets turn over on federal schedules, payer policy turns over on the payer’s schedule — and one lands Wednesday. Effective September 16, 2026, EviCore by Evernorth updates the esophagogastroduodenoscopy (EGD) guidelines supporting prior authorization and medical necessity determinations for Horizon Blue Cross Blue Shield of New Jersey’s Gastroenterology Program.
That is the recurring lesson of every October. The claim that fails is rarely the one carrying an unfamiliar new diagnosis code — coders are trained and vigilant on new codes. It is the routine, correctly coded claim that hits a payer edit nobody read, because payer policy updates do not come with a CMS press release and a 90-day runway.
If you bill obstetrics, start the contract work now, not in December. The global maternity codes are deleted on January 1, 2027, and every commercial contract priced against them needs a four-phase replacement structure. Model your own antepartum-plus-labor-plus-delivery-plus-postpartum total before you see the payer’s proposed crosswalk. Practices that walk into that conversation without a number will be handed one.
The 2025 CAQH Index, released February 19, 2026, is the closest thing the industry has to an audited scoreboard on administrative automation. Its headline finding is genuinely good news: U.S. healthcare avoided $258 billion in administrative spending in 2024 by conducting transactions electronically — a 17% increase in cost avoidance year over year, achieved despite the continuing fallout of the 2024 Change Healthcare cyberattack. Medical administrative spend fell 9%; dental fell 4%.
The second finding is the one to budget against: $21 billion in savings remains available from fully automating the transactions still handled manually or partially manually.
The aggregate numbers are abstract. The per-transaction numbers are not:
Denial benchmarks have settled into an uncomfortable band. The average initial denial rate is approximately 11.8% across payer types, with some categories above 19%; hospital denial rates approach 10% broadly, and the marketplace-plan average has reached nearly 20%. Unresolved denials alone can represent up to $5 million in annual revenue loss for a single organization.
Against that, one worked example is worth more than any benchmark table. Cleveland Clinic runs a 15% incoming denial rate — worse than the industry average. In 2025 its revenue cycle team overturned 92% of those denials, bringing net controllable loss to under 2%.
The lesson for a ten-physician practice is not that it should hire Cleveland Clinic’s denial team. It is that the incoming denial rate and the net loss rate are different numbers, and most practices only track the first. A 15% denial rate with a 92% overturn is a healthier revenue cycle than an 8% denial rate with a 40% overturn — and the second practice will report the better-looking KPI to its board.
The kiosk figure is the actionable one, because time-of-service collection is the only patient-balance intervention with no aging curve attached. Every dollar not collected at the desk enters a 30-, 60-, or 90-day pipeline that costs money to run and loses value the whole way down.
| KPI | Industry reality | Your target |
|---|---|---|
| Days in A/R — solo / 1–3 providers | 45–55 typical | ≤ 45 |
| Days in A/R — 4–10 providers | 40–48 typical | ≤ 40 |
| Days in A/R — 10+ providers | 35–45 typical | ≤ 35 |
| Clean claim rate | Benchmark 95%; guidance 95–98% | 98%+ |
| Initial denial rate | 11.8% average; marketplace near 20% | ≤ 5% |
| Denial overturn rate | Rarely tracked | ≥ 85% — track it monthly |
| Net controllable loss | Rarely tracked | < 2% of net revenue |
| A/R over 90 days | Varies widely | ≤ 15% of total A/R |
| Time-of-service collection rate | Most practices under 30% | ≥ 50%, and rising |
| Eligibility verification — automated share | Largest remaining CAQH gap | ≥ 95% electronic; zero routine phone checks |
| Prior authorization — electronic share | $10.97 manual vs $5.79 electronic | ≥ 75% electronic by Q1 2027 |
Every practice tracks its denial rate. Almost none track the two numbers that determine what a denial actually costs: overturn rate and net controllable loss. Cleveland Clinic’s 15% denial rate produces under 2% net loss because 92% of denials come back. Add both metrics to your monthly package this quarter — they are computable from data you already have, and they will change which denials you decide to work.
A market report published September 1, 2026 sizes the healthcare revenue cycle management software market at $40.58 billion in 2026, reaching $57.35 billion by 2030 — profiling Epic Systems, Optum, Oracle Health, R1 RCM and 17 other vendors. The stated growth drivers are AI-enabled intelligent automation, cloud-native infrastructure, and real-time claims processing.
| Market | Current | Forecast |
|---|---|---|
| RCM software | $40.58B (2026) | $57.35B by 2030 |
| Medical billing outsourcing | $14.90B (2024) | $44.30B by 2033 — 13.06% CAGR |
| Organizations outsourcing ≥ 1 RCM function | 97% | 70% plan to expand |
| Organizations reporting RCM staffing gaps | 63% — roughly two in three | Billing turnover above 22% annually |
97% of healthcare organizations now outsource at least one RCM function, and 70% plan to expand outsourcing in the coming year. The pricing is well established: 4% to 9% of net collections on a percentage model, or $4 to $12 per claim on a per-claim model.
The reason is not primarily cost. It is labor. 63% of providers report staffing gaps in their RCM departments; independent tallies put roughly two in three organizations short-staffed in revenue cycle. Billing department turnover runs above 22% annually. Certified coders start at $55,000–$75,000 in major markets, with experienced CPC-credentialed staff frequently above $90,000.
Run the arithmetic a five-physician practice actually faces. One certified coder at $75,000 plus benefits is roughly $95,000 fully loaded, against a 22% annual chance of having to re-recruit and re-train the role. At a 6% outsourcing rate, that same $95,000 buys coverage on approximately $1.6 million of net collections — with no turnover exposure and no single-point-of-failure risk when one person takes vacation in October.
That is not an argument for outsourcing. It is an argument for pricing your in-house function honestly, including turnover cost, before comparing.
Procode AI launched an AI-powered RCM offering for surgical billing, backed by new funding and an acquisition — the now-familiar pattern of acquiring an established billing business and embedding AI into its existing workflows rather than selling software into a cold market. Meanwhile, venture funding for AI-powered RCM continues to surge, and private equity and strategic acquirers are concentrating capital on companies with AI-driven automation and scalable technology.
Becker’s 11th Annual Health IT + Digital Health + RCM Conference runs September 14–17 in Chicago — the same days the CY 2027 PFS comment window closes. Practices sending staff should assign someone to file the comment letter before departure.
The RCM software market is growing at roughly 9% a year and the outsourcing market at 13% — and outsourcing is growing faster for a reason that has nothing to do with software quality. It is growing because 63% of practices cannot staff the function. Any technology decision made this year that does not explicitly account for a 22% annual turnover rate in the team meant to operate it is a decision made on the wrong variable.
Hospital price transparency enforcement of the updated requirements began April 1, 2026 — six months ago as of this issue. The revised rules, finalized in the CY 2026 OPPS/ASC final rule, changed the data elements hospitals must include in publicly available machine-readable files, with the stated aim of ensuring hospitals provide meaningful, accurate information about what they actually charge.
CMS is not finished. It has issued a Request for Information seeking feedback on strengthening MRF requirements, specifically on:
The free-text field item is the tell. Free text is where non-compliance hides in a technically compliant file — a machine-readable file that is readable by machines but not comparable across hospitals satisfies the letter of the rule and defeats its purpose. Congressional pressure is running in the same direction, with Senators Braun and Hassan having formally called on CMS to enforce the rule. CMS publishes its enforcement activities and outcomes as a public dataset on data.cms.gov — which means any practice, competitor, or plaintiff’s attorney can see enforcement history without filing a request.
CMS-0057-F requires four FHIR APIs by January 1, 2027 — 112 days from today. The regulated parties are payers: Medicare Advantage organizations, Medicaid and CHIP managed care entities, state Medicaid agencies, and QHP issuers on the federally facilitated exchanges. No physician practice is a regulated entity under this rule.
That is precisely why practices are underprepared for it. The Provider Access API and the Prior Authorization API are built to be consumed by provider systems — the PA API is specifically designed to let a provider’s system query a payer’s prior authorization requirements and submit structured requests electronically instead of through a portal or a fax.
The compliance obligation belongs to the payer. The benefit only materializes if your practice management system or EHR can talk to the API. A practice whose vendor has no published roadmap for FHIR PA submission will spend 2027 watching a fax-elimination mandate go into effect while continuing to send faxes.
There is one question worth asking your EHR vendor in writing this month: what is your delivery date for CMS-0057-F Prior Authorization API and Provider Access API consumption, and is it included in our current subscription or priced separately? The answer to the second half is the one that will surprise people.
| Item | Status / date |
|---|---|
| Hospital price transparency — updated requirements | Enforcement live since April 1, 2026; MRF RFI open |
| CMS-0057-F — PA decision timeframes (72h expedited / 7d standard) | In force since January 1, 2026 |
| CMS-0057-F — first public PA metrics reports (CY 2025 data) | Were due March 31, 2026 |
| CMS-0057-F — four FHIR APIs | January 1, 2027 — 112 days |
| CY 2027 PFS comment deadline | September 14, 2026 — 3 days |
| FY 2027 ICD-10-CM effective | October 1, 2026 — 20 days |
| CPT 2027 effective (maternity restructure, 92507 deletion, 58 Cat III codes) | January 1, 2027 — 112 days |
| Medicare telehealth flexibilities and work GPCI floor | Extended through December 31, 2027 |
Your practice has no compliance obligation under CMS-0057-F and no compliance obligation under hospital price transparency — and that is the risk, not the relief. Both rules restructure the environment your revenue cycle operates in while giving you no deadline of your own to organize around. Put the January 1, 2027 API date on your own operational calendar anyway, with a named owner and a vendor answer in writing, or it will arrive as someone else’s accomplished fact.
