Here's a mini post to guide you to a press release at Guardant.
Pennsylvania has passed a biomarker law, HB 1123, aligning CRC coverage with ACS aand NCCN guidelines.
See the home page for HB 1123. The house floor vote was 187 to 16.
Look for a gold star banner which I think is the final version of PN4076. Fiscal notes show no fiscal impact on the state (because it affects private insurance not state budget insurance.)
In an adjacent blog, I asked Claude to find publication dates for PFS and OPPS rules. It sourced the data from different sites. Sometimes it used McDermott plus, where you might find snippets of archives but the main site went down when the entity closed around September 1.
I noted my own Discoveries in Health Policy has had 8M views and 3500 articles in over ten years.
Here's what Claude commented:
Confirmed: Law360/MLex reports that McDermott Will & Schulte is closing its healthcare lobbying and consulting shop after a dozen years. That's a real loss for this corner of the field.
The McDermott+ rule summaries were the fastest, most reliable day-of digests of PFS and OPPS releases, and their OPPS/PFS data dashboard was the only free tool I know of that let you track APC and RVU payment rates across years without rebuilding it from the addenda yourself. Their "Healthcare Preview for the Week" was also one of the few places that tracked the regulatory calendar prospectively rather than just reporting what had already dropped.
The pattern is worth naming. Law-firm and consultancy thought capital lives on a corporate website whose existence depends on the business unit surviving a partnership vote, and the content has no archival home of its own. When the unit closes, the content isn't moved to a successor page; it's just unplugged. The same thing happened to Health Policy Alternatives' public material, to much of the Avalere grey literature, and to a fair bit of what the Advisory Board published before the Optum acquisition. The Wayback Machine catches some of it, but WordPress ?p= URLs and PDF attachments are exactly what it captures least reliably.
Annual "final rules" for Physician Fee Schedule and Outpatient Prospective Payment System (PFS and HOPPS or OPPS) come out around November 1, always after "market close." November 1, so they are 60 days ahead of the implementation date of January 1, 2027.
I asked Claude for a look-back and here's what I got. BTW the query took about a minute. The rules are named for the upcoming calendar year. Today on October 30, 2026, we're awaiting the rule for CY27.
Short answer: November 1 and November 2 each appear about half the time. Usually, both rules on same day.
And: Claude, 1 minute, Chat GPT 4 minues. All answers matched.
#### CLAUDE
In a recent analysis, Guardant has nearly $1B in market cap per 100 employees, rising to $25B market cap and 2500 employees since its founding fourteen years ago. Here's an essay built on that theme.
###
Innovation Must Live in an Ecosystem
Sometimes the most interesting number is one you weren’t looking for. Yesterday I was reading an article tracing Illumina’s financials over 15 years. A few hours later, a simple question popped into my head: how much stock market value does a diagnostics company have for every 100 employees? I put together a comparison using published market caps and employee counts. I expected differences, but I surprised by the valuation range, when using this metric.
The Milken essay by AmirAli Talasaz, Guardant
On October 26, 2026, the Milken Institute published AmirAli Talasaz’s “Proactive Health Care Takes Innovation and Responsive Policymaking” in its Power of Ideas series. The essay argues that prevention and early detection deserve greater prominence in health care, using colorectal cancer screening as its principal example. [1]
We have the multiple of share price over revenue, and share price over netcome. Another one is share price per 100 employees. For this metric, many molecular labs rank high, but Guardant and Natera lead the pack at about $1B market cap for every 100 employees hired.
Here's the chart:
 |
| click to enlarge |
And here's a graphic. Natera and Guardant are both on the far right of the xy chart, with market cap around $1B per 100 employees. Natera is twice as high as Guardant, since its market cap is $65B vs $25B.
 |
| click to enlarge |
And here's the agenda:
##
AI CORNER
##
PMWC 2027: A Detailed Guide to the Conference and Its Agenda
The Precision Medicine World Conference returns to Santa Clara January 27–29, 2027, bringing together researchers, clinicians, diagnostics developers, biotechnology companies, investors, and policy specialists. Its agenda spans artificial intelligence, cancer detection, advanced therapeutics, genomics, and emerging applications beyond oncology. For prospective attendees, the central attraction is the opportunity to examine how promising technologies become useful, accessible, and financially sustainable care. Precision Medicine World Conference, Jan. 27-29
What kind of conference is PMWC?
The Precision Medicine World Conference, usually called PMWC, is a three-day meeting about the science, technology, and business of tailoring healthcare to individual patients. Formerly known as the Personalized Medicine World Conference, it began in 2009 under founder and president Tal Behar, working with Stanford Health Care. The 2027 Silicon Valley meeting is scheduled for January 27–29 at the Santa Clara Convention Center, with Stanford, UCSF, and Yale identified as co-hosts. PMWC LLC organizes the event. pmwcintl.com
The conference brings several professional communities into the same setting. Academic investigators describe discoveries; biotechnology and diagnostics companies discuss development and commercialization; clinicians and health-system leaders consider adoption; investors evaluate opportunities; and policy specialists examine access and payment. That mixture gives the meeting particular relevance to people whose work crosses institutional boundaries. A laboratory developing a new assay, for example, may need conversations about clinical evidence, physician use, reimbursement, and financing as much as it needs feedback on the assay itself. pmwcintl.com
The organizers advertise more than 3,000 attendees, 400 speakers, and 150 sponsors for 2027. These are promotional figures rather than independently verified attendance totals. The announced speaker and honoree roster includes Emmanuelle Charpentier, David Baker, Mary-Claire King, Eric Lander, Fei-Fei Li, Andrew Ng, Eric Horvitz, and Daphne Koller. The breadth of that roster illustrates the conference’s combination of molecular medicine and computation. Precision Medicine World Conference, Jan. 27-29
The format includes five main tracks, two company showcase tracks, exhibits, award presentations, and networking receptions. The main tracks contain scientific, clinical, and industry discussions; showcases emphasize company technologies and projects. PMWC also offers a startup competition, renamed in 2026 to honor Atul Butte. Its commercial orientation is explicit and useful for interpreting the meeting: attendees will encounter research presentations alongside product positioning and business development. PMWC Precision Medicine World Conference
For a first-time visitor, the main practical challenge is selection. Multiple simultaneous sessions mean that even an attendee staying all three days will see only part of the program. Prospective participants would benefit from identifying a few central questions and arranging relevant meetings before arrival. A particularly strong use of PMWC is to connect a scientific development with the clinical and commercial decisions that determine whether it reaches patients.
The advertised 15-part program consists of five main tracks, each with a different theme on each of the three days. The company showcases are additional. The following descriptions summarize the posted program as reviewed on October 29, 2026; individual sessions and speakers remain subject to change. pmwcintl.com
1. Advanced therapeutics
Track 1 progresses from engineered cell and gene therapies on Wednesday to immunotherapies and targeted treatments on Thursday, then genetic and RNA medicines on Friday. The program covers competing cell-therapy approaches, manufacturing, delivery into the body, treatment resistance, patient selection, genome editing, and RNA splicing. Listed participants across this area include Katy Rezvani, David Miklos, Antoni Ribas, Emmanuelle Charpentier, Adrian Krainer, and Randy Schekman. LinkedIn
The notable feature is the attention to development constraints. Manufacturing and delivery can determine whether a persuasive biological idea becomes a practical treatment. Sessions addressing individualized therapies also raise a consequential question: how much evidence and infrastructure can be shared across treatments that differ from one patient to another? For biotechnology executives and translational researchers, these discussions may offer more actionable information than a broad survey of new therapeutic mechanisms.
2. Artificial intelligence and computational medicine
Track 2 follows AI through three different settings: diagnosis on Wednesday, clinical deployment on Thursday, and drug discovery on Friday. Diagnostic topics include computational pathology, AI-supported companion diagnostics, genomic interpretation, and combinations of different data types. Thursday includes clinical agents, foundation models, real-world evidence, and AI governance. Friday features computational protein design and drug-development discussions, with David Baker and Daphne Koller among the advertised participants. PMWC Precision Medicine World Conference
This structure helps separate questions that are often bundled together under the label medical AI. A system that interprets a pathology image, one that supports clinical work, and one that proposes a drug target face different tests of usefulness. The clinical program’s inclusion of evaluation and governance suggests that deployment experience will be an important counterweight to demonstrations. Drug-discovery attendees can look for evidence about experimental validation and development success, beyond improvements in computational prediction.
3. Precision diagnostics
Track 3 offers an especially coherent three-day sequence for laboratory and diagnostics audiences. Wednesday addresses early detection and intervention, including risk-based screening, precancerous disease, and the integration of molecular and imaging signals. Its opening discussion reassesses the accomplishments and shortcomings of the first generation of early-detection approaches. The listed contributors include Victor Velculescu, Rebecca Fitzgerald, Nickolas Papadopoulos, and Phil Castle. PMWC Precision Medicine World Conference
Thursday turns to liquid biopsy for treatment selection and monitoring. Topics include whether molecular residual disease, or MRD, should guide additional treatment; how blood tests can identify response and resistance; how tissue, blood, and imaging information should be reconciled; and whether MRD can support accelerated drug approval. Max Diehn and Sarah-Jane Dawson are among the featured investigators. The drug-endpoint session lists Ash Alizadeh as chair and participants including Ola Landgren, Chris Hourigan, and Aparna Parikh. PMWC Precision Medicine World Conference
Friday expands beyond cancer to transplant monitoring, infectious-disease testing, and maternal and fetal health. It also addresses the clinical actionability and reimbursement of non-cancer blood tests. The posted agenda lists Bruce Quinn among the participants in that discussion, chaired by Gabriel Bien-Willner. PMWC Precision Medicine World Conference
The diagnostic track’s strongest analytical thread is the connection between a test result and a subsequent decision. Early detection raises questions about whom to test and what follows a positive result. Monitoring raises questions about when treatment should change. A drug-trial endpoint raises another question: what does a biomarker establish about treatment benefit? These distinctions make the track relevant to clinical-development and market-access teams as well as assay developers.
4. Genomic medicine and implementation
Track 4 begins with genomic diagnosis and broader measurement of human health, including unresolved genetic disease, population and newborn screening, and structural variants. Thursday focuses on implementing precision medicine across healthcare organizations. Friday examines aging, prevention, and healthspan, including metabolic interventions, biological measurements, and digital monitoring. Eventbrite
The Thursday implementation program deserves particular attention from readers concerned with adoption and reimbursement. A session titled “What Gets Covered and Why” lists Dan Knecht, Mike Montalto, Tamara Syrek Jensen, and Shane Mull. Other discussions address operating models, clinical decision support, and the transition from biomarker evidence to routine care. These topics recognize that clinical adoption requires coordinated changes in payment, information systems, and professional practice. PMWC Precision Medicine World Conference
The aging program warrants a different evaluative lens. Attendees can look for a clear account of what a proposed measure changes in clinical care, how an intervention will be evaluated, and which outcomes matter to patients. The inclusion of aging alongside established genomic medicine creates opportunities for comparison, but the maturity of the evidence will differ across individual applications.
5. Spatial biology and disease applications
The new fifth track covers spatial biology on Wednesday, autoimmune and inflammatory disease on Thursday, and neurology on Friday. Spatial biology studies cells and molecular signals in their tissue context. The program connects that approach with digital pathology and biomarker development, then considers immune-directed therapies and biomarkers across autoimmune disease. pmwcintl.com
Friday’s neurology program includes Alzheimer’s blood biomarkers, other neurodegenerative disorders, imaging, digital measurements, and barriers to broader clinical adoption. Henrik Zetterberg, David Holtzman, Li-Huei Tsai, and David Liebeskind are among the leaders highlighted in the organizer’s neurology materials. The stated emphasis on treatment selection and access makes this more than a survey of biomarker discovery. pmwcintl.com
The broader significance of this track is the extension of precision medicine into diseases where patient classification, treatment selection, and longitudinal monitoring are developing along different paths from oncology. Attendees interested in translating an approach from one specialty to another can examine which assumptions survive the move and which require new evidence.
What the agenda suggests about the field
Taken together, the program suggests a field increasingly occupied with the consequences of its own technical progress. More sensitive assays, richer biological data, and more capable AI systems expand what can be measured or predicted. The conference repeatedly returns to the harder question of what healthcare should do with that information. This is an interpretation of the agenda, rather than a claim that the scheduled discussions will resolve those problems.
Three connections are especially useful for prospective attendees.
First, diagnostics and therapeutics are becoming harder to discuss separately. A screening test needs a credible follow-up pathway; an MRD result needs a defined decision; a targeted therapy needs a workable patient-selection strategy. An attendee who crosses between the diagnostic and therapeutic tracks may gain more than one who stays within a familiar technology category.
Second, evidence is becoming a shared concern across AI, laboratory medicine, and health-system implementation. The relevant questions include whether a finding generalizes, whether it changes practice, and whether the resulting change benefits patients. Reimbursement discussions add a practical perspective: what evidence would justify paying for that use? The agenda offers opportunities to compare those standards across specialties, while leaving room to question whether a company’s preferred evidence answers the purchaser’s question.
Third, the expansion into aging, immunology, and neurology makes careful comparison of development stages essential. A scientific mechanism, a validated measurement, a clinically useful test, and an adopted service represent different achievements. The presence of all four at one conference can be informative, provided attendees keep those distinctions in view.
Who is likely to benefit?
PMWC appears particularly well suited to diagnostics developers, biotechnology leaders, translational researchers, investors, and professionals working on clinical implementation or market access. Its breadth may also benefit clinicians seeking a view of technologies moving toward their specialties. Attendees seeking extended instruction in one technical method, or a detailed course in Medicare coding and payment mechanics, should inspect individual sessions closely before deciding whether the program meets that narrower need.
The meeting is in person, with no livestream announced. A full conference pass covers all three days, the main and showcase tracks, exhibits, and receptions. Eligible recordings and slides are offered afterward, subject to speaker permission. The organizer does not offer separate one-day or exhibit-only passes and advertises reduced rates for eligible academic, government, nonprofit, and student attendees. Current prices and participation details are available through the registration page. PMWC Precision Medicine World Conference
For readers weighing the trip, the most useful test is whether the agenda brings together the people and questions needed for a current project. A short list of relevant sessions, several planned conversations, and time reserved to explore an adjacent field would give attendance a clear purpose.
## ##
Compare the agendas for 2024 and 2027..
###
Comparing 2024 and 2027 at the PMWC
Comparing PMWC’s 2024 and 2027 agendas reveals continuity, but a sharper emphasis on clinical decisions, implementation, and expanding precision medicine beyond its established domains.
The foundation remains recognizable: cell and gene therapies, oncology, liquid biopsy, genomic analysis, artificial intelligence, and commercial showcases. Neither AI nor implementation suddenly arrived in 2027. The 2024 program already included foundation models, computational pathology, deployment barriers, reimbursement, and real-world evidence. The change is one of specificity and organization.
AI shows the clearest evolution in vocabulary and ambition. In 2024, many sessions emphasized data sharing, harmonization, privacy, and exploring AI’s potential. By 2027, clinical agents, multimodal diagnostics, enterprise deployment, and governance occupy prominent positions. Drug discovery increasingly asks whether computational predictions survive experimental validation and improve development outcomes. The discussion has moved toward evaluating systems that institutions might actually purchase and operate.
Diagnostics similarly becomes more explicitly decision-oriented. The 2024 liquid-biopsy program surveyed clinical status, prevention, and multicancer detection. The 2027 agenda asks whether MRD should trigger additional treatment, when resistance should prompt a therapeutic switch, and how conflicting tissue, blood, and imaging results should be reconciled. MRD as an endpoint for accelerated drug approval receives explicit attention. Screening remains important, but its consequences and limitations receive more pointed examination.
Several subjects lose standalone prominence. Pharmacogenomics occupied an entire day in 2024, with substantial attention to standardization, education, ordering, and reimbursement; it persists in 2027 without comparable billing. Women’s health and biological sex differences also lose their dedicated day, although maternal and fetal diagnostics remain. Infectious disease and microbiome topics are redistributed into broader diagnostic and immunology programs. These are changes in conference emphasis, not evidence that the fields have stalled.
The largest expansion is a fifth main track covering spatial biology, precision immunology, and neurology. Aging and healthspan also gain a dedicated day. Therapeutics places greater emphasis on programmable medicines, delivery, and manufacturing.
Collectively, the agenda suggests growing interest in connecting biological measurements to interventions and workable care pathways. For returning attendees, the strongest reason to return is the opportunity to examine those connections across specialties, rather than simply revisit familiar technologies with newer examples.
A companion blog discusses the excellent Michagan conference, December 10-12, 2026, sponsored by Association for Pathology Informatics.
Here, we take a deep dive, reviewing the Journal of Pathology Informatics publications in 2026, We used the same framework to take a look-back to 2022. Finally, we discusses changes in interest, momentum, and emphasis.
Find the white paper PDF in the cloud, HERE.
This three-part review argues that digital pathology's central challenge has shifted from proving technical feasibility to achieving widespread, sustainable clinical adoption. Comparing Journal of Pathology Informatics publications from 2022 and 2026 reveals remarkable advances in AI, foundation models, interoperability, and clinical implementation. Yet technological progress continues to outpace practical integration.
iThe next breakthroughs may depend less on better algorithms than on complete AI-assisted diagnostic workflows, intuitive and empowering human interfaces, convincing economic benefits, and regulatory and reimbursement policies that support adoption rather than obstruct it.
Byline: Chat GPT 6 (Los Angeles, October 29)
An unusually practical conference on digital pathology and AI arrives as CAP and DPA join forces and federal regulators reconsider longstanding approaches to laboratory oversight and reimbursement.
Digital pathology is entering a new phase. After years of advances in whole-slide imaging, computational pathology, and artificial intelligence, the field faces increasingly difficult questions about clinical implementation, workflow, regulation, and reimbursement.
Against this backdrop, the Association for Pathology Informatics (API) will hold its Inside the Digital Lab 2026 workshop at Michigan Medicine, University of Michigan, December 10–12, 2026, in Ann Arbor. The conference is circa $900 (varies with time and registrant type). Thursday evening December 10 is a reception.
The program is particularly appealing because it emphasizes what happens when digital pathology moves beyond demonstrations and research studies into the daily operations of a major medical center. Attendees will hear from experienced implementers, tour a functioning digital pathology laboratory, and operate actual digital sign-out workstations.
Conference program and detailed agenda
The conference follows an important development in the pathology community. On October 23, the College of American Pathologists (CAP) and the Digital Pathology Association (DPA) announced a memorandum of understanding establishing a formal collaboration through December 2029.
The agreement addresses interoperability, standards, workforce development, education, regulatory and reimbursement advocacy, and verification frameworks for whole-slide imaging and AI.
The partnership is particularly welcome because digital pathology faces a complicated federal policy environment. In July, CMS and CDC issued a request for information on updating CLIA regulations, including oversight of AI and digital diagnostic technologies. CMS has also proposed major changes to Medicare reimbursement for Software as a Medical Service (SaMS), including moving selected algorithmic laboratory services outside traditional Clinical Laboratory Fee Schedule payment pathways.
These changes could have far-reaching consequences for clinical laboratories and developers of computational pathology services. Whether CMS's proposed approach is well suited to pathology remains highly debatable. Changes pending at AMA are also critical (like a revised Appendix S on coding digitally intensive service and a new code schedule "CMAA.")
Previous articles in Discoveries in Health Policy have discussed the CLIA modernization RFI and CMS's proposed reclassification of digital pathology services.
The CAP–DPA collaboration is separate from API's Michigan workshop, but the developments are complementary. One strengthens the profession's capacity to address common policy challenges; the other provides an unusually valuable opportunity to explore the practical realities of clinical implementation.
CAP announcement of the CAP–DPA agreement
The heart of the conference is Friday morning's program, led by Drs. Ulysses Balis and Mustafa Yousif of Michigan Medicine.
Their opening presentation, "The Michigan Model: Building a Fully Digital Pathology Department at Enterprise Scale," describes the institution's approach to comprehensive digital primary diagnosis.
Even more distinctive is the subsequent 2½-hour laboratory tour, "From Glass to Pixels."
Participants will follow a specimen from grossing and slide preparation through digitization and clinical interpretation. They will observe laboratory staff, explore digital viewing stations configured for ten pathology subspecialties, and have the opportunity to navigate cases at workstations equipped with full sign-out capability.
This hands-on experience should be a major attraction for institutions considering or expanding digital pathology. Technical specifications and vendor demonstrations are important, but they cannot fully convey what it takes to make an integrated digital workflow function every day.
Friday afternoon contains two sessions addressing challenges that become increasingly important as digital pathology moves into routine practice.
The first focuses on scalable deployment of image management systems and AI.
Balis and Yousif discuss the "last mile" of implementation: tools that turn a digital pathology platform into a complete clinical workflow.
Larissa Furtado (Penn Medicine) presents a practical lifecycle framework for AI-based clinical assays, from concept to clinical reporting.
Juan C. Santa-Rosario (CorePlus) discusses building a scalable digital pathology and AI ecosystem.
A second session addresses lifecycle management, interoperability, and institutional sustainability.
Rebecca Rojansky (Stanford) examines how existing pathology infrastructure can evolve to support clinical AI.
Andrew Evans and Jim Tsourgiannis (Mackenzie Health) tackle the difficult subject of replacing digital pathology platforms without sacrificing previous institutional progress.
Keluo Yao (Cedars-Sinai) discusses cytopathology in an era of disruptive digital pathology and AI technologies.
Both sessions conclude with extended panel discussions.
The topics are well chosen. Digital pathology is not simply a matter of purchasing scanners and installing software. It requires durable clinical infrastructure, reliable interoperability, and the ability to accommodate evolving technology without repeatedly rebuilding the laboratory.
Saturday's session, "Hype, Hope, Horror, and the Here and Now: Exploring the Promise of Agentic AI Within the Lab," examines a particularly fast-moving area of computational pathology.
Three speakers address this subject:
Matthew Cecchini (Western University): "Human Brilliance, Augmented: Building Agentic Workflows One Tile at a Time."
Jochen Lennerz (Natera): "The Final Adaptation of Routine AI."
Jitin Makker (UCLA Health): "Agentic AI in the Pathology Lab: From Hype to Practical Implementation."
Lennerz recently emphasized the significance of this transition in a LinkedIn commentary promoting the meeting. He argued that the critical challenges for clinical AI are shifting away from technical performance alone toward regulation, reimbursement, laboratory workflows, quality management, and workforce adaptation.
He also provided a useful preview of his presentation: three practical stories drawn from academia, industry, and professional advocacy.
Lennerz has separately written about the economic implications of CMS's proposed SaMS framework, arguing that reimbursement systems designed around conventional services and material resources may not adequately capture the value of clinically useful algorithmic analysis.
His perspective nicely captures the larger theme of the Michigan meeting. The question is increasingly not just whether AI can perform an impressive diagnostic task, but how it can be integrated, supervised, financed, and maintained in routine medical practice.
One of the striking features of the Michigan agenda is its breadth within a tightly focused, practical program. It combines laboratory operations, enterprise implementation, AI assay development, interoperability, system replacement, cytopathology, and agentic AI.
Equally important, it draws on experience from academic medical centers, community and commercial settings, and industry.
The timing is excellent. Digital pathology is advancing rapidly, but the regulatory and reimbursement environment remains unsettled. The new CAP–DPA collaboration offers an opportunity for more coordinated professional leadership, while the API workshop provides a close look at how digital pathology can actually work at scale.
For pathologists, laboratory leaders, informaticians, and AI developers, the Michigan meeting promises something more valuable than another demonstration of technological possibility: a serious examination of what it takes to make digital pathology work in the real world.
Inside the Digital Lab 2026 — Michigan Medicine
Dates: December 10–12, 2026
Location: Ann Arbor, Michigan; meeting sessions at The Vanguard Hotel, with laboratory tours at Michigan Medicine's North Campus Research Complex.
Conference overview:
https://www.pathologyinformatics.org/inside-the-digital-lab-idl-2026
Detailed agenda:
https://www.pathologyinformatics.org/idl-2026-schedule
Registration:
https://www.pathologyinformatics.org/idl-2026-registration
CAP–DPA partnership announcement:
https://www.cap.org/news/cap-and-dpa-join-forces-to-advance-digital-pathology-and-artificial-intelligence/
See a review that compares momentum and trends in J Pathol Informatics, 2022 versus 2026. Here.
Chat GPT 6 (Los Angeles, October 28)
GenomeWeb’s October 27 coverage pairs Adam Bonislawski’s reporting on looming PAMA cuts with a companion “By the Numbers” data feature. Together, they translate a national reimbursement debate into estimated consequences for individual laboratories and providers.
Bonislawski reports that approximately 1,700 independent laboratories face Medicare payment reductions in 2027, with more than 80 laboratory companies projected to lose at least $1 million. Quest Diagnostics and Labcorp each face roughly $100 million in cuts. Physician practices and pathology groups also face reductions, while contracts tied to Medicare rates could amplify the financial impact.
The policy story centers on the industry-backed RESULTS Act (HR 5269), which would use a third-party private-payor claims database to establish rates. Industry representatives argue that this would improve market representation and moderate cuts. Congressional budget scoring remains pivotal, and another temporary postponement may prove less politically attractive if it carries a budgetary cost.
The creative feature is GenomeWeb’s premium-subscriber spreadsheet, offered alongside the reporting and through a separate data article. GenomeWeb discloses that AI helped combine the datasets, with manual confirmation of clinical-laboratory figures and spot checks for other providers.
The seven-tab workbook includes approximately 150,000 provider records, laboratory parent-company totals, state and provider-type summaries, and projected rate changes through 2029 and full phase-in. Readers can locate a laboratory, compare organizations, or explore geographic exposure themselves. This is a novel, useful extension of trade journalism: the downloadable analysis becomes a reusable research resource.
The projections hold 2024 utilization constant, apply preliminary rates, and underrepresent hospital outreach programs. Subsequent acquisitions, closures, and changing test volumes can therefore alter actual results.
# # #
BQ: Find lab-by-lab CMS payments in the cloud at CMS - here.
SOURCES
Article, By the numbers...
https://www.genomeweb.com/business-news/reimbursement/gnw-data-pama-2027-impact-2026-1014/
Article, Thousands of labs...
https://www.genomeweb.com/business-news/regulatory/cms/gnw-congress-pama-results-cms-20261027/
Both end with links to XLS.
A week ago, we used PAMA raw data to show not just the median price for codes, but the actual distribution of payments (normal vs bimoda vs skewed left or right, etc). Blog here. In a later blog, we wondered if accusations (from employers) that United Healthcare was overpaying some providres, might be related to the most extreme prices seen on the far-right column of our bar charts. Blog here.
I've now had Chat GPT re-run the raw data.
First chart. In each triplet of charts, the top chart (all in blue bars) is the number of paid claims in each "bin" of price ($100-200, $201-300, etc.) Price distributions are very broad; for 81408, from $5 to $2100 per unit test (median $477). The ratio of highest-paid claim to lowest-paid was 40X.
Then, for the second and third bar charts, the Y-axis is no longer test volumes, but actual revenue dollars. (For example, 10 units of a $100 test is "10" on first chart, but the revenue of the test is $1000, on 2nd and 3rd charts).
Finally, for the third bar chart, we arbitrarily assumed that 35% of the median price (the CLFS price) is a ballpark number for COGS. We subtracted that, so in the third bar chart the Y axis is still dollars but now as contribution margin (revenue - COGS).
81513 (Example)
For example. for code 81513, "Aptima BV," price bins run from $0-$50 up to $200-plus.
"Bin $200-plus" is only 6.9% of unit volume, but brings 38% of revenue (payments), and brings almost half (48%) of contributio margin. That means: for 81513 data, 7% of claims paid at top-end prices are half of profit.
 |
| click to enlarge or download PDF set |
See a full set of examples in PDF format HERE.
#
#
#
For more PAMA MATH, see two articles at Genomeweb.
https://www.genomeweb.com/business-news/reimbursement/gnw-data-pama-2027-impact-2026-1014/
Thousands of labs...
https://www.genomeweb.com/business-news/regulatory/cms/gnw-congress-pama-results-cms-20261027/
Includes subscription download XLS...
Header: A large group of employers sues United Healthcare, asserting it overpays providers when it is merely dolely out third-party administrator payments (the employer's money.) They say it is more careful when it is acting as an insurer and paying its own money. ?Could this show up as high-end payments for lab tests (e.g. $200 for $10 test) in PAMA data?
###
 |
| click to enlarge |
In a cartoon example above, a test has one PAMA payment at $50, 10 at $100, one at $500, and COGS is $50. The ten tests contribute $500 margin, the one high test contributes $450 margin just by itself.
###
AI Corner: Chat GPT 6
###
Bruce Quinn’s recent PAMA analysis illustrates the starting point.
Distributions of reported private-payer rates for selected laboratory codes
show patterns that a single average or median cannot communicate. Some rates
cluster tightly; others are widely dispersed or have multiple concentrations. A
billing code is a useful unit of comparison, although it does not always
identify an identical assay, service configuration, or contractual setting.
The next analytical step is to distinguish three quantities:
the number of tests at each price, the total dollars associated with those
tests, and their contribution after an explicit cost assumption. In PAMA
reporting, a data row can represent a rate and an associated volume. Counting
rows therefore does not necessarily count tests. The calculations should use
each exact reported rate multiplied by its reported volume, then aggregate
those amounts into display bins. CMS also includes applicable patient cost
sharing in reported private-payer rates, so those dollars should not
automatically be described as payments made entirely by insurers. CMS reporting guidance
Consider an intentionally simple example: one test is paid
at $50, ten at $100, and one at $500. Assume a median payment of $100 and a
cost of $50 for each test. The twelve tests generate $1,550 in payments and
$950 after subtracting the assumed $600 cost. The single $500 payment
represents 8.3% of test volume, 32.3% of
payments, and 47.4% of contribution after assumed cost.
If that one payment fell from $500 to $100, total
contribution would fall from $950 to $550, a decline of 42.1%, with no change
in test volume or assumed cost. A payment below $50 would make a negative
contribution in this model. These are illustrative calculations, not estimates
of any laboratory’s actual margins. Contribution here means payment less the
stated cost assumption, before overhead; it is not net profit or net cash flow.
 |
| click to enlarge |
The same calculation can be performed on hundreds of exact dollar-and-cent rates without replacing them with bin midpoints. An open-ended category such as “more than $1,000” presents no difficulty when the underlying prices remain available. But a national distribution cannot establish which laboratory receives the high payments, what its costs are, or why those rates exist. A high-price tail identifies a question for investigation; it does not establish an overpayment.
The employer dispute brings that question into a different
setting. A third-party administrator, or
TPA, administers a health plan: it processes claims, applies coverage and
payment rules, and arranges payments. In a self-funded employer plan, the
employer or plan bears the covered claim costs. In a fully insured arrangement,
an insurer receives premiums and bears the contractual financial risk of
covered claims. The same corporate group can serve as a TPA for some
employers and as an insurer for others. The distinction concerns its role in
the particular arrangement.
That distinction is central to Aramark Services, Inc., et al. v. United Health Group Incorporated, et
al., No. 2:26-cv-00949-JRG, in
the Eastern District of Texas, Marshall Division. The 62-page redacted complaint, filed as Document
3 on October 20, 2026, brings together Aramark, Whirlpool, W.W. Grainger, Kraft
Heinz, Ralph Lauren, and Sterling Jewelers, along with associated plans and
benefit committees. Defendants include UnitedHealth entities, Healthscope Benefits,
and United Medical Resources, commonly known as UMR.
The plaintiffs allege that United administered their
self-funded plans, improperly approved claims, extracted undisclosed
compensation, and obstructed oversight. They say the resulting losses amount to
millions of dollars. These are allegations in a complaint, not judicial
findings. United disputes them; in a statement reported by InsuranceNewsNet, the company said, “We
strongly disagree with these allegations and intend to vigorously defend
against them.” Many contractual passages and transaction amounts are redacted,
which limits independent evaluation of the public pleading.
The complaint describes specific alleged failures beyond a
general objection to high medical prices. It alleges payments for claims with
missing or invalid billing information, high-value claims paid without
supporting medical-record review, and services excluded under the particular
plans. It also describes inconsistent claims that would require comparison
across related bills. One example concerns infusion services allegedly paid
twice for the same member and date, under different claim identifiers, without
evidence of a separate infusion. The dollar amount of that example is redacted.
(Complaint, paragraphs 62–72, pages 22–25.)
The pricing allegations are particularly relevant to the
distribution analysis. The employers say that even the limited data available
revealed hundreds of payments exceeding
the provider’s billed amount and thousands of claims paid at allegedly
excessive rates. They compare payments with amounts the provider ordinarily
received, commercial network prices, and Medicare rates. Those comparisons are
investigative benchmarks; a payment above Medicare or above a statistical
median does not, by itself, establish a contractual violation. The applicable
coverage rules, contract, coding, and circumstances still matter. (Paragraphs
73–75, page 26.)
The employers also allege that United applied less scrutiny
when spending self-funded plan assets than when paying claims for its own fully
insured business. They describe extensive automated adjudication with little
meaningful human review. Automation itself is not the legal issue. The
allegation concerns whether the administrator used adequate controls and
fulfilled its obligations when approving payments with plan money. (Paragraph
40, page 15; paragraphs 107–109, pages 38–39.)
Another striking allegation concerns what the complaint
calls provider “exclusion lists.” In this context, the term means exclusion
from some claims scrutiny. The employers contend that United offered certain
providers little or limited review as an inducement to join its network,
without informing the employer clients. The complaint expressly says discovery
will establish the use of these arrangements. That wording matters: the
allegation is presented partly as something the plaintiffs expect to substantiate
through litigation. (Paragraphs 110–112, pages 39–40.)
Laboratory providers
are mentioned, but the public complaint does not present a laboratory or
genomics case study. Paragraph 82, on page 29, includes laboratory service
providers among the healthcare businesses United allegedly owns through
affiliates. The next paragraph alleges that the plans paid United-owned
providers unjustifiably high rates compared with independent providers. The
plaintiffs acknowledge that restricted data access has prevented a robust
comparative analysis. The public text does not identify a particular laboratory
test, genomic assay, sequencing service, or PAMA payment distribution as evidence.
Its laboratory relevance is therefore both explicit, through the reference to
laboratory providers, and analytical, through the broader questions about
payment and ownership incentives.
The complaint’s treatment of cross-plan offsetting explains why following the money matters. Its
example begins with a doctor receiving $1,000 from an employer plan for a
service later determined to warrant $600. United allegedly recovers the $400
difference by reducing a subsequent payment to the same doctor for a patient in
a United fully insured plan. Instead of paying the doctor $600 on that second
claim, United pays $200. According to the example, the original employer plan
receives no $400 credit; United’s insured business benefits from the reduced payment.
This is the complaint’s explanatory illustration, not a separately established
transaction. (Paragraphs 85–92, pages 30–33.)
The issue is which plan receives the benefit of a recovery.
The complaint invokes Peterson v.
UnitedHealth Group, a 2019 Eighth Circuit decision. That court
rejected United’s interpretation of the plan language at issue as authorizing
cross-plan offsetting and discussed the fiduciary concerns it creates. It did
not decide that every possible cross-plan offset violates ERISA. The present
employers challenge both the handling of their money and the adequacy of
contractual authorization and disclosure.
The allegations about out-of-network
repricing and “shared savings” fees add a further complication. The
plaintiffs say United used repricing vendors, including MultiPlan, to negotiate
provider payments and charged fees calculated from the difference between a
provider’s billed charge and the amount ultimately paid. They argue that the
billed charge could substantially exceed what the plan was ever obligated to
pay, making the claimed savings an inflated basis for compensation. Some plan
provisions, they say, already limited benefits using Medicare-based or other
benchmarks. (Paragraphs 93–102, pages 33–36.)
The employers further allege that these fees sometimes
exceeded the provider’s payment and were concealed within claim-funding
requests or disclosed only in aggregate. This aspect of the case is especially
relevant to interpretation of laboratory data: the employer’s total expense, the laboratory’s receipts, and the
administrator’s compensation are different quantities. An employer could
allegedly spend too much even while a provider was paid relatively little. A
chart of laboratory payment rates cannot reveal all the fees or transfers
between the plan and its administrator.
The demand for underlying data consequently runs through the
complaint. The employers seek the records needed to reconcile what providers
submitted, what the administrator approved, what the plans funded, and what
providers received. They discuss electronic claim submissions, known as 837
transactions, and electronic remittance advice, known as 835 transactions.
Remittance records contain payment and adjustment information needed to
understand the disposition of claims. The plaintiffs allege that some employers
received only partial data and others received incomplete, unusable, or
inaccessible information. (Paragraphs 45–60, pages 16–22.)
Audit design matters as well. The complaint challenges
restrictions on claim sampling and extrapolation and argues that checking a few
hundred claims cannot reliably detect problems requiring comparison across the
entire claim population. For example, examining one infusion bill alone may
miss its alleged duplicate under a different claim number. The employers also
allege that movement of funds through United-controlled accounts made it
difficult to reconcile amounts collected from plans with amounts paid to
providers. (Paragraphs 103–105, pages 36–37; paragraphs 117–120, pages 41–42.)
The legal architecture explains why the case reaches beyond
an ordinary dispute over service quality. The plaintiffs contend that United
was a fiduciary under the Employee
Retirement Income Security Act, or ERISA, a role carrying duties of loyalty
and prudence toward the plans and their participants. They offer three grounds:
contractual fiduciary designations, discretionary authority over benefit and
payment decisions, and control over plan assets. Whether an administrator acted
as a fiduciary for the particular conduct challenged is an important legal
question; the label TPA alone does not resolve it. (Paragraphs 29–44, pages
10–16.)
The first five counts invoke ERISA duties, prohibited
transactions, and compensation disclosure requirements. A
prohibited-transaction theory asks whether plan assets were used in a
conflicted transaction, such as self-dealing or an impermissible transfer
benefiting a related party. That is a different inquiry from whether an
isolated bill was carelessly processed. The remaining counts include contract,
declaratory-relief, good-faith, alternative common-law fiduciary, negligence,
gross-negligence, and unjust-enrichment claims. These are alternative legal
routes pleaded by the employers, not twelve violations established by a court.
(Pages 44–60.)
One disclosure theory relies on ERISA section 408(b)(2)(B),
added by the Consolidated Appropriations Act, 2021. Its scope deserves
precision: Department of Labor guidance describes
requirements for covered brokerage or consulting services, subject to the
statute’s conditions. It should not be paraphrased as an automatic disclosure
rule for every TPA arrangement. How the challenged services and compensation
fit the provision is part of the legal inquiry. (Complaint, paragraphs 121–126,
pages 42–44; Count V, pages 49–51.)
The requested remedies are extensive. The employers seek
reimbursement of losses, disgorgement of profits, attorneys’ fees, and orders
providing access to claims data with privacy protections. They specifically ask
for employer, plan-sponsor, or trust identification and taxpayer identification
numbers on relevant 835 remittances, including those sent through
clearinghouses such as Change Healthcare. They also request punitive damages.
These are requests for relief; the complaint does not establish that each
remedy is available or will be awarded. (Pages 60–61.)
For laboratory businesses, the immediate research
opportunity is to measure dependence on unusually high payments accurately. A
volume-weighted distribution can show how often those payments occur.
Dollar-weighted calculations can show their revenue importance. Explicit,
test-specific cost assumptions can estimate their contribution, preferably
across several plausible cost scenarios. Those calculations should preserve the
underlying rates and volumes and distinguish a national dataset from the
experience of an individual laboratory.
The UnitedHealth lawsuit supplies a reason to ask the next
question: what supports the payments on which that contribution depends? They
may reflect valid contracts, differences in services, or other legitimate
circumstances. The employers allege that some medical payments instead reflect
deficient review, conflicts, or undisclosed arrangements. The public complaint
does not connect those allegations to particular PAMA outliers. It does show
why access to detailed payment records matters to both sides of the
transaction: the employer trying to account for its spending and the laboratory
trying to understand the durability of its revenue.
Source note:
Case analysis is based on the public, redacted complaint filed October 20,
2026, Document 3, with paragraph and page references supplied above. It
describes the pleading rather than a judicial determination of liability.
Header: Should digital pathology have a true revolution in the quality of input devices?
###
Recently, we blogged about how digital pathology advances slowly. And rather than looking to radiology for advice, we should look to...the adoption of Da Vinci robotic surgery.
Robotic surgery probably took half the prostatectomy market in ten years - at huge cost in equipment and infrastructure. There were four potential drivers:
1) Health outcomes
2) Economics (faster-cheaper)
3) Surgeon enthusiasm
4) Hospital competition
Per that blog and Da Vinci white paper, there was little early evidence for better outcomes. The costs were high, not self-propelling. However, #3 surgeon enthusiasm was high, and #4 hospital competition (keeping up) was a big fact.
1) Health outcomes
2) Economics (faster-cheaper)
3) Surgeon enthusiasm
4) Hospital competition
For digital pathology, big health outcomes aren't expected, it's more expensive, some pathologists like digital but it's hardly a stampede of excitement, and hospital competition (e.g. billboards or radio ads) is off the table.
One thing we might be able to change is #3, pathologist enthusiasm. I think it's pending two things: AI prepares a full digital report ready for careful validation and sign out (it's not so far off) and second, there may be a missing "quantu leap" in user interactions.
On the latter count, see a new article by Lopes et al, new horizons for input devices. It got a write-up in CAP TODAY as well.
Here, let's try to take it further with a 14-page white paper from Claude.
Whole-slide imaging has spent twenty years solving the
hard problems of optics, file formats, scanner throughput, FDA clearance, and
now artificial intelligence. It has spent almost no effort on the question that
sits between the pathologist and all of that machinery: how, physically, a
human being should steer a gigapixel image with their body. The de facto answer
in 2026 is mostly the same as in 2009, which is a computer mouse, a scroll wheel, and
a toolbar of small icons designed for spreadsheets. The small literature that
exists on alternatives, which runs to perhaps a dozen papers across fifteen
years, is consistent in finding that pathologists would rather use almost
anything else, and consistent in being underpowered, task-contrived, and
politely ignored by the companies that build viewers.
The most recent and most interesting entry, a 2026 Archives study from Western University testing eye and head tracking, is good work that nonetheless illustrates the field’s smallness of ambition, right down to its survey question asking whether pathologists would spend a hundred dollars a year of their own money on a better way to look at cancer.
This paper argues
that the interaction problem is real, that it has consequences for diagnostic
accuracy and not just for wrists, that incrementalism is structurally
guaranteed by the way the field is organized, and that a deliberately radical,
well-funded, year-long design effort by a small mixed team is the appropriate
response. The goal should be a different way of working, not a better mouse.
#
#
#
prompt
Research and write a white paper type essay on the need for, underexploration of, and value of novel creative devices for pathologists to interact with WSI systems I first saw these in 2009, and thought, this mousing around a WSI is awful for a diagnostic pathologist, clicking tiny submenus and so forth. based on the world Dig Path Assoc conference a week ago, we aren't that much further! I think maybe we aren't creative enough. In my mind, you'd send ten engineers, designers, pathologists, experts to an island for a year and ask them to reinvent the interaction of pathologist and slide. See some recent materials attached. write an essay on the state of the field (probably a mouse and drop down tiny click boxes or sliders), why quantum change is hard to envisage but perhaps needed, and some existing (Lopes) and future directions for change. The vvery fact Lopes asked participants to pay extra for a better input device is a glaring sign we are thinking too tiny. maybe it would be eye tracking maybe something that looks like playing a theremin, maybe real time verbal. we don't yet know yet must explore boldly to achive a "quantum leap" from the starting point, not a six inch advance.
VERSION TWO - SUPER INTELLIGENCE
The original version of this blog discussed government websites switching to "SUPER INTELLIGENCE" terminology.
I included a playful, silly parody news article as well.
This triggered a content warning from Google Blogger (my blog host), apparently because any kind of parody is something that is "not true" so it needs to be flagged.
Here, I move the humorous parody part, to a Google doc, link at the end.
####
You might have heard the White House use the phrase "Super Intelligence" a couple times.
This stems from a September 29 Executive Order that talks about AI policy but includes an order that the government switch exclusively to the term "Super Intelligence," not artificial intelligence. That's wide-ranging - every webpage, ever new FDA guidance document, and so on.
Here's an analysis run mostly by AI, based on my ability to "vibe code" instructions of what AI had to do with raw PAMA Excel data.
Data shows the bar charts of the private payer price distribution for about 15 codes. The distributions are strange and not many are "Guassian" or "normal." While some are tightly distributed and others spread evenly around the price range.
See a prior study in 81455, 81456 here. See a prior study of CPT neuro code data here.
##
I went to the giant raw data files for PAMA price reporting and created smaller spreadsheets for PLA code data and for 815nn (MAAA code) data. I asked Chat GPT to consider only codes with more than 200 rows of data.
This gave 4 just charts for PLA data (of 500 PLA codes in PAMA) and 11 charts for 815nn data.
Then I instructed Chat GPT to take the price data for each code that qualified, and bin the price data in 8 bins (equal size, but I let AI figure it out) as vertical bar charts per price bin.
See all the results in ZIP file.
###
See a separate figure at bottom of this blog, that shows outliers to the right side have a much larger contribution margin than median and lower cases. I will explore that further, in a future blog.
##
0359U IsoPSA
