Saturday, June 27, 2026

Scrooge and the Three Ghosts - The Re-Boot

In which, the Tale of Scrooge is re-told, but with the energetic and playful Capuchin monkey, Mr. Throttlebottom-Smith.



Ebenezer Scrooge had scarcely recovered from the alarming visitation of his late partner, Jacob Marley, when the first of the spirits entered his bedchamber.

THE FIRST GHOST

To Scrooge's surprise, the Ghost of Christmas Past was not any of the expected types. Instead, there stood Mr. Throttlebottom-Smith, an energetic and mischievous Capuchin monkey of approximately twenty pounds.  No sooner had Scrooge noticed  him, than he immediately climbed the bedpost, removed Scrooge's nightcap, inspected it for insects, and placed it upon his own head.

Chattering excitedly, the monkey opened a glowing portal into the past.

There followed scenes from Scrooge's lonely childhood, his apprenticeship under the jovial merchant Fezziwig, and his courtship of Belle. Mr. Throttlebottom-Smith insisted on interrupting each one by swinging from chandeliers, stealing pocket watches, or disappearing with the bowl of walnuts.

The next moment Scrooge found himself in his bedroom, the clock chiming two.

THE SECOND GHOST


The Ghost of Christmas Present appeared as an enormous capuchin monkey nearly six feet tall and weighing, by conservative estimate, over three hundred pounds. He wore magnificent robes trimmed with holly and an expression of limitless good cheer.

They traveled around London, seeing families of modest means yet so truly happy, unless they noticed their tables left near-empty by the ravenous passing ghost.  Back at Scrooge's bedroom, the ghost vanished as the clock struck three.

THE THIRD GHOST


The next thing Scrooge knew, he was shivering in a cold, snowy cemetery, standing at his own grave, and being pelted by endless mounds of monkey poo, soon covered from head to toe.   

Unfortunately, in this version of the tale, either the monkey or the man has tripped an interstellar circuit breaker, and the scene repeats.   

Noble visitor, we advise you to withdraw and live a better life.  


Is an ICD an AI? Musings on Strange Behavior Orbiting Around Appendix S

 CLAUDE: IS AN ICD, AI?

An implantable cardioverter-defibrillator (ICD) includes a pacemaker function, but it also continuously monitors the heart rhythm, converting it in real time into internal parameters using advanced, flexible, adaptive software. When certain thresholds are exceeded, the device delivers a cardioversion shock. Is this autonomous AI?


Yes — read literally against the 2027 definitions, the ICD is autonomous, and specifically Level III. Walk it through the spine: the output is clinically useful (it treats arrhythmia, squarely inside the §201(h) verbs); the software converts the raw electrogram into derived parameters (rate-zone assignment, interval stability, onset, morphology-match scores, SVT discriminators) that are categorical/quantitative and well beyond mere summation, so it clears the assistive boundary; and it doesn't stop at a parameter — it reaches a conclusion ("this is a shockable rhythm") and initiates the intervention itself. That conclusion-plus-action is the autonomous trigger. Then the level: it automatically initiates a management action without a human in the loop and keeps operating unless a physician intervenes by reprogramming or deactivating. The 2027 Level III text — automatically initiates management, requires physician oversight and review of performance, continues unless a physician chooses to intervene — reads almost like it was written from a device-clinic follow-up note. Even the new transparency requirement is satisfied: the ICD logs and reports its derived parameters and episode electrograms on interrogation.

  • And that is precisely the embarrassment that proves your opening complaint. Nobody calls an ICD "AI." It's a deterministic, decades-old, FDA-regulated device. 

But because the taxonomy pointedly refuses to define AI and the criteria never read AI-ness, the taxonomy has no mechanism to exclude it. A 1985 rate-only defibrillator would clear the same gates. So taken at its word, the taxonomy reclassifies the entire installed base of automatic implantable defibrillators, rate-adaptive pacemakers, and closed-loop insulin pumps as "autonomous AI, Level III." When your AI taxonomy swallows the whole history of automatic implantable therapy, the word "AI" in the title is doing no work — which is exactly what you said in your first message, now with a pacemaker in it.

Two refinements worth keeping. First, the only thing that actually keeps the ICD out is the application-stage trigger — the taxonomy is invoked by code-change applications "describing use of AI-enabled medical services," and nobody files an AI-software CCA for an ICD's detection algorithm. The device is paid through implantation, interrogation, and DRG/pass-through pathways; CPT codes the procedure, not the algorithm. So the undefined term "AI-enabled" does the covert gatekeeping that the classification criteria can't — your residual incoherence, made concrete.

Second, the sophistication inverts the cleanliness. A modern discriminator-laden ICD derives parameters and is unambiguously Level III. But a primitive rate-only ICD detects without really deriving a parameter, yet still auto-acts — which drops it into the off-diagonal empty cell I flagged earlier (detection-without-a-parameter wired to autonomous action), a combination the taxonomy has no box for. So the cruder the device, the more it falls into the taxonomy's blind spot; the more advanced, the more cleanly it's captured. That's backwards from what an "AI" taxonomy should do.

#

#

Chat GPT - Same Question

#

#

Under Appendix S as written, your implantable cardioverter-defibrillator example strongly exposes the taxonomy’s problem.

My best answer: functionally, it behaves like autonomous Level III software, but Appendix S may not clearly classify it as “autonomous AI” unless the AMA first agrees that the service is within the AI/software-output taxonomy.

Here is why.