Two MGMA figures, taken together, explain the independent practice environment more completely than any consolidation statistic:
The long-run version of the same problem: operating cost per FTE physician rose more than 63% from 2013 to 2022, while the Medicare conversion factor rose 1.7%. And in CY 2027, as proposed, the conversion factor goes down again — to $33.17 or $32.84 depending on APM status.
There is no efficiency program that closes a 61-point spread. What there is, instead, is a set of choices about which costs a practice is willing to stop carrying manually.
60% of respondents to MGMA’s 2026 Regulatory Burden Report are in independent practices, and the report ties rising regulatory burden and falling reimbursement directly to physician burnout, consolidation, and reduced patient access.
The structural asymmetry is simple arithmetic. Every item in this issue — the September 14 comment deadline, the FY 2027 code turnover, the CPT 2027 maternity restructure, the CMS-0057-F API dependency, the Texas gold-carding recalculation, the EviCore EGD guideline update — costs approximately the same number of staff hours to absorb whether a practice has 4 physicians or 400. A health system amortizes that across a compliance department. A four-physician group amortizes it across the practice manager’s evenings.
| Benchmark | Where most independents are | Where you should be |
|---|---|---|
| Days in A/R | 45–55 (1–3 providers) | ≤ 40, trending down quarter over quarter |
| Clean claim rate | Around 95% | 98%+ |
| Initial denial rate | 11.8% average | ≤ 5% |
| Denial overturn rate | Untracked at most practices | ≥ 85%, reported monthly |
| Eligibility verification | Mixed manual and electronic | ≥ 95% electronic; zero routine phone checks |
| Prior authorization — electronic submission | Portal and fax dominant | ≥ 75% electronic before January 1, 2027 |
| 12-month PA approval-rate dossier by payer | Does not exist | Built and refreshed quarterly — required for gold carding |
| Time-of-service collection rate | Under 30% | ≥ 50% |
| Administrative FTE per physician | Frequently 3+ | Measure monthly; trend it down |
| Cost to collect | Rarely calculated | Calculate it this quarter, fully loaded, including turnover |
| EHR vendor CMS-0057-F roadmap | Unknown | In writing, with a date and a price |
The 63%-versus-1.7% gap is not going to be closed by negotiating better rates or working denials harder. It gets closed by removing labor from transactions that should never have required labor — the 24-minute eligibility call, the $10.97 manual prior authorization, the patient balance that enters a 90-day pipeline because nobody asked for it at the desk. Every one of those has a documented automated alternative and a measurable per-transaction saving. That is the whole strategy.
| Specialty | This week’s change | The number |
|---|---|---|
| Primary Care | PE/hour rewrite and facility-differential RFI both aimed at office-based cost structure; new shared medical appointment and ACP clinical-staff codes proposed | Comment closes Sept 14 |
| Cardiology | Ambulatory Specialty Model heart failure arm; FY 2027 dilated cardiomyopathy specificity | ±9%, mandatory, Jan 1, 2027 — 112 days |
| Orthopedics | Largest projected CY 2027 downside; ASM low back pain arm; 2026 efficiency adjustment already applied | −7% proposed CY 2027 |
| Oncology | Radiation delivery tier consolidation is being reclassified downward by payers | ~50% of high-complexity claims flagged or denied |
| Radiology | Efficiency adjustment on imaging; Aetna dose-compliance reduction live since Sept 1 | −15% on modifier CT without Smart Dose |
| Gastroenterology | EviCore EGD prior-authorization guideline update for Horizon BCBSNJ | Effective Sept 16, 2026 |
| Behavioral Health / SLP | 92507 deleted; SBIRT and cessation in final timed-code transition year | Telehealth secured through Dec 31, 2027 |
| OB / Maternal-Fetal | Global maternity codes deleted; four-phase reporting | Jan 1, 2027 — first restructure in ~30 years |
Primary care is the specialty with the most at stake in Monday’s comment deadline and the least obvious hook to comment about. Its exposure is not a single code — it is the practice expense methodology. Phasing out specialty-specific PE/hour data anchored to 2007 or earlier, adjusting indirect PE allocation, and formally asking whether the facility versus non-facility differential is still appropriate are three proposals that jointly reprice the economics of seeing a patient in an office. On the constructive side, CMS proposes separate coding and payment for shared medical appointments and two new HCPCS codes for clinical-staff advance care planning, with 99497 and 99498 restricted to practitioner-delivered services — a small but genuine expansion of what a primary care team can bill for work it already performs.
The Ambulatory Specialty Model is now 112 days out. It is mandatory, two-sided risk, with no opt-out and no hardship exemption, running five performance years from January 1, 2027 through 2031 with payment years 2029–2033. CMS projects approximately 8,600 physicians in selected geographies — roughly a quarter of core-based statistical areas and metropolitan divisions — managing about 600,000 heart failure and low back pain episodes. Participants are scored on quality, cost, improvement activities, and interoperability and are exempt from MIPS reporting for the duration. On the coding side, October 1 brings dilated cardiomyopathy subdivision and added specificity for rare cardiac conditions — documentation changes that cardiology notes are not currently written to support.
Orthopedic surgery carries the largest projected decrease in the CY 2027 proposed rule at −7%, concentrated on specialties that bill office visits alongside same-day procedures and on operating-room specialties whose flagship codes were revalued. This follows a CY 2026 final rule that applied a 2.5% efficiency adjustment to surgical procedures, orthopedic services, diagnostic imaging, and outpatient interventions, producing roughly −5% overall impacts for orthopedics that year. Orthopedics is also drafted into the ASM low back pain arm, which covers orthopedic surgery, neurosurgery, interventional pain, pain management, and PM&R. Two consecutive years of structural cuts plus mandatory risk is the most adverse combination facing any specialty in this issue.
The most under-reported RCM story of the quarter belongs to radiation oncology, and it is a payer-behavior story rather than a coding story. Effective January 1, 2026, CMS consolidated external beam treatment delivery into three complexity tiers — 77402 (Level 1), 77407 (Level 2), 77412 (Level 3) — the largest radiation oncology overhaul in more than a decade, with the technical component of image guidance now bundled into the delivery codes rather than separately billable.
The codes are working as designed. The payers are not processing them as designed. Practices report that cases involving high-complexity care are being reclassified by insurers into lower-tier codes, and that roughly 50% of high-complexity claims are being flagged for additional review or denied. The gap between CMS’s projections and practice experience is not coming from the code structure; it is coming from adjudication. For a radiation oncology practice, the operational response is documentation-specific: the tier determinants — multiple isocenters, active motion management — must be explicit in the treatment record on every Level 3 claim, because a downgrade appealed without that documentation will not be overturned. Medical oncology, separately, gains secondary malignant neoplasm codes identifying specific metastatic sites on October 1, which is a medical-necessity defense asset.
Two pressures, one immediate and one structural. Immediately: Aetna’s 15% payment reduction on radiology services billed with modifier CT that do not meet “Smart Dose” standards compliance has been live since September 1 and applies claim by claim. It is an equipment-attestation and dose-registry problem wearing a billing problem’s clothing, and it is not appealable on clinical grounds. Structurally, diagnostic imaging was named in the CY 2026 efficiency adjustment and remains exposed to the CY 2027 practice expense rewrite, where imaging’s high indirect-cost allocation makes it sensitive to any change in PE methodology.
The nearest deadline in this section is Wednesday, September 16, 2026, when EviCore by Evernorth updates the esophagogastroduodenoscopy guidelines underpinning prior authorization and medical necessity determinations for Horizon Blue Cross Blue Shield of New Jersey’s Gastroenterology Program. GI practices outside New Jersey should still read it: EviCore guideline revisions propagate across the plans that contract with EviCore, and an EGD criteria change is a high-volume event for any GI practice. More broadly, GI billing complexity in 2026 is being driven by prior authorization and medical necessity review rather than by code changes.
The stability news holds: Medicare telehealth flexibilities are extended through December 31, 2027, which removes the recurring cliff that has disrupted behavioral health scheduling for three years. For CY 2027, CMS proposes adding smoking and tobacco cessation and SBIRT to the timed behavioral health code upward adjustment in its final transition year — the last opportunity for that adjustment to be extended to additional services. Speech-language pathology faces a harder change: 92507 is deleted January 1, 2027, replaced by new codes. It is the most frequently reported SLP service in most practices, which means fee schedules, superbills, authorization templates, and payer contracts all reference a code that will not exist.
CPT 2027 deletes the bundled global maternity codes and replaces them with service-level reporting across antepartum, labor management, delivery, and postpartum, effective January 1, 2027 — the first restructure in roughly three decades. The clinical rationale is sound: global codes have long mispriced fragmented prenatal care. The revenue cycle consequence is that every commercial obstetric contract priced against a global code needs a new rate structure in 112 days, and OB practices that have not opened that conversation with their top three payers are already behind the payers’ own crosswalk modeling.
Ten moves for the week of September 11, 2026 — each tied directly to something in this issue.