CMS’s New “Software as a Medical Service” Coding Category, Explained

CMS’s Calendar Year (CY) 2027 Hospital Outpatient Prospective Payment System (OPPS) proposed rule, published July 23, 2026, does something Medicare payment policy has never done before: it creates a standalone billing category for algorithmic software. The proposal designates 36 HCPCS codes as “Software as a Medical Service” (SaMS) and introduces a new O1 status indicator to pay for them separately under OPPS. For coders and compliance teams who have spent years fitting AI-driven diagnostic tools into codes built for human-performed procedures, this is the first sign that CMS is building a lane meant specifically for software.

What the Proposed Rule Actually Changes

The SaMS category covers software-based technologies that support clinical decision-making through algorithmic analysis — retinal imaging analysis, coronary fractional flow reserve derived from CT imaging, quantitative brain MRI analysis, algorithmic ECG risk assessment, and fracture-risk modeling among them. CMS is also proposing to move 10 algorithmic clinical laboratory analyses off the Clinical Laboratory Fee Schedule and into OPPS payment categories, treating them as diagnostic tests rather than lab tests. As Ropes & Gray’s July 30, 2026 analysis of the rule notes, CMS is reassigning 21 codes from clinical APCs to New Technology APCs while trying to hold CY 2026 payment levels roughly steady during the transition.

The O1 Status Indicator

The new “O1” indicator marks a claim line as Software as a Medical Service, paid separately under OPPS without the multiple-procedure discounting that applies to many other services. Functionally, it works like the existing “S” status indicator for procedures billed independent of a primary service — but O1 exists solely to flag that the value being paid for is an algorithm’s output, not a clinician’s time or a physical input.

Why CMS Says This Is Necessary

CMS frames CY 2027 as a transition year. The agency says existing payment structures don’t map well onto software whose cost scales with development and validation rather than per-unit labor or materials, and it is asking for comment on how to build “a more comprehensive and appropriate payment methodology for SaMS” going forward. In other words: this rule is a placeholder, not a finished framework, and the version CMS finalizes for CY 2028 or later could look substantially different depending on what commenters raise now.

Why This Matters for Coders and Compliance Teams

Even coders who never touch one of the 36 designated codes directly should pay attention, because the SaMS category previews how CMS intends to classify algorithmic outputs generally. Practical implications include:

• New code-to-APC mapping to learn. The 21 codes moving to New Technology APCs and the 10 codes shifting off the Clinical Laboratory Fee Schedule will need updated crosswalks in coding software and charge description masters before CY 2027 takes effect.
• Documentation standards for algorithmic output are still undefined. The rule doesn’t yet specify what supporting documentation justifies an O1 claim line, which means compliance teams should expect sub-regulatory guidance (or audit findings) to fill that gap over time.
• Denial risk during the transition. Payers unfamiliar with the O1 indicator may initially deny or bundle these claims incorrectly, similar to what happened when early digital-therapeutic CPT codes launched with inconsistent payer recognition.
• A comment window that’s still open. The deadline to submit comments on the OPPS/ASC proposed rule (CMS-1850-P) is August 31, 2026 — coders and RCM leaders with real experience billing algorithmic services have a direct channel to shape how the final rule handles documentation and payment mechanics.

What This Signals for AI-Driven Coding Tools

SaMS as proposed applies to diagnostic and clinical-decision-support software, not to coding automation itself. But the underlying logic — that an algorithm’s output can be a billable, separately reimbursable service distinct from the procedure it supports — is the same logic that will eventually have to apply to autonomous coding and CDI tools if they take on more clinical judgment. Coding teams evaluating AI-assisted platforms should watch how CMS ultimately defines documentation and audit requirements for SaMS, since those standards are a reasonable preview of what regulators will expect from any AI system that materially affects a claim.

What to Do Before the Comment Deadline

Coding and compliance leaders don’t need to wait for the final rule to start preparing. Reviewing whether any of the 36 designated codes or 10 relocated lab codes appear in current claim volume, flagging them for crosswalk updates, and drafting a comment on documentation expectations before August 31 are all reasonable steps to take now. Vendors and health systems that stay quiet during the comment period are more likely to be surprised by whatever CMS finalizes.

Coding accuracy and audit-readiness only get harder as payment categories multiply and rules shift year over year. Medikode’s automated medical coding platform (https://www.medikode.ai/) is built to keep pace with exactly this kind of regulatory change, applying current code sets and payer rules automatically so coding teams aren’t relearning crosswalks by hand every time CMS issues a new rule.