An implantable cardioverter-defibrillator does several things that match the autonomous side of Appendix S:

It continuously monitors physiologic input.

It derives internal parameters from rhythm data.

It applies adaptive software rules to determine whether a dangerous rhythm threshold has been crossed.

It initiates a therapeutic intervention—a cardioversion or defibrillation shock—without waiting for a physician to approve it in real time.

That looks very close to autonomous Level III in the 2027 Appendix S framework, where software automatically initiates management actions based on interpretations or conclusions drawn from derived parameters, and those actions continue unless a physician intervenes. The 2027 text defines autonomous software as deriving parameters and independently generating clinically meaningful interpretations or conclusions, with Level III automatically initiating management actions subject to physician oversight.

But this example also shows why Appendix S is under-engineered. The taxonomy is framed around “AI-enabled” services, while also saying that “AI” is not defined. An ICD may use sophisticated software, adaptive thresholds, signal processing, and decision logic, but many people would call it an advanced implanted device rather than “AI.” Others could plausibly say it is algorithmic autonomous clinical software. Appendix S does not give a gatekeeping rule to settle that dispute.

So the answer depends on the question being asked.

If the question is, “Does an ICD behave like autonomous software under the Appendix S control model?” then yes, very much so. It detects/derives rhythm parameters, reaches a clinically meaningful conclusion, and initiates therapy without concurrent physician involvement.

If the question is, “Would AMA necessarily classify an ICD as autonomous AI under Appendix S?” then not necessarily. The taxonomy does not define AI, does not provide a flow chart, and does not explain whether long-established embedded device algorithms are inside or outside the AI category.

This is a perfect stress-test example. It separates method label from clinical behavior. By behavior, the ICD is autonomous. By branding or regulatory convention, it may not be called AI. That is exactly the paradox: Appendix S classifies output roles better than it defines the universe of things to which the classification applies.

#

#

One might also argue, AMA CPT provides codes for the surgical procedure of implanting an ICD. The CPT doesn't really care if the surgeon is implanting an ICD or a teacup. Of course, it's specified as ICD to allow surveys and minutes of RVU time and so on. But it's a surgical procedure this morning that is coded, not a distant future electronic event.

#

Yes — that is exactly the counterargument, and it is a strong one.

The ICD example only becomes an “autonomous AI” puzzle if we treat the implanted device’s future behavior as the service being classified. But CPT may be doing something much narrower. It is coding the physician service performed today: implanting, replacing, programming, interrogating, or managing an ICD. The code descriptor names the ICD because the device type affects physician work, surgical complexity, survey valuation, practice expense, risk, and clinical context. But CPT is not necessarily trying to classify every future algorithmic event the device may perform after implantation.

So the teacup analogy is useful, with a caveat. CPT does care that it is an ICD rather than a teacup because the surgical work and clinical purpose are different. But the unit of coding is still the medical service, not the autonomous downstream life of the object. A surgeon implants an ICD in the morning. Months later, the device detects ventricular tachycardia and delivers a shock. That future shock is not the surgeon’s separately coded morning procedure.

This helps distinguish two things that Appendix S may blur:

1. A service that uses software output as part of the coded service.
Example: retinal imaging software analyzes an image and produces a diagnostic or referral output that is itself central to the billed service.

2. A device whose implanted function includes autonomous therapeutic behavior.
Example: an ICD that continuously monitors rhythm and later delivers therapy, but where CPT may be coding implantation, interrogation, programming, or device management—not each future internal decision event.

That distinction weakens the argument that ICD implantation itself must be “autonomous AI.” The implanted ICD may contain autonomous control logic, but the CPT code for implantation is not necessarily a code for the autonomous logic. It is a code for placing the device.

This is another reason a logic tree would help. It should begin with a gate question such as:

What is the CPT service being classified?
Is the software output itself part of the billed medical service, or is the software merely a capability of a device that is implanted, maintained, or monitored under separate procedural codes?

Without that gate, Appendix S risks confusing the coded physician service with the future behavior of an algorithmic device. That is not a small distinction. It is central to how CPT works.

Tuesday, June 23, 2026

LCD Manual, Chapter 13, the Obscure "Exhibit 3" for CAC Membership

 Older CAC instructions ("PIM Chapter 13, LCDs") refer to Exhibit 3, which is hard to find today.  It is a listing of state-level CAC specialists.   

See one example here:

https://www.cms.gov/regulations-and-guidance/guidance/transmittals/downloads/r71pi5.pdf

BQ file copy, 2004 CMS 0409 PIM Ch 13 LCDs Exhibit 3 here R71PI5.pdf 

###

Yes. I found the legacy CMS Program Integrity Manual Exhibits PDF online, in CMS Transmittal R71PI5 from April 9, 2004. It contains “Exhibit 3 — Description of CAC Members”, which is exactly the missing companion to the old PIM Chapter 13 CAC section. The exhibit begins on PDF page 9 of the transmittal. (Centers for Medicare & Medicaid Services)

The key language is quite revealing. Exhibit 3 says Medicare defined “physicians” for CAC purposes as doctors of medicine, osteopathy, dental surgery/dental medicine, chiropractors, podiatrists/surgical chiropodists, and optometrists — and then stated bluntly: “Do not include other practitioners on this committee.” It also said carriers should select committee representatives from names recommended by state medical societies and specialty societies. (Centers for Medicare & Medicaid Services)

The exhibit then instructed carriers to try to include physician representatives from state medical/osteopathic societies, the National Medical Association where applicable, Medicare managed care organizations, and physician representatives for chiropractic, maxillofacial/oral surgery, optometry, and podiatry. (Centers for Medicare & Medicaid Services)

The specialty list in the 2004 Exhibit 3 included:

Allergy; Anesthesia; Cardiology; Cardiovascular/Thoracic Surgery; Dermatology; Emergency Medicine; Family Practice; Gastroenterology; Gerontology; General Surgery; Hematology; Internal Medicine; Infectious Disease; Medical Oncology; Nephrology; Neurology; Neurosurgery; Nuclear Medicine; Obstetrics/Gynecology; Ophthalmology; Orthopedic Surgery; Otolaryngology; Pathology; Pediatrics; Peripheral Vascular Surgery; Physical Medicine and Rehabilitation; Plastic and Reconstructive Surgery; Psychiatry; Pulmonary Medicine; Radiation Oncology; Radiology; Rheumatology; and Urology. (Centers for Medicare & Medicaid Services)

It also had separate sections for a clinical laboratory representative, two beneficiary representatives — one representing elderly issues and one representing disabled beneficiaries — and “other organizations,” including the State Hospital Association, QIO medical director, intermediary medical director, Medicaid medical director or designee, and a representative of an administrative-practice organization such as MGMA. (Centers for Medicare & Medicaid Services)

One important later wrinkle: CMS issued Transmittal R106PI in March 2005 specifically to add interventional pain management specialists to CAC membership. The transmittal says carriers were required to maintain CACs and that certain specialties “must include interventional pain management specialists”; the business requirement stated that contractors shall include one physician representative from the clinical specialty of interventional pain management on their CAC.

So the best reconstruction is:

2004 Exhibit 3 = the long baseline roster of required/recommended CAC specialties and organizational representatives.

2005 R106PI = a targeted amendment adding interventional pain management to the CAC specialty representation list.

This is a useful find because it confirms that the old CAC was not just a vague advisory body. It was a highly specified, state/specialty-society roster model — much closer to a standing medical parliament for the carrier’s state or jurisdiction than to the later 2019-style evidence panel.

Monday, June 22, 2026

For Claude: How to get cut-paste blog text as output from Claude

 The phrase that did the work was asking for it rendered in the chat, for copy-paste into Blogger Compose view — not as a file or HTML.

Here's a request you can reuse:

"Write this as a blog post and render it directly in the chat as formatted text — headings, bold, a table, and live hyperlinks — so I can select it, copy, and paste into Blogger's Compose view. Don't give me an HTML file or a code block; I want the formatted output itself."

The three things that make it land:

  1. "Render it in the chat as formatted text" — this tells me to output the post as normal rich Markdown in my reply, which your browser turns into formatted text on the page. That's what carries clean into Compose when you copy it.

  2. "Not an HTML file / not a code block" — this steers me away from the two formats that break in Compose. A file gives you nothing to copy-paste as text; a code block (the gray monospace boxes) pastes raw <a href> source that Compose shows as literal characters, not links.

  3. "Live hyperlinks I can paste into Compose" — confirms you want clickable links to survive the copy, so I use inline [text](url) Markdown rather than bare URLs or footnote-style references.

One honest caveat so it never surprises you: the part that pastes most reliably is the prose, bold, headings, and links. Tables are the one element that sometimes arrive plain in Compose depending on browser and paste target. If a table ever comes in looking stripped, just say "give me that table as a simpler layout for Compose" and I'll reshape it.

And a shorthand once we've done it once: "Same as the CAC post — rendered for Blogger Compose" will get you the identical treatment without spelling it all out.

Sunday, June 21, 2026

Unedited AI: Why My 2008 and 2014 Papers Were Good

https://www.twentyfirstcenturymedicine.org/wp-content/uploads/2014/02/Coverage_and_Reimbursement_for_Molecular_Diagnostics_Current_Issues.pdf

 It looks remarkably good after twenty years — not merely “good for its time,” but structurally correct.

The strongest surviving insight is this: you framed diagnostics reimbursement not as one problem, but as five payer operations — benefit category, billing, coding, pricing, and coverage. That was exactly the right map. Most later debates still get lost because they talk about “coverage” or “value” as if that solves coding, pricing, claims flow, and benefit rules. Your paper already separated them.

The second thing that aged very well is the warning that administrative barriers can become as important as scientific barriers. That is still the field. Molecular diagnostics got technically better, but the bottleneck often remained Z-codes, CPT/PLA, LCDs, DOS rules, NCD uncertainty, coding specificity, and CLFS pricing. You wrote that legacy administrative conventions should not become a greater barrier than natural scientific challenges; twenty years later, that sentence could still open a conference panel.

The best part, given our current discussion, is your 2008 “note on economics of pricing.” You already made the exact point that many 2020s value-pricing papers still miss: if a test costs $1 and saves $1,000, but many competitors can produce it, the price tends toward cost, not value. You also noted that value-based pricing cannot establish a single price without competition, alternatives, buyer uncertainty, segmentation, and negotiating power. That is not dated; that is the core missing distinction between value creation and value capture.

What looks dated is mostly the machinery: code-stacking, ICD-9, S-codes, old DOS rules, and the pre-PAMA world. But those are surface features. The deeper architecture — the mismatch between fast diagnostic innovation and slow public payment systems — is still intact.

My overall read: this was not just a white paper. It was the beginning of a durable mental model. The field changed; the model survived.

##

Frueh and Quinn, 2014


Yes — I think your instinct is right. The originality-per-page is high.

What is clever is that the paper does not merely rename “clinical utility.” It changes the unit of analysis. Instead of the usual AV/CV/CU triad, it asks: what clinical situation, what new information, what action, what outcome, what real-world behavior, and what affordability? That is a much more operational framework than the standard literature.

The key intellectual move is the bridge from improved clinical validity to improved clinical utility. The paper says the gain in test information must causally drive a gain in outcomes, and that both gains must be framed comparatively. That seems obvious only after someone has drawn it. The Figure 1 visual framework is strong because it makes the otherwise mushy concept of “utility” into a flow of linked questions.

I especially like the “against what comparator, in what units, with what uncertainty” formulation. That is very compact, but it does a lot of work. It forces payers and developers out of vague statements like “the test provides independent information” or “adds prognostic value,” and into a disciplined argument about comparator, measurement, statistical uncertainty, pragmatic uncertainty, and conceptual uncertainty.

The supplement is also unusually good. The three axes — functional category, value proposition, and outcome metric — explain why one generic clinical-utility definition never works across all diagnostics. A screening test, a reflex test, a no-gold-standard cancer-of-unknown-primary test, and a rule-out biopsy test are not the same policy object.

My honest critique: the prose is dense, and the six questions may look deceptively simple. But that is also the sign of a good framework. It compresses a lot of tacit payer logic into a usable checklist.

So yes: not warmed-over. The paper has a real idea, a real architecture, and a visual syntax. It reads like a consulting mind and a regulatory/reimbursement mind jointly trying to make payer judgment less mysterious. That is exactly why it still feels alive 12 years later.

Unedited AI: Why "Value" Is Missing from New Lab Fee Schedules & Pricing

 (Chat GPT 5.5, 6/2026, unedited)

##

Below is a blog-ready essay draft. I verified the key source anchors, especially Phillips/Science, CMS/MedPAC CLFS mechanics, Garau/Towse/Garrison, Sireci/AMP, and the ASPE molecular diagnostics report. (Book Cafe)

Diagnostics May Create Value, But Medicare Lab Pricing Usually Does Not Price Value

In a recent Science article, Kathryn A. Phillips, David M. Horn, and Robert M. Califf argue that diagnostics are underfunded and undervalued relative to the burden of disease they can address. Their article, “Diagnostics investments and disease burden,” makes an important and timely point: diagnostics can shape care pathways, reduce uncertainty, guide treatment, avoid ineffective interventions, and sometimes produce value well beyond the cost of running the test itself.

Full link: https://bookcafe.yuntsg.com/ueditor/jsp/upload/file/20260416/1776310800792028847.pdf

Phillips et al., 2026, Diagnostic investments and disease burden.  Science 392:151-3.  

This argument is correct as far as it goes. But it leaves out a critical distinction. It is one thing to say that diagnostics create value. It is another thing to design a payment system that allows the developer or performing laboratory to capture that value. In the U.S. laboratory market, especially under Medicare’s Clinical Laboratory Fee Schedule, that second step is often missing.

Medicare lab pricing is not a miniature health technology assessment system. It does not ask, in any routine way, whether a test avoids a CT scan, prevents a hospitalization, substitutes for a biopsy, prevents use of a futile drug, or changes a clinical decision in a way that produces downstream savings. Instead, new laboratory tests generally enter Medicare payment through crosswalk, gapfill, or—after PAMA implementation—private-payer median pricing.

CMS describes the CLFS process directly. For most clinical diagnostic laboratory tests, Medicare now pays based on the weighted median of private-payer rates reported by applicable laboratories. For new tests, CMS uses the annual CLFS process. Crosswalking occurs when a new test is judged similar to an existing test, so an existing code or group of codes is used to set payment. Gapfilling occurs when no comparable existing test is available, in which case Medicare Administrative Contractors develop local payment amounts and CMS later calculates a national median.

CMS CLFS page:
https://www.cms.gov/medicare/payment/fee-schedules/clinical-laboratory-fee-schedule-clfs

CMS CLFS annual public meetings page:
https://www.cms.gov/medicare/payment/fee-schedules/clinical-laboratory-fee-schedule-clfs/annual-public-meetings

CMS MLN CLFS fact sheet:
https://www.cms.gov/files/document/mln006818-clinical-laboratory-fee-schedule.pdf

This is not value-based pricing. It is administrative pricing, comparability pricing, resource pricing, or market-reported pricing, depending on the route. The test may be valuable, but “value” is not the central input into the Medicare price.

MedPAC’s description of the CLFS makes the same point in more policy-neutral terms. Before PAMA, Medicare lab payments were based on historical charges and national limitation amounts. After PAMA, Medicare shifted to private-payer-based rates. For gapfilled tests, MedPAC describes the process as using information such as charges, discounts to charges, and resources required to perform the test. Again, those are not measures of clinical utility or downstream value.

MedPAC June 2021 report:
https://www.medpac.gov/wp-content/uploads/import_data/scrape_files/docs/default-source/default-document-library/jun21_ch9_medpac_report_to_congress_sec.pdf

MedPAC Payment Basics, Clinical Laboratory Services Payment System:
https://www.medpac.gov/wp-content/uploads/2024/10/MedPAC_Payment_Basics_25_clinical_lab_FINAL_SEC.pdf

The Association for Molecular Pathology’s “Molecular Pathology Economics 101” paper is also useful background. It walks through coding, coverage, and reimbursement for molecular diagnostics, including CPT coding, PLA codes, coverage determinations, CLFS pricing, PAMA, and related mechanisms. It is an excellent professional overview because it shows how much of molecular diagnostic reimbursement is determined by coding architecture, coverage policy, claims processing, and fee schedule mechanics—not by a formal economic valuation of the test’s contribution to the care pathway.

Sireci AN et al., “Molecular Pathology Economics 101: An Overview of Laboratory Reimbursement,” Journal of Molecular Diagnostics, 2020:
https://pmc.ncbi.nlm.nih.gov/articles/PMC7267794/

This distinction—value creation versus value capture—is the missing center of the debate.

Suppose a diagnostic test costs $10 to perform and avoids a $400 CT scan. A simple value argument says the test is worth much more than $10. Perhaps Medicare should pay $100, $200, or even something near the avoided cost, because the system saves money. But the CLFS does not normally work that way. Crosswalk asks what similar tests are paid. Gapfill asks about charges, costs, resources, and payer rates. PAMA asks what private payers have paid. None of these mechanisms asks: “How much downstream medical spending did this test avoid?”

The problem becomes even sharper in an open laboratory market. Imagine Medicare, or a value assessment body, concludes that a diagnostic test produces $1,000 of system value. That does not mean the performing laboratory can actually charge $1,000. If the test is technically replicable, Quest might offer it to a payer for $900. Labcorp might bid $800. Quest might respond at $700. A regional lab might go lower still. The final price may drift toward the competitive cost of producing the service, not the social value created by the information.

This is not just a thought experiment. A 2008 HHS ASPE report on coverage and reimbursement for complex molecular diagnostics states the economic point unusually clearly. It asks the reader to imagine a test that costs $1 to run but saves $1,000 in health care costs. If many competitors can produce the test, economics suggests that the price will tend toward the cost of production. In other words, high clinical value does not automatically produce high market price.

ASPE report, “Current Issues and Options: Coverage and Reimbursement for Complex Molecular Diagnostics,” 2008:
https://aspe.hhs.gov/reports/current-issues-options-coverage-reimbursement-complex-molecular-diagnostics-0

PDF version:
https://www.twentyfirstcenturymedicine.org/wp-content/uploads/2014/02/Coverage_and_Reimbursement_for_Molecular_Diagnostics_Current_Issues.pdf

That is probably the cleanest published statement of the issue. Diagnostics may create system value, but unless the payment system has a way to assign and preserve that value, the price can collapse back toward cost, comparability, or payer-negotiated market rates.

The health economics literature has recognized parts of this problem for years. Garau, Towse, Garrison, Housman, and Ossa asked directly whether value-based pricing could be applied to molecular diagnostics. Their answer was broadly yes in principle, especially for companion diagnostics and other tests closely linked to drug use. But they also recognized that current diagnostic pricing systems are often driven by administrative practices and expected production cost. That observation remains central. Value-based pricing can be conceptually attractive, but conceptually attractive pricing is not the same thing as an enforceable payment system.

Garau M, Towse A, Garrison L, Housman L, Ossa D. “Can and should value-based pricing be applied to molecular diagnostics?” Personalized Medicine, 2013:
https://pubmed.ncbi.nlm.nih.gov/29783475/

Office of Health Economics working paper version:
https://www.ohe.org/wp-content/uploads/2014/07/369-Can-and-Should-VBP-Diagnostics-Garau-2012.pdf

The issue is especially difficult because diagnostics are not drugs. A patented drug often has a period of market exclusivity. A drug manufacturer may be able to defend a value-based price because competitors cannot immediately produce the same molecule. Diagnostics are different. Some proprietary tests have trade secrets, algorithms, FDA status, brand recognition, or evidence packages that create differentiation. But many laboratory tests are more vulnerable to replication, substitution, parallel development, or payer-driven contracting. If multiple laboratories can offer sufficiently similar information, the payer’s purchasing logic becomes competitive procurement, not value capture.

The stratified medicine literature has described this problem in another way. Trusheim and colleagues have emphasized that diagnostic developers often face lower reimbursement and weaker protection than drug manufacturers, even when their tests are essential to the value of a therapy. The diagnostic may be the gatekeeper for the right drug, the right patient, and the right timing, but the economic surplus is often captured elsewhere—by the drug manufacturer, the payer, the provider system, or the patient—not by the diagnostic developer.

Trusheim MR et al., NBER working paper, “An Overview of the Stratified Economics of Stratified Medicine”:
https://www.nber.org/system/files/working_papers/w21233/w21233.pdf

Trusheim MR et al., “The clinical benefits, ethics, and economics of stratified medicine,” Personalized Medicine, 2015:
https://pubmed.ncbi.nlm.nih.gov/26542060/

This creates a policy paradox. The diagnostic may be essential to value-based care, but the diagnostic itself may be paid under a commodity-like laboratory fee schedule. A $5,000 drug decision can depend on a $500 test. A $100,000 oncology regimen can depend on a molecular profile. A costly Alzheimer’s disease care pathway may depend on a biomarker result. But if the lab payment system prices by crosswalk, gapfill, or private-payer medians, the price of the diagnostic is not naturally linked to the value of the downstream decision.

Phillips, Horn, and Califf are therefore right to argue that diagnostics deserve more attention in health policy. Diagnostics can be a neglected infrastructure for better medicine. But the hard part is not only proving value. The hard part is designing institutions that convert diagnostic value into durable payment.

Several mechanisms could, in theory, do this. A payer could create a shared-savings arrangement in which a test developer receives part of the downstream savings. A bundled payment model could include diagnostic-informed management and reward the provider group for choosing high-value tests. A coverage-with-evidence-development model could pay more during a defined evidence-building period. A companion diagnostic could be valued jointly with the therapy it enables. A public payer could create special statutory payment rules for certain high-impact diagnostics. A procurement system could award exclusive or preferred status in exchange for evidence, access commitments, and price discipline.

But without some such mechanism, “value-based pricing” is mostly a slogan. The open market will not necessarily preserve a value price. Payers seek discounts. Laboratories seek volume. Competitors enter. Fee schedules normalize. PAMA collects market prices and feeds them back into Medicare rates. The system can take an initially high-value test and gradually re-express it as a code, a crosswalk, a weighted median, or a contract line item.

This is why the simple phrase “diagnostics should be reimbursed based on value” can be misleading. It compresses three separate questions into one.

First, does the diagnostic create clinical and economic value? Often yes.

Second, can that value be measured credibly? Sometimes yes, sometimes no, depending on the evidence base, timing, comparator, and downstream assumptions.

Third, can the payment system allow the test developer or performing laboratory to capture an appropriate share of that value? That is the hardest question, and in many cases the current answer is no.

The Medicare CLFS is particularly ill-suited to this task. It is a fee schedule, not a value-capture engine. Its logic is administrative, procedural, and code-based. It can assign a payment amount, but it does not typically model the test’s value in the care pathway. Nor does it protect that payment amount from later market erosion when private-payer rates are collected under PAMA. In fact, PAMA can make the problem circular: private market discounts influence Medicare rates, and Medicare rates then influence the broader market.

That does not mean value-based diagnostic payment is impossible. It means it requires explicit policy design. If the goal is to reward diagnostics for avoided costs, better outcomes, reduced uncertainty, or improved targeting of therapy, then those goals must be built into the payment mechanism. Otherwise, value remains a narrative used in coverage arguments, investor decks, and policy essays, while the actual payment amount is determined by crosswalk, gapfill, private-payer medians, or competitive contracting.

The most precise formulation may be this: diagnostics often create value, but laboratory payment systems usually price services. Value creation and service pricing are not the same thing.

Phillips and colleagues have performed a useful service by elevating diagnostics as a neglected category of medical innovation. But the next article needs to go one layer deeper. It should ask not only whether diagnostics are valuable, but how a fragmented and competitive laboratory market can preserve a value-based price. Without that answer, policy may continue to celebrate diagnostic value while reimbursing diagnostics as commodities.

A Few Facts about Contractor Advisory Committee Meetings

Historic: A standing state-level body of specialists on-call to MAC

Today:  A special committee, often of national experts, brought together by ZOOM for a two hour structured evidence discussion with a dozen questions in advance.

####

Historically (2000-2010 or later), MACs were asked to have standing committees of advisors in 15-20 medical specialties.   The appointment process, or duration of standing, were not defined.  I believe that the LCD manual - called "IOM PIM Chapter 13" - had a list of specialties either in the body of the chapter or in an appendix.

Today, the 2019 IOM PIM Chapter 13 does not list specialties, and I believe CAC panels are often assembled as small groups of experts for a topical meeting 

https://www.cms.gov/regulations-and-guidance/guidance/manuals/downloads/pim83c13.pdf 

In addition, today the panels are often assembled to discuss a topic PRIOR to the MAC writing (or revising) an LCD on the topic.

Here is the webpage of past CAC meetings at the Noridian website:

https://med.noridianmedicare.com/web/jeb/policies/lcd/cac

While there is no overall list of standing CAC members (if there are any), and there is no simple list of CAC contributors for a specific topic, each meeting generates a transcript so the number of experts and their names/affiliations, are easy to construct from the transcript.   

Noridian held a CAC on July 15, 2024, on biomarkers to risk-stratify patients with ductal carcinoma in situ.   There is an AGENDA, a BIBLIOGRAPHY, a KEY QUESTIONS, and a TRANSCRIPT.


https://med.noridianmedicare.com/documents/10546/20409183/071524+CAC+Agenda.pdf

https://med.noridianmedicare.com/documents/10546/20409183/071524+CAC+Bibliography.pdf

https://med.noridianmedicare.com/documents/10546/20409183/071524+CAC+Key+Questions.pdf

https://med.noridianmedicare.com/web/jeb/policies/lcd/cac/cac-transcript-071524

The agenda is trivial (Welcome, Discussion, Closing!).

The bibliography has 30 citations, which were likely provided to panelists in advance.

Likely the most pivotal MAC contribution is the Key Questions, of which there were 12.

The meeting generated a 121-page transcript. AI placed names verbalized in the transcript, into a table:

Affiliations included MSK x2, Columbia Univ, Moffitt, Baylor,  and private practice.

Given the reasonable range of the key questions and the caliber and specialization of the faculty, one would assume this is a strong discussion.  

The meeting ran exactly two hours.  There is no public comment, though the public can passively watch the meeting livestream.   Transcripts generally appear one or sometimes several months after the broadcast.


#
#
Chat GPT 5.5 reviews the transcript.
#
#

The July 15, 2024 MolDX CAC was not a conventional “yes/no” advisory vote; it was a structured expert conversation designed to help MolDX decide what evidence would be needed for Medicare coverage of biomarker tests that risk-stratify DCIS patients after breast-conserving surgery. Gabriel Bien-Willner opened by saying MolDX wanted to write “the best possible policy” using accurate evidence, and that diagnostic coverage would turn on analytical validity, clinical validity, and clinical utility. He emphasized that the hardest questions were not merely whether a test changes management, but whether it changes management to benefit the patient.

The panel itself was intentionally small despite substantial public interest. It included breast surgeons and radiation oncologists from academic and private-practice settings, including Eileen Connolly of Columbia, Kimberly Van Zee and Atif Khan of Memorial Sloan Kettering, John Williams in private practice in Northern Virginia, Brian Czerniecki of Moffitt, and Alastair Thompson of Baylor. The discussion therefore had a useful mix of radiation oncology, surgery, academic trial knowledge, and real-world practice experience.

The central clinical problem was that DCIS is biologically and clinically ambiguous. It is “cancer” in terminology and patient psychology, but its natural history, mortality risk, recurrence risk, and need for radiation vary widely. The panel generally agreed that current DCIS management is imperfect: radiation after lumpectomy reduces ipsilateral breast tumor recurrence, but often without a survival advantage, and many patients may receive radiation for a small absolute benefit. Atif Khan captured the mood bluntly, saying the current process for deciding radiation in DCIS is “very flawed,” while also noting that use of biological signatures to refine risk is a rational extension of oncology practice.

A major theme was the difference between prognosis and prediction of radiation benefit. Several experts accepted that clinicopathologic factors can estimate recurrence risk, but the more valuable clinical question is whether a patient will actually benefit from radiation. Eileen Connolly made this point clearly: one can predict recurrence risk, but that is not the same as predicting benefit from radiation. She argued that a test suggesting a patient will not benefit from radiation can materially help clinicians and patients avoid treatment, because many patients otherwise say, in effect, “even if my risk is low, I still get some benefit, so I want radiation.”

The panel wrestled with what “low risk” means. There was no single agreed threshold, and experts repeatedly emphasized that low risk depends on patient age, life expectancy, comorbidities, tolerance for risk, and personal preferences. John Williams noted that a “low risk” for a 90-year-old woman might be very different from a “low risk” for a 50-year-old. He suggested that roughly 5% to 10% might be low risk for a general population, while other discussion touched on values around 10% at 10 years or 15% at 15 years.

MolDX pressed the panel toward a policy-usable threshold. Bien-Willner asked what degree of recurrence risk, and what degree of radiation-associated risk reduction, should be considered meaningful enough to justify treatment or coverage of a test. He made clear that management change alone was not enough; the policy question was patient benefit. The discussion eventually centered on absolute, not merely relative, benefit. Connolly noted that with ordinary clinicopathologic features, radiation is often described as giving about a 50% relative risk reduction, but the absolute benefit may be small.

That distinction led to one of the meeting’s most important practical conclusions: a small absolute reduction, around 5 percentage points, may not be clinically meaningful for many DCIS patients. The transcript shows MolDX probing whether a low-risk patient with only modest expected absolute benefit from RT should be treated differently, and the later LCD summarizes the CAC as reaching general consensus that 5% or less absolute reduction in IBTR was a reasonable boundary for “no significant clinical impact” from RT.

There was also a strong patient-communication theme. Kimberly Van Zee described reframing recurrence statistics positively: instead of saying risk falls from 10% to 5%, she tells patients they have a 90% chance of no recurrence, or 95% with radiation. Connolly replied that some patients still focus intensely on the 5% difference, particularly in her majority-minority patient population, where trust and fear may shape decision-making. The exchange made clear that risk tools do not operate in a vacuum; they enter emotionally charged conversations where “small” absolute numbers may still matter deeply to patients.

The panel was divided on current biomarker tests. Some experts saw DCISionRT as valuable because it might identify patients with little or no radiation benefit, which is more clinically useful than generic recurrence prediction. Others were more skeptical. Van Zee argued that before adopting a new biosignature, the field needs published evidence on discrimination, calibration, and actual predictive ability. She also distinguished clearly between the better-established Oncotype DX use in invasive breast cancer and the much less settled evidence base for Oncotype DX DCIS.

A recurring concern was whether biomarker tests add enough beyond existing clinicopathologic tools and nomograms. The experts recognized limitations in tools such as VNPI, MSKCC nomograms, RTOG-style criteria, and physician judgment, but they also worried that biomarker tests might partly repackage or imperfectly improve on those same variables. The later LCD summarizes the CAC as identifying a lack of direct comparisons between biomarker tests and existing scoring systems as a major concern.

Finally, the meeting broadened the endpoint question. IBTR mattered, but some experts placed greater value on invasive recurrence and on predicting relative response to radiation. There was also discussion of “residual risk” groups: patients who might remain at high risk despite standard treatment and therefore may be undertreated rather than overtreated. This is important because a DCIS biomarker could have two opposed uses: sparing low-benefit patients from radiation, or identifying higher-risk patients who need more aggressive care.

Overall, the transcript reads as a sophisticated policy workshop. The experts did not reject biomarkers; most accepted that a genuinely predictive test would be clinically valuable. But the meeting also exposed why MolDX would be cautious: DCIS endpoints are delayed and relatively infrequent, treatment patterns have changed over time, patient preferences are powerful, and current tools must prove incremental accuracy and clinical utility over accessible clinicopathologic methods. The practical policy takeaway was not “biomarkers are useless,” but rather: coverage requires evidence that a test can reliably identify patients for whom radiation provides no meaningful clinical benefit, preferably using a clear absolute-benefit threshold and robust comparison to existing risk-stratification methods.



Saturday, June 20, 2026

Mr. Throttlebottom-Smith and HOUND OF THE BASKERVILLES

 

 


The Hound of the Baskervilles and the Simian Method

In The Hound of the Baskervilles, Conan Doyle gives us not the familiar pair of Holmes and Watson, but a less stable and more revealing trio: Sherlock Holmes, Dr. Watson, and Mr. Throttlebottom-Smith, an energetic capuchin monkey whose residence at Baker Street is never satisfactorily explained. Holmes refers to him only as “one of my irregulars, though of the arboreal division,” and leaves the matter there.

Watson dislikes him immediately. The feeling is mutual. Watson regards the monkey as an affront to the dignity of detection; Mr. Throttlebottom-Smith regards Watson as a slow rival with accessible waistcoat pockets. Holmes, who is amused by anything that produces evidence, encourages both just enough to make the rivalry permanent.

Mr. Throttlebottom-Smith and DAS KAPITAL

 


Karl Marx, Das Kapital, and the Unsettled Question of Mr. Throttlebottom-Smith

Karl Marx’s London years were not a quiet scholarly interval between insight and publication. They were decades of illness, debt, pawn tickets, family grief, smoke, boils, manuscripts, and books borrowed under conditions that would alarm a modern librarian. Into this crowded household came Mr. Throttlebottom-Smith, a Capuchin monkey of considerable energy, uncertain manners, and a master’s degree from the Columbia School of Journalism.

The degree was difficult to explain. Mr. Throttlebottom-Smith could not read, write, speak, or file copy. Yet the certificate in a small leather tube attached to his collar testified to his advanced professional training. Marx at first dismissed the matter as bourgeois credentialism in its purest form. With time, he softened. The monkey, Marx often observed, had an advantage over other journalists: it did not pretend to understand political economy.

Mr. Throttlebottom-Smith and DOWN AND OUT (Orwell)

 


George Orwell, the Manchester Times, and the Paris Adventures of Mr. Throttlebottom-Smith

George Orwell did not go to Paris merely as a young Englishman in search of poverty, authenticity, and copy. He went as a semi-official correspondent of the left-wing Manchester Times, a paper so advanced in its sympathies that it believed hunger should be investigated personally, though preferably by a junior contributor. As the contributor, Orwell's assignment was to report on the working poor of Paris: dishwashers, hotel drudges, Russian émigrés, and clerks fallen on hard times.

There was one editorial condition. Orwell was to be accompanied everywhere by Mr. Throttlebottom-Smith, an energetic Capuchin monkey on loan from the primate house of the Manchester Zoo. The arrangement was part journalism, part educational outreach, part circulation scheme. The newspaper had discovered modern promotion, and the zoo had discovered that a capuchin photographed better than a hungry socialist.

Friday, June 19, 2026

Essay: Pawlows Reflexe [German]

Pawlows Reflexe

Vokabeln unten


Iwan Petrowitsch Pawlow wurde 1849 in Rjasan geboren, einer Provinzstadt 200 Kilometer südöstlich von Moskau, die heute eine regionale Hauptstadt mit einer halben Million Einwohnern ist. Als Sohn eines Dorfpriesters wuchs Pawlow in einem Haushalt mit Kerzen und Ikonen, Suppen und Samowaren auf. Als Kind litt er an mehreren Krankheiten. Er war intensiv, aufmerksam, streitlustig und, laut einem Zeugnis der fünften Klasse, „stark durch Essen motiviert“.

Mit einundzwanzig trat Pawlow in die Universität von St. Petersburg ein, um Physik und Mathematik zu studieren. Doch bald übte die Biologie die stärkere Schwerkraft aus — und ersparte ihm jede weitere wirkliche Vorlesung über Schwerkraft. Er wechselte zur Medizin, studierte an der Kaiserlichen Medizinischen Akademie und wurde später Direktor eines physiologischen Labors. Seine frühen Arbeiten betrafen die Bauchspeicheldrüse und die Nerven des Herzens. Aber bald interessierte ihn vor allem die Verdauung vor der Verdauung: die Art, wie der Körper sich auf Essen vorbereitete, bevor das Essen überhaupt da war. Der Geruch von Brot, die Schritte eines Assistenten, das Klappern einer Schüssel konnten Magensaft und Speichel hervorrufen. Der Magen ... hatte ein Gedächtnis??

Aber der Weg zu seinen größten Experimenten war über viele Jahrzehnte mit Fehlschlägen gepflastert. Eine Pudelrasse erwies sich als zu melancholisch und starrte ihr Futter schweigend an wie ein kleiner russischer Dichter. Spaniels waren vielversprechend, aber theatralisch. Terrier reagierten nur, indem sie laut Blähungen machten.

Um 1903 herum, als Pawlow Anfang fünfzig war, kam endlich der Erfolg. Eine Klingel konnte den Körper davon überzeugen, dass das Mittagessen angekommen war.

Diese Entdeckung machte Pawlow berühmt, aber das brachte seine eigenen Gefahren mit sich.

Nur wenige wussten es, aber seit den 1880er Jahren entwickelte Pawlow eine strenge Mittagsgewohnheit. Jeden Tag aß er dieselbe Mahlzeit: einen Pirog und eine kleine Teetasse Borschtsch, aber erst nachdem die Mittagsglocken von St. Piotr geläutet hatten. Wenn ein junger Assistent zu früh mit einem Löffel klirrte, unterbrach Pawlow das ganze Mittagessen, sammelