On July 2, 2026, the Centers for Medicare and Medicaid Services (CMS) released the calendar year (CY) 2027 Hospital Outpatient Prospective Payment System (OPPS) proposed rule and on July 14, 2026, subsequently released the CY 2027 Physician Fee Schedule (PFS) proposed rule.¹ These proposed rules establish a newly named Medicare payment category, Software as a Medical Service (SaMS), for software that supports clinical decision-making through algorithmic analysis.² This is CMS’ first effort to pay for software-based clinical services in a structural way as their own category rather than as a cost buried inside reimbursement for another procedure. The consequential impact is that CMS would create a new reimbursement vehicle for software and AI-enabled products to be reimbursed as standalone services in the future. The proposal is worth close attention from stakeholders because Medicare reimbursement often determines the success of software or AI products intended for clinical purposes. Comments are due August 31, 2026, on the OPPS rule and September 14, 2026, on the PFS rule.

This alert explains why software products have historically struggled to fit within Medicare’s existing payment systems, and what the proposed rules would change, leave unresolved, and mean for companies building or investing in this space.

How Medicare Pays and the Historic Misalignment for Software

By way of background, Medicare Part B pays for physician and practitioner services through the PFS and for hospital outpatient services through the OPPS in the following manners:

• The Physician Fee Schedule. Medicare assigns each service a payment amount based on relative value units (RVUs), which reflect three components: physician work, practice expense, and professional liability insurance costs. Expenses related to software have generally been reimbursed as a practice expense. The practice expense component is intended to account for non-physician resources needed to furnish the service, such as equipment, supplies, and clinical staff time.
• Hospital Outpatient Prospective Payment System. Medicare groups services into Ambulatory Payment Classifications (APCs), which determine the payment rate for the service. Each service is also assigned a status indicator, a code that tells Medicare Administrative Contractors whether and how the service is reimbursed.

However, these approaches do not align with medical software as used today, and with AI-enabled products in particular. CMS says as much in the rules, noting that these systems were not designed for technologies whose value comes from scalable algorithms rather than material inputs. Much of the value created by AI for medical purposes comes from reducing or replacing physician interpretation, yet physician work remains the largest driver of payment under the PFS. A fully autonomous AI service could therefore involve little to no physician work, producing little reimbursement under the ordinary logic of the fee schedule despite providing significant clinical value.

CMS has addressed this in part by incorporating the value of software into a practice expense, the non-physician cost bucket, which includes equipment, supplies, and clinical staff time. CMS did this for autonomous retinal-imaging analysis (Current Procedural Terminology (CPT) 92229). While viable in part, this approach also has limitations, including generally low reimbursement. The practice expense methodology rests substantially on practice-cost survey data collected in the late 2000s, before software services of this kind existed. In addition, CMS has historically declined to accept per-click software fees as a direct cost input.³

Changes from the Proposed Rules

The proposed rules rename the category CMS previously called “Software as a Service” to Software as a Medical Service, defined as software-based technologies that support clinical decision-making through algorithmic analysis, and distinguish SaMS from prescription digital therapeutics and from remote physiologic and therapeutic monitoring.⁴ The change reflects a decision to treat clinical software as a category of medical service with its own payment approach.

The OPPS Proposed Rule. The proposed rule introduces three important measures:

1. CMS would designate 36 Healthcare Common Procedure Coding System (HCPCS) codes as SaMS. These codes include services such as AI-based retinal imaging analysis, coronary fractional flow reserve derived from CT, quantitative brain MRI analysis, algorithmic ECG risk assessment, fracture-risk modeling, and similar services.⁵
2. CMS would create a new status indicator, “O1,” to identify SaMS on outpatient claims. A status indicator is the code that tells Medicare contractors whether and how a service is paid. The proposed O1 indicator would function the same as the existing “S” indicator, meaning the service would receive separate payment and would not be subject to multiple-procedure discounting. CMS also seeks comment on whether an indicator more like “T” might be more appropriate for SaMS.
3. CMS would move the 21 SaMS codes currently paid under clinical ambulatory care into New Technology APCs. CMS describes this as interim while it develops a longer-term methodology, which indicates it may better align payment with clinical outcomes.

Notably, each designated code describes a specific, already-defined clinical service, such as retinal imaging analysis for one disease or fracture-risk modeling from a particular scan. In other words, most of these codes already exist and are reimbursed by Medicare today. The proposal primarily changes how CMS categorizes and, in some cases, pays for those services. In addition, CMS is not proposing a single Medicare payment category for AI or clinical software generally. Instead, Medicare reimbursement continues to be tied to discrete clinical services with reimbursement determined on a code-by-code basis.

The PFS Proposed Rule. The PFS proposal is more limited. Unlike the OPPS, the PFS does not use APCs or status indicators. As noted above, payment is determined through RVUs or, where a national fee has not been established, contractor pricing. Accordingly, the principal SaMS-related change in the PFS proposal is the adoption of the SaMS terminology and the proposal to pay certain algorithm-only analyses that are being removed from the Clinical Laboratory Fee Schedule through contractor pricing, as noted below. The proposed rule also contains a Request for Information (RFI) asking whether CMS should establish a consistent Medicare payment framework for SaMS technologies across care settings.

Algorithmic Clinical Laboratory Tests

Clinical laboratory tests are currently paid under a separate payment system from the PFS and OPPS called the Clinical Laboratory Fee Schedule (CLFS). Payment is available only when the test is performed by a laboratory that is certified under the Clinical Laboratory Improvement Amendments (CLIA). CMS proposes reclassifying standalone algorithmic analyses of existing lab data (e.g., genomic reanalysis and AI-based digital pathology reviews) from the CLFS to other Medicare payment methods. Specifically:

• For the OPPS, CMS proposes assigning 10 existing codes to OPPS payment categories, reflecting the agency’s view that these services are not laboratory tests once the underlying data has already been generated.
• For the PFS, CMS is proposing not to assign a fixed national payment rate. Local Medicare Administrative Contractors would instead determine payment.

