(Rapid AI Blog)
This summer, in both OPPS and PFS proposed rules, CMS proposed to "nix" software-intensive lab services from the CLFS and perhaps even from CLIA. (Concurrently, but no coincidence, CLIA held a Request for Information on moderning CLIA, including open-ended questions about the status of digital-only and software-intensive services.)
This is a 180-degree position relative to AMA CPT (and CAP at AMA) which requires CLIA licenses, letters from CLIA medical directors, etc., for these software-intensive lab services. See: 180 Degrees.
Below, see summaries of the comments of several major organizations. At bottom, you'll find links to each PDF.
##
Chat GPT
##
CMS’s SaMS Proposal Meets
Broad Laboratory Pushback
First: The CMS proposal in brief
In the CY 2027 OPPS proposed rule, CMS takes a consequential step beyond simply creating a new label, “Software as a Medical Service” or SaMS. For a subset of services involving algorithmic analysis of previously generated laboratory data, CMS proposes that the analysis should no longer be treated as a clinical diagnostic laboratory test paid on the Clinical Laboratory Fee Schedule. Instead, CMS would classify these services with other diagnostic software, move ten existing HCPCS codes from the CLFS to New Technology APCs under the OPPS for 2027, and establish separate payment through the new “O1” status indicator. CMS also proposes that future codes meeting its SaMS-laboratory definition generally follow the same route. CMS’s rationale includes the view that the secondary analysis does not inherently require a CLIA-certified laboratory, that proprietary algorithm costs are difficult to evaluate through CLFS crosswalk and gapfill, that the CLFS lacks beneficiary cost-sharing and budget neutrality, and that similar algorithmic services should be treated consistently whether their input is a CT image or laboratory-generated data.
Major Comments
College of American Pathologists (CAP)
CAP opposes moving algorithm-only analyses of laboratory data off the CLFS. Its central position is that these analyses should remain classified as clinical laboratory services while CMS and stakeholders develop a more durable payment framework. CAP is somewhat unusual, however, in openly acknowledging that neither traditional laboratory nor hospital payment systems are especially well designed for software whose value comes from proprietary algorithms, data interpretation, and relatively non-material costs. It therefore leaves the door open to a future purpose-built payment framework that is budgetarily separate from the OPPS and better reflects clinical value. CAP’s letter has an important two-track nuance: it supports New Technology APCs and the new O1 status indicator as an interim mechanism for SaMS services already living in the OPPS world, while separately and specifically opposing removal of laboratory SaMS from the CLFS. CAP also opposes multiple-procedure discounting and unconditional packaging, and points out errors in CMS’s laboratory list, including the inclusion of 81416, which requires sequencing, and the obsolete 0208U.
American Clinical Laboratory Association (ACLA)
ACLA mounts perhaps the most comprehensive challenge to the proposal, combining regulatory, technical, operational, and patient-access arguments. Its fundamental proposition is that digitizing specimen-derived information does not somehow sever that information from the laboratory testing process: a digital pathology image, genomic data file, or other digital representation remains derived from the patient specimen, and the laboratory’s validated algorithm may be an integral part of producing the clinically meaningful result. ACLA argues that CMS should first finish its concurrent CLIA RFI concerning data-only facilities, algorithms, and postanalytic activities and only then decide how these services should be classified for payment. It also attacks CMS’s ten-code list as technically unreliable because CMS inferred what tests do from descriptors alone; ACLA notes that 0511U involves extensive wet-lab tumor culture, 81416 requires a newly collected and sequenced comparator specimen, and 0208U has already been deleted.
ACLA also emphasizes the practical consequences of moving these services to the OPPS. The performing laboratory can bill Medicare directly today; under an OPPS model, the hospital might instead have to bill for an analysis that the hospital neither performed nor possesses, potentially forcing laboratories into vendor relationships with hospitals. ACLA highlights unresolved date-of-service questions and objects strongly to introducing beneficiary coinsurance for tests that Congress placed on a fee schedule generally paid without patient cost-sharing. As an alternative to reclassification, ACLA offers one of the docket’s most concrete compromises: keep the tests on the CLFS but, if CMS is concerned about unregulated software vendors, condition Medicare payment for designated SaMS laboratory services on performance by a CLIA-certified, appropriately accredited laboratory.
Association for Molecular Pathology (AMP)
AMP strongly rejects the conceptual separation of the algorithm from the laboratory test. It describes data analysis as an inseparable professional component of a continuous diagnostic testing service and argues that when software generates a reportable laboratory result, is necessary to interpret the result, and contributes to patient care, a CLIA-certified laboratory should remain responsible for validity. AMP therefore asks CMS to clarify that the CLIA concept of a laboratory encompasses facilities that examine information derived from human specimens, including genomic, molecular, and “dry bench” activities. From there the payment conclusion follows directly: software-based laboratory tests should remain on the CLFS rather than being shifted to New Technology APCs. AMP also pushes back on the notion that CLFS valuation is inherently inadequate, pointing to PAMA’s private-payer reporting framework, and argues that the proposed change is particularly premature while CMS is simultaneously asking foundational CLIA questions about data-only facilities and the federal government is developing broader AI policy.
Coalition for 21st Century Medicine (C21)
C21 develops the most detailed structural case for keeping these analyses in the laboratory system. It argues that when a CLIA-certified laboratory performs an analysis on specimen-derived data, the service fits naturally within the existing definition of laboratory activity. In some cases, C21 stresses, the supposedly “raw” sequencing or imaging data have little or no clinical utility until the algorithm transforms them into an actionable result; separating the computational step from the laboratory test can therefore be clinically artificial. C21 also points to existing CMS practice, including CLIA edits on algorithmic codes and CMS guidance addressing laboratories that examine digital data and images, as evidence that computational activity is already embedded in the CLIA environment.
C21 considers the CLFS more suitable for payment because CMS and its MACs already have experience pricing algorithm-heavy tests, proprietary information can be handled confidentially, and PAMA private-payer data provide a market-based mechanism that may be more meaningful than hospital charges and costs. It also rejects CMS’s program-integrity rationale: if utilization problems actually emerge, C21 argues, Medicare already has LCDs, NCDs, NCCI edits, and audits. Coinsurance is not a program-integrity tool; it simply transfers part of the cost to patients. If CMS nevertheless proceeds with an OPPS approach, C21 asks for a viable pathway for laboratories themselves to bill Medicare, permanent separate payment rather than packaging, a report-based date of service, and no multiple-algorithm payment reduction.
Physician Fee Schedule Pathology Payment Coalition
The PFS Pathology Payment Coalition reaches essentially the same bottom-line conclusion but places particular emphasis on taxonomy and implementation.
It supports the term SaMS as an improvement over SaaS, yet warns CMS not to treat very different technologies—algorithm-only LDTs, laboratory SaMD, and imaging SaMD—as though they were interchangeable simply because software is involved. The Coalition recommends waiting for the CLIA RFI before establishing new payment classifications and argues that laboratory algorithms based on specimen-derived outputs remain inextricably linked to the underlying laboratory procedure regardless of whether the analysis occurs immediately or months later. It also identifies errors in Table 62, including 81416 and the deleted 0208U, and concludes that the CLFS has considerably more experience and better mechanisms than OPPS for valuing these tests.
The Coalition then lays out a fallback plan if CMS proceeds anyway. Laboratories must retain the ability to bill directly; otherwise, hospital outpatient departments would be billing for services they did not actually perform. Multiple-procedure discounting should not apply because separate algorithms answer separate clinical questions and do not necessarily share encounter-level resources. CMS should define the date of service around completion or reporting of the algorithmic analysis, and it should create a faster, predictable process for assigning genuinely new algorithmic services to New Technology APCs if that system is ultimately used.
A striking degree of consensus
The most striking feature of these comments is how little disagreement there is on the central issue. CAP, ACLA, AMP, C21 and the PFS Coalition all oppose CMS’s proposal to remove the identified laboratory analyses from the CLFS. None accepts CMS’s basic premise that the absence of a new wet-lab step makes a patient-specific computational laboratory analysis cease to be laboratory testing.
The organizations repeatedly converge on the same conceptual point: the clinically relevant unit is the diagnostic testing process, not the physical location of the last computation. Sequencing data, digitized pathology slides and other specimen-derived data do not suddenly become generic information simply because they have been converted into bits. In many modern tests, the computational analysis is precisely what converts those bits into a result that a clinician can use.
There is nearly as much agreement on regulatory sequencing. CMS is simultaneously asking through its CLIA RFI whether data-only facilities should require CLIA certification and what role algorithms play in the testing process. The commenters see an obvious sequencing problem in declaring through payment policy that these activities are not laboratory services before the agency has completed its regulatory inquiry into whether they are laboratory activities.
A second strong theme is that the proposal may create more administrative complexity rather than less. The performing entity is generally a laboratory, while OPPS is built around hospital outpatient departments. Moving the code does not move the underlying work. Several groups therefore envision the anomalous outcome of a hospital billing Medicare for a diagnostic analysis actually developed, validated and performed elsewhere by a laboratory. Direct billing, date-of-service rules, enrollment, beneficiary status and payment after discharge all become harder rather than easier.
The commenters also largely reject CMS’s invocation of transparency, budget neutrality and beneficiary cost-sharing as reasons to leave the CLFS. They acknowledge that algorithm valuation can be difficult, but argue that crosswalk, gapfill and PAMA can be improved rather than abandoned. Several suggest confidential submission of proprietary algorithm-development or cost information. And they view the absence of CLFS coinsurance as a deliberate feature of laboratory payment, not evidence that these tests belong elsewhere.
Finally, several comments expose a surprisingly basic implementation problem: CMS’s proposed Table 62 does not appear to reliably identify what CMS says it is identifying. The inclusion of a deleted code, an exome-sequencing code requiring another specimen, and—in ACLA’s example—a highly wet-lab-intensive tumor culture service makes it difficult to treat the list as evidence that a stable new payment category has already been defined.
Where the groups actually diverge
The differences are mostly matters of emphasis rather than competing policy positions. AMP is particularly categorical that computational interpretation belongs within CLIA when it generates or is integral to a laboratory result. ACLA goes further than the others in offering a practical interim regulatory guardrail: keep the service on the CLFS but require CLIA certification plus recognized accreditation for designated high-complexity software-based laboratory services.
C21 and the PFS Coalition devote more attention to designing a workable fallback system if CMS proceeds despite their objections—direct laboratory billing, separate payment, appropriate date-of-service rules and no multiple-procedure discounting. CAP likewise accepts O1 and New Technology APC treatment for SaMS that genuinely belongs in the OPPS environment.
CAP arguably leaves the largest opening for a future policy different from today's CLFS. It expressly calls for development of a durable SaMS payment framework that could better reflect clinical value and remain budgetarily separate from the OPPS. Thus CAP's position is not simply that every software-intensive diagnostic must remain on the CLFS forever. Its position is that CMS should not solve the genuine long-term SaMS payment problem by prematurely removing laboratory services from the laboratory system.
The composite stakeholder “ask” to CMS
Taken together, the comments can be distilled into a fairly coherent community recommendation:
Do not finalize the CY 2027 reclassification of laboratory SaMS from the CLFS to OPPS. Keep existing laboratory algorithm services—and, for now, new ones—within the laboratory payment framework.
Finish the CLIA work first. Determine what constitutes a laboratory in the era of digital pathology, genomics, bioinformatics and “data-only” testing, and then align Medicare payment policy with that regulatory conclusion.
Develop a technically sound definition before creating a payment category. Do not identify SaMS merely by scanning CPT descriptors for the absence of wet-lab terminology; correct Table 62 and distinguish algorithm-only LDTs, laboratory SaMD, imaging SaMD and genuinely stand-alone software.
Preserve the performing laboratory’s ability to bill Medicare directly. Avoid creating a system in which hospitals bill for outside laboratory analyses they neither develop nor perform, and resolve date-of-service and enrollment rules before any migration occurs.
Do not use coinsurance, packaging or multiple-procedure discounting as surrogate program-integrity tools. Medically unnecessary or duplicative algorithms can be controlled through coverage policy, coding edits, audits and other established Medicare mechanisms.
Improve CLFS rate-setting rather than assuming OPPS is more transparent. CMS can request additional proprietary information confidentially, continue using crosswalk and gapfill where appropriate, and incorporate PAMA private-payer data.
If CMS remains concerned about unregulated software vendors, tie laboratory payment to appropriate CLIA oversight rather than removing the service from the laboratory benefit. ACLA's accreditation proposal provides one concrete model.
If some SaMS ultimately does move into OPPS, make the transition explicit and protective: separate O1 payment, no packaging or arbitrary discounting, predictable New Technology APC procedures, and a defined pathway by which the actual performing entity gets paid.
The overall message from the laboratory and pathology community is therefore not resistance to software or AI. Quite the opposite. The comments portray software, bioinformatics and digital image analysis as increasingly intrinsic parts of modern laboratory medicine. Their objection is to using a payment rule to draw a new regulatory boundary before CMS has determined where that boundary actually belongs.
##
OPPS COMMENTS CY2027
https://www.regulations.gov/docket/CMS-2026-2344/comments
##
CAP 9p
https://www.regulations.gov/comment/CMS-2026-2344-1657
##
AMP 3p
https://www.regulations.gov/comment/CMS-2026-2344-1307
##
ACLA 8p
https://www.regulations.gov/comment/CMS-2026-2344-1933
##
C21 14p
https://www.regulations.gov/comment/CMS-2026-2344-1242
##
PFS Pathology Payment Coalition [McDermott] 5p
https://www.regulations.gov/comment/CMS-2026-2344-1408
##
##
AdvaMed Comments, not discussed above, are generally right on the same track; page 18 of 53, here.
###
Bruce Quinn’s comment takes a somewhat different approach from the organizational submissions, focusing less on the economics of CLFS versus OPPS and more on whether CMS has adequately defined the category it proposes to regulate.
He argues that “SaMS” remains conceptually unstable: modern diagnostic services routinely combine software, hardware, laboratory processes, imaging, and professional interpretation, making it difficult to identify a principled boundary between a software service and the diagnostic service of which software is an integral part. He also points to apparent inconsistencies in CMS’s Table 62, including codes that clearly entail wet-laboratory work, as evidence that the proposed classification has not yet been operationalized reliably.
On the legal and administrative side, Quinn raises a broader concern that removing services from the category of clinical laboratory tests may have consequences well beyond the fee schedule, potentially affecting laboratory billing authority, enrollment, date-of-service rules, referral restrictions, and other Medicare provisions built around the laboratory-test classification. His recommendation is therefore procedural and conservative: treat the 2027 proposal more like an RFI, defer implementation for at least a year, and use that period for stakeholder meetings, additional requests for information, and development of a more coherent definition and regulatory framework before changing payment status.
