CMS has released its 110 proposals for new lab codes for 2027. On this web page, scroll down for Agendas and Important Materials to find the proposals. CMS takes comment until October 21.
Chat GPT provides an analysis (giving me what I asked for, without my having to count row by row.)
The CMS Pricing Proposals, September 2026
110 codes are in play.
CMS proposes crosswalking 61 and gapfilling 49.
The striking finding is that CMS follows the panel’s exact majority recommendation for fewer than half the codes, although many departures preserve crosswalking while changing the comparator or multiplier.
The basic counts
| Code group | Crosswalk | Gapfill | Total |
|---|---|---|---|
| PLA codes | 53 | 49 | 102 |
| Other codes | 8 | 0 | 8 |
Total | 61 (55%) | 49 (44%) | 110 |
PLA codes constitute 93% of the list. Answering the two directional questions explicitly:
- Of proposed gapfills, 100% are PLA codes: 49/49.
- Of PLA codes, 48% are proposed for gapfill: 49/102.
Not all the items have been assigned a "final code" on which we set PLA and non-PLA decisions. Six non-PLA entries still show “TBD” in column C; I counted their distinct temporary codes in column B. Also, the 110 entries include four reconsiderations, so “110 codes under review” is more precise than “110 new codes.”
Most crosswalks are straightforward...57 out of 61!
Of the 61 proposed crosswalks, 57 (93%) use one code at one unit, without multiplication, addition, or subtraction. Only four are more complicated:
| Code | CMS proposed crosswalk |
|---|---|
| 0575U | 0005U × 0.5 |
| 0668U | 81229 + 86301 |
| 0600U | 87633 + 87632 |
| 0666U | 87633 + 87632 |
Thus, there are three additive crosswalks and one fractional crosswalk.
None uses an upward multiplier.
How often does CMS follow the expert panel?
Agreement needs to mean the same crosswalk code and arithmetic, or the same gapfill recommendation. Simply counting column I’s statements that CMS “agrees” would overstate agreement.
| Comparison with the panel’s majority recommendation | Codes |
|---|---|
| Same recommendation | 53 |
| Different crosswalk code, multiplier, or combination | 35 |
| Panel crosswalk → CMS gapfill | 19 |
| Panel gapfill → CMS crosswalk | 2 |
| No majority for a specific recommendation | 1 |
| Total | 110 |
Among the 109 codes with a specific majority recommendation, CMS agrees exactly on 49% and departs on 51%. Basicaly, CMS agrees vs disagrees 50/50.
The exception is 0657U: votes split 4/2/3 among three crosswalks, with three votes for gapfill. CMS chooses gapfill. That rejects the panel’s collective preference for crosswalking, but there was no majority for any particular comparator.
The disagreements between CMS and its Experts reveal several recurring priorities.
1. Missing private-payor data disrupts otherwise acceptable crosswalks.
For 13 codes, CMS explicitly cites the absence of private-payor data (in the currently active PAMA surveys) a relevant comparator. Examples include 0698U, where CMS substitutes 0523U for the panel’s 0570U, and MRD codes 0642U and 0647U, where CMS substitutes 81420 for 0307U. For 0635U, an atopic-dermatitis gene-expression test, CMS instead proposes gapfill because it could not identify a suitable replacement.
This is a substantial theme: a comparator can appear technically appropriate yet still fail CMS’s pricing rationale. These 13 cases overlap other disagreement categories; they are not an additional tally.
2. CMS repeatedly rejects multiplying prices to reflect additional test content.
Six infectious-disease antibody codes—0580U, 0615U, and 0636U–0639U—received unanimous panel recommendations using multipliers of two to five. CMS retains the base comparator but eliminates every multiplier, citing immunoassay efficiencies. Holding the comparator’s price constant, that produces amounts 50%–80% below the panel’s recommendation.
The same concern appears in sequencing. For 0672U–0676U, CMS rejects adding proband and comparator-code prices and proposes gapfill, citing shared informatics and reporting efficiencies. For 0687U, it removes the additional half-unit of 81266 because it believes 81265 already includes comparator-specimen resources.
3. Technical platform can outweigh similarity in clinical purpose.
The largest repeated disagreement concerns 12 methylation-based risk tests, 0616U–0627U, spanning conditions including dementia, cardiovascular disease, and psychiatric disorders. The panel favored 0565U; CMS selects 0318U, explaining that microarray technology is a better match than NGS. In 11 of these rows, 0318U explicitly received zero panel votes; it is not listed as a voting option in the remaining row.
Platform differences also drive gapfill decisions for 0690U—digital PCR versus an NGS comparator—and 0613U, where CMS distinguishes the new test’s NGS component from the proposed comparator’s PCR methods.
4. CMS favors gapfill when it considers the resource comparison insufficiently specific.
Examples include MRD codes 0688U and 0689U, rapid/ultrarapid sequencing codes 0657U–0659U, and kidney-disease code 0653U. For 0653U, CMS distinguishes analysis of a portion of the exome from the panel’s proposed comparator, which CMS describes as whole-exome sequencing. For liver methylation codes 0611U and 0612U, CMS says the descriptors do not clearly establish the method needed to support the proposed NGS crosswalk.
The recurring message is that broad clinical resemblance does not establish comparable resources.
5. CMS also overrides gapfill recommendations when it sees a usable existing method.
There are only two such reversals. For prostate-risk code 0609U, CMS selects 81539 because the descriptors are nearly identical, despite a 10–1 panel preference for gapfill. For 0604U, CMS rejects unanimous gapfill and selects the generic LC-MS/MS comparator 83789, reasoning that the method is common on the CLFS.
A further notable override is 0697U, an obesity-related genetic test: CMS adopts the laboratory’s proposed comparator 0349U, which received zero votes, over the panel’s unanimous choice of 0466U.
What It All Means
Taken together, these decisions show CMS emphasizing method, descriptor specificity.
This year CMS often lowered pricing by referring "shared processing efficiencies" for multiple analytes.
One special consideration this year, if the panel recommended a code on the fee schedule which has no parallel private payer PAMA data - CMS generally "nixed" that as a viable crosswalk, and kept looking.
Strong panel support—even unanimity—often does not overcome those CMS selection rules..
(BQ - And CMS prefers to work only from the actual text of the code, not other arguments the lab may have, like GB of sequencing, etc.)