CMS’s rationale is that these analyses do not require a CLIA-certified laboratory to perform once the underlying test results or images have been produced. Importantly, CMS is also considering applying this approach more broadly to other algorithm-only analyses. As a result, companies developing AI-enabled diagnostic, pathology, genomic, or other data-analysis products should evaluate whether their services could be affected and consider submitting comments on the proposal.

The Broader Reimbursement Picture

Beyond the hospital outpatient setting addressed in the CY 2027 OPPS proposed rules, there are other developments that are beginning to shape reimbursement for software and AI. CMS solicited comment in the CY 2026 PFS final rule on how to pay for software under the PFS, acknowledging that software does not fit comfortably within practice expense categories built around clinical labor, supplies, and equipment. Most recently, the CY 2027 PFS proposed rule also opens a request for information on technology- and AI-enabled primary care and the Annual Wellness Visit, and proposes a new Merit-based Incentive Payment System improvement activity crediting clinician use of AI; both are signals of CMS’s direction rather than payment rules.⁶ Additionally, the American Medical Association’s CPT Appendix S taxonomy classifies AI services as assistive, augmentative, or autonomous based on the software’s output and the role that output plays in clinical interpretation.⁷

Lastly, the pending Health Tech Investment Act would create a statutory Medicare pathway for “algorithm-based healthcare services,” directing placement in New Technology APCs based on manufacturer-submitted cost data and barring reassignment for a minimum of five years.⁸ As drafted, eligibility would turn on the service being delivered through a U.S. Food and Drug Administration (FDA)-cleared or -approved device, which would omit clinical decision support software that satisfies the statutory criteria for exclusion from the definition of device.⁹ For a company building non-device clinical decision support, the payment path runs through existing evaluation and management, care management, or SaMS routes instead. The bill remains pending, and companies should not plan a reimbursement strategy around its enactment.

Looking Ahead

The SaMS proposals are best viewed as a first step rather than a final reimbursement solution. They are interim by design as CMS develops the permanent methodology in future rulemaking. At the same time, the proposals do not eliminate the growing expectation among Medicare and commercial payers that software developers demonstrate not only analytical validity or FDA-cleared performance, but also meaningful clinical utility and measurable impact on patient care.

To better understand how these proposals may affect your business or for assistance in addressing these regulatory proposals, please contact Wilson Sonsini attorneys Jodi Daniel and Ty Kayam, or any member of Wilson Sonsini’s Healthcare and FDA Regulatory practice.

---

[1] Centers for Medicare & Medicaid Services, Medicare and Medicaid Programs: CY 2027 Payment Policies Under the Physician Fee Schedule and Other Changes to Part B Payment and Coverage Policies (CMS-1848-P) (issued July 14, 2026; released for public inspection) (hereinafter “CY 2027 PFS Proposed Rule”). The Federal Register citation should be confirmed against the published version. Comments are due September 14, 2026.

[2] Centers for Medicare & Medicaid Services, Medicare and Medicaid Programs: Hospital Outpatient Prospective Payment and Ambulatory Surgical Center Payment Systems; Proposed Rule for CY 2027 (CMS-1850-P), 91 Fed. Reg. 41,734 (July 7, 2026) (released for public inspection July 2, 2026) (hereinafter “CY 2027 OPPS Proposed Rule”).

[3] CY 2021 PFS final rule, 85 Fed. Reg. 84,472 (Dec. 28, 2020) (finalizing contractor pricing for CPT code 92229 and declining to include the RUC-recommended $25 per-use AI analysis fee as a direct practice expense input); see also CY 2022 PFS final rule, 86 Fed. Reg. 64,996 (Nov. 19, 2021) (establishing national values for CPT code 92229 by direct crosswalk to CPT code 92325).

[4] CY 2027 OPPS Proposed Rule, section X.B.1; CY 2027 PFS Proposed Rule (proposing the same terminology change from SaaS to SaMS).

[5] Id. Table 61 (proposed new technology APC and status indicator assignments for proposed SaMS HCPCS codes).

[6] Medicare and Medicaid Programs: CY 2026 Payment Policies Under the Physician Fee Schedule (CMS-1832-F) (final rule issued October 31, 2025), 90 Fed. Reg. 49,266 (Nov. 5, 2025); 21 C.F.R. §§ 882.5801, 882.5803. The CY 2025 PFS final rule established the DMHT G-codes (G0552–G0554) for devices under 21 C.F.R. § 882.5801; the CY 2026 final rule expanded the policy to include ADHD devices under § 882.5803.

[7] American Medical Association, CPT Appendix S (Artificial Intelligence Taxonomy for Medical Services and Procedures). Revisions accepted by the CPT Editorial Panel at its May 2026 meeting, effective with the 2027 CPT code set, retain the three categories but key classification to the software’s output(s) and their clinical role rather than “work performed by the machine.”

[8] Health Tech Investment Act, S. 1399, 119th Cong. (2025); H.R. 6197, 119th Cong. (2025). S. 1399 was introduced April 9, 2025, and referred to the Senate Committee on Finance; H.R. 6197 was referred to the Committees on Energy and Commerce and Ways and Means. Neither bill has advanced out of committee as of this writing.

[9] Federal Food, Drug, and Cosmetic Act § 520(o), 21 U.S.C. § 360j(o) (excluding certain clinical decision support software functions from the device definition).