50-word summary
CMS’s July 2026 CY2027 Physician Fee Schedule proposal reproduced striking Maryland data suggesting CPT 88305 may be overvalued -- without naming the source.
Discoveries in Health Policy has discovered that the trail leads to Maryland Health Care Commission officials Mesta, Chappel, and Jacobs in Health Affairs.
Their open-access article adds useful detail—and methodological questions that pathologists will find worth raising now.
150-word summary
CMS’s July 2026 CY2027 Physician Fee Schedule proposal drew attention by reporting Maryland claims data that make the 25-minute physician-time assumption for CPT 88305 look implausible. CMS called the submitter only an “interested party.” The underlying work can now be traced to the Maryland Health Care Commission and an open-access Health Affairs Forefront article by Shankar Mesta, Andre Chappel, and Douglas Jacobs. The article reproduces CMS’s extraordinary 88305 findings—1,763 provider-days exceeding eight nominal hours, 587 exceeding 24 hours, and about 37 hours when other services are included—and adds methodological detail. It also exposes questions commenters should raise before the September deadline: handling of professional versus technical components, pathology date-of-service conventions, provider attribution, and the small, nonrepresentative Urban Institute study behind the famous two-minute estimate. Brett Matsumoto’s 2020 American Economic Review critique of “impossible hours” analyses provides useful ammunition for demanding claims-level sensitivity tests before CMS acts this fall.
This blog: Authored by Chat GPT 5.6.
Source:
Found: The Maryland Source Behind CMS’s 88305 Challenge
An open-access Health Affairs article fills in the backstory to CMS’s extraordinary “37-hour workday” data—and gives both sides of the 88305 debate material worth putting into the public record now
The 88305 story in the July 2026 proposed CY2027 Medicare Physician Fee Schedule is becoming more interesting, not less.
CMS devoted several paragraphs of the proposed rule to striking Maryland claims data suggesting that the long-standing physician-time assumption for CPT 88305 may substantially exceed real-world practice. CMS reported 1,763 provider-days on which nominal 88305 intraservice time exceeded eight hours, including 587 days on which 88305 alone added up to more than 24 hours. When other billed services were added, CMS reported an average nominal workload of about 2,217 minutes—roughly 37 hours—on those high-volume days. CMS also highlighted an older Urban Institute study that measured a median 88305 intraservice time of only two minutes, versus the 25 minutes carried in the Medicare Physician Fee Schedule.
What CMS did not do was identify the nominator by name. Throughout the discussion, the Federal Register refers simply to an “interested party” or “the nominator.”
The College of American Pathologists has subsequently identified that nominator as the Maryland Health Care Commission (MHCC). But there is another piece of the paper trail that appears to have received relatively little attention: on June 5, 2026—more than a month before the proposed rule—three MHCC officials, Shankar Mesta, Andre Chappel, and Douglas Jacobs, published an open-access Health Affairs Forefront article laying out essentially the same analysis in considerably more detail.
Read Mesta, Chappel and Jacobs in Health Affairs
The article is titled “How States’ All-Payer Claims Databases Can Help CMS More Accurately Value Services.” Mesta is MHCC’s chief of Cost and Quality; the article expressly states that the authors serve at MHCC and that MHCC submitted the analysis to CMS through the Potentially Misvalued Codes process. Maryland’s own website also lists a February 2026 report carrying the same potentially-misvalued-code analysis. Thus, strictly speaking, the formal CMS source was the MHCC nomination, while Mesta et al. is the contemporaneous public exposition of that work. (Health Affairs)
For anyone preparing comments on 88305 before the September 14, 2026 deadline, the Health Affairs article is worth retrieving now. (Centers for Medicare & Medicaid Services)
Mesta et al. tells considerably more of the story
CMS summarizes the Maryland numbers rather dryly. Mesta and colleagues provide the argument behind them.
The article begins with an explicit criticism of Medicare’s traditional dependence on surveys performed through the AMA Relative Value Scale Update Committee. It cites a GAO finding that the median RUC survey contained only 52 responses and had a median response rate of 2.2 percent. The authors argue that empirical data should increasingly be used to test whether survey-derived physician times remain plausible in current practice. (Health Affairs)
Their proposed test is wonderfully simple. Take claims attributed to a provider on a given day, multiply each service by the intraservice physician time embedded in the PFS, and add the minutes. The calculation assumes the PFS times are correct. Therefore, when that assumption generates a 20-, 30-, or 40-hour workday, something in the model deserves another look.
Importantly, Maryland did not blindly screen the entire fee schedule. It began with 60 codes previously studied empirically by the Urban Institute and 326 surgical codes evaluated by RAND—services for which empirical studies had already suggested shorter times than the PFS. Maryland then looked in its 2023 all-payer claims database for billing patterns that could independently test those discrepancies. That selection method makes the study a useful confirmatory exercise, but not an unbiased census of overvaluation throughout the PFS. (Health Affairs)
And 88305 was not merely one of the results. It was the spectacular outlier.
Nearly 2,000 work-days exceed 8 hours - for 88305 alone.
Maryland found 1,763 provider-days above eight nominal hours of 88305 alone, versus only 40 such days for 88307 and much smaller numbers for the other nominated services. On 587 occasions, 88305 itself exceeded 24 nominal hours. When other services were added, Mesta et al. report an average of about 37 hours. (Health Affairs)
The article also makes a point CMS softens. Mesta and colleagues observe that gastroenterologists and dermatologists appeared frequently among these high-volume billers. They explicitly say this is not evidence of fraud or inappropriate billing. But they go on to speculate that the relatively favorable valuation may have encouraged some specialists to establish laboratories and capture pathology revenue themselves, potentially creating an incentive toward excess biopsy utilization. CMS reports the GI and dermatology finding but does not reproduce that more provocative inference. (Health Affairs)
That distinction is worth preserving. The claims analysis can test whether a time assumption is plausible. By itself, it does not establish unnecessary biopsies, self-referral abuse, or improper billing.
One curious difference between CMS and Health Affairs
There is also a small but potentially meaningful discrepancy between the two presentations.
Mesta’s exhibit labels its second-stage calculation as average intraservice time of all codes on the high-volume days. Thus, the Health Affairs table gives 37 hours for 88305, 23 hours for 15734, 12 hours for 19318, 30 hours for 19380, and so forth. (Health Affairs)
CMS describes its calculations somewhat differently. The proposed rule says that for 88305 the 37 hours included other codes and pre-service time as well. For several surgical codes, CMS consequently reports substantially larger totals than the Health Affairs exhibit—for example, about 32 rather than 23 hours for 15734, 41 rather than 30 hours for 19380, and 52 rather than 38 hours for partial hepatectomy.
This may simply reflect additional calculations contained in MHCC’s underlying nomination letter that were not reproduced in the Health Affairs table. But commenters would be justified in asking CMS to state precisely which time elements were used in each calculation and to release or summarize the analytic specifications sufficiently for replication.
For 88305, the distinction does not rescue a 37-hour result. It matters because CMS is considering a valuation policy, and reproducibility matters.
There are real questions to ask about the Maryland method
The strength of the Maryland result is its sheer magnitude. If an assumption routinely produces impossible workdays, it is difficult simply to shrug at it.
But an impossible calculated workday does not tell by itself which assumption is wrong. The embedded PFS time may be wrong. The claims may be attributed to a provider in a way that does not correspond to personal physician work. Component billing may be misunderstood. The date attached to the claim may not be the day the pathologist actually reviewed the slide. Several effects can coexist.
This is where commenters opposed to a major 88305 revaluation should concentrate their effort. The best response is not that claims-based reality checks are conceptually illegitimate. It is to show specifically where the Maryland calculation does—or does not—translate claims into physician work correctly.
One obvious question is professional versus technical component billing. Surgical pathology can be billed globally, as a professional component with modifier 26, or as a technical component. A technical-only 88305 unit clearly cannot be assigned 25 minutes of pathologist interpretation. The published Health Affairs article says Maryland multiplied the “quantity of services submitted by a rendering provider” by PFS time, but it does not explain in the article how component modifiers, split billing, duplicate global/component representations, or other claims mechanics were normalized. That does not prove Maryland got this wrong; it identifies something that needs to be shown.
A second issue is particularly important in pathology: date of service. CMS’s own billing guidance states that the technical component of surgical pathology is dated to specimen collection. For a globally billed pathology service, Medicare permits the provider to report either the date the professional interpretation is completed or the date the technical component was performed. A professional-component-only claim uses the date the review and interpretation was completed. (Centers for Medicare & Medicaid Services)
That means a pathology claims file can potentially place multiple global services on a specimen-collection date even when the microscopic interpretations were not literally all performed on that calendar day. For an analysis whose unit of observation is the provider-day, that is not a trivial issue. Commenters should ask whether MHCC distinguished global, PC and TC claims and whether it tested the result using only claims whose date of service most closely corresponds to actual professional interpretation.
CAP has now raised these same categories of concern publicly. Its August 5 response argues that the Maryland analysis does not adequately account for Medicare billing rules, reassignment requirements, date-of-service conventions, and limitations of claims data in representing pathology intraservice work. (College Of American Pathologists)
Those objections should not remain merely in a CAP press release. To the extent they can be quantified, they belong in the formal CMS comment record.
And then there is the famous “two minutes”
The 25-versus-2-minute comparison is powerful rhetorically, but the two-minute number deserves some context.
The source is a 2016 CMS-sponsored Urban Institute report, Stephen Zuckerman et al., “Collecting Empirical Physician Time Data: Piloting an Approach for Validating Work Relative Value Units.” Its table really does show 88305 with a PFS intraservice time of 25 minutes and a median empirical time of 2 minutes—a ratio of 12.5 to 1. CMS cites this Urban report directly in the proposed rule.
Read the Urban Institute report
But this was a pilot, not a national 88305 time-and-motion survey. Urban approached nearly 20 potential organizations and ultimately collected direct-observation data at only three sites; the authors explicitly describe them as a convenience sample that “should not necessarily be viewed as representative.” Service volumes were low for many studied codes. Urban itself also singled out 88305 as clinically heterogeneous: one code can represent one or multiple tissue samples, with very different work depending on tissue source and the nature of the request.
Thus, two different propositions should not be conflated:
The current 25-minute assumption can generate implausible aggregate workloads. Maryland supplies striking evidence for that proposition.
The correct national typical 88305 time is two minutes. The Urban pilot is much weaker evidence for that much more specific proposition.
A commenter could accept the first proposition enthusiastically while challenging the second.
Matsumoto: useful ammunition for commenters
Anyone preparing a technical critique of the Maryland method should also retrieve Brett Matsumoto’s 2020 letter-to-editor in the American Economic Review, “Detecting Potential Overbilling in Medicare Reimbursement via Hours Worked: Comment.”
Read Matsumoto in the American Economic Review
Matsumoto was responding to a 2017 paper by Hanming Fang and Qing Gong. Fang and Gong had taken Medicare utilization data, assigned physician times to services, and found roughly 2,300 physicians whose claims implied more than 100 hours of Medicare work per week. (American Economic Association) [Firewall] "Detecting Potential Overbilling in Medicare Reimbursement via Hours Worked."
Matsumoto showed why some of those fantastic numbers could be artifacts of claims architecture rather than superhuman physicians. Aggregated public Medicare data can count separately billed portions of a service in ways that make the apparent service count misleading. His most dramatic examples involved global surgery and ophthalmology, where postoperative-only services could create extraordinarily inflated apparent procedure counts. He therefore turned to detailed claims data and adjusted for billing features including professional and technical components and other modifiers. (American Economic Association)
Matsumoto is not an 88305 paper, and his cataract-billing examples do not automatically invalidate Maryland’s pathology findings. That would be an overreach.
But he provides almost a ready-made methodological question for CMS:
Before converting an “impossible hours” calculation into a national RVU change, has CMS verified the result using claims-level information capable of distinguishing the billing components and attribution rules that created the apparent service count?
That is excellent ammunition for a comment letter.
Fang and Gong’s 2020 reply is also worth reading. They acknowledged the service-overcounting issue but found that their qualitative conclusions survived their corrections. They also made the important observation that regulators themselves possess much richer claims data than outside researchers and therefore need not remain trapped by limitations of aggregated public files. (AEA Publications)
That seems particularly apt here. CMS has the Medicare claims. It can perform the PC-only, global-only, modifier-specific and date-of-service sensitivity analyses itself.
A simple example [BQ]. Dr. Wilson bills CMS for 6000 units of 88305 in this database - about 2000 hours of work-time per RUC. However, he may have supervised a lab that did 5500 units of technical component (blocks, slides) for outside third-parties, which were immediately shipped out by Dr. Wilson. Dr. Wilson actually signs out 500 units of 88305 -26 (professioinal) per year, only 10 per week. Simple views of public CMS data will attribute "6000" units of 88305 to Wilson, cobbled together whether they be technical billing lines, professional billing lines, or global billing lines.
Another pathology paper points in Maryland’s direction
There is also an interesting 2017 paper from the American Journal of Clinical Pathology: Daniel Cloetingh, Rodney Schmidt and Christina Kong, “Comparison of Three Methods for Measuring Workload in Surgical Pathology and Cytopathology.”
This was not a Medicare payment study and should not be presented as one. It compared alternative methods for measuring actual pathology workload at Stanford. But its findings are remarkably relevant to the present debate.
The authors concluded that RVUs tend to favor subspecialties with high volumes of small specimens. They specifically found GI and GU workload lower when measured with the Royal College of Pathologists complexity system than when measured with RVUs, observing that both fields have large numbers of 88305 biopsies. They went further and stated that RVUs tend to overestimate workload in GI practices rich in small, mostly uncomplicated biopsy specimens. At the same time, they emphasized the enormous heterogeneity concealed inside 88305: a simple GI mucosal biopsy and a medical renal core biopsy can carry the same 88305 code despite radically different interpretive complexity. (OUP Academic)
That paper supports neither “25 minutes is right” nor “2 minutes is right.” It supports the deeper proposition that 88305 is an unusually crude container for heterogeneous pathology work, and that high-volume small-biopsy practice can look disproportionately productive when workload is measured by RVUs.
Put the arguments into the record now
CMS has not yet cut 88305 because of the Maryland analysis. In July it asked whether these services should be changed for CY2027 or in future rulemaking, and whether changes should affect physician time alone or corresponding work RVUs as well.
That makes the next several weeks important.
The Mesta article is open access, unusually readable, and written by the officials responsible for the Maryland analysis. It adds context that the Federal Register does not. Organizations intending to comment on 88305 should retrieve it, cite it, and respond to its actual methodology rather than only to CMS’s abbreviated version.
Whether CMS itself will formally cite Mesta et al. in the November final rule cannot be known. But the odds that CMS will have to grapple with the Maryland analysis are obviously high. If commenters believe there are defects in component handling, provider attribution, pathology dates of service, case-mix assumptions, or reliance on the Urban two-minute estimate, the time to put those criticisms into the administrative record is now. If CMS ultimately relies on the Maryland evidence, significant methodological objections placed before the agency will require a reasoned response in the final-rule process.
Conversely, supporters of revaluation have plenty of material here too. Mesta’s 1,763 high-volume provider-days, the 587 nominal 24-hour days, Urban’s two-minute pilot result, and the independent pathology workload literature make it difficult to maintain that the existing 25-minute assumption should simply be accepted because it has been there for years.
The emerging debate is therefore becoming considerably better than a simple fight between CMS and organized pathology. The interesting question is now empirical:
How long does contemporary 88305 professional work really take—and can Medicare construct a valuation that recognizes both the extraordinary speed of high-volume routine biopsy practice and the genuine complexity hidden inside the same code?
That is a question worth answering before somebody simply changes 25 minutes to another number.
###
Regardless of the fine print (BQ). Regardless of the argumentation chosen, you still have an 88305 H&E and four immunostains some of which are negative) tallying about 4x20 or 80 minutes, about an hour-and-a-half of pathologist time - suggesting per RUC that 2-3 such cases and half the pathologists' workday is shot.
50-word summary
CMS’s July 2026 CY2027 Physician Fee Schedule proposal reproduced striking Maryland data suggesting CPT 88305 may be overvalued without naming the source. The trail leads to Maryland Health Care Commission officials Mesta, Chappel, and Jacobs in Health Affairs. Their open-access article adds useful detail—and methodological questions worth raising now.
150-word summary
CMS’s July 2026 CY2027 Physician Fee Schedule proposal drew attention by reporting Maryland claims data that make the 25-minute physician-time assumption for CPT 88305 look implausible. CMS called the submitter only an “interested party.” The underlying work can now be traced to the Maryland Health Care Commission and an open-access Health Affairs Forefront article by Shankar Mesta, Andre Chappel, and Douglas Jacobs. The article reproduces CMS’s extraordinary 88305 findings—1,763 provider-days exceeding eight nominal hours, 587 exceeding 24 hours, and about 37 hours when other services are included—and adds methodological detail. It also exposes questions commenters should raise before the September deadline: handling of professional versus technical components, pathology date-of-service conventions, provider attribution, and the small, nonrepresentative Urban Institute study behind the famous two-minute estimate. Brett Matsumoto’s 2020 American Economic Review critique of “impossible hours” analyses provides useful ammunition for demanding claims-level sensitivity tests before CMS acts this fall.
Source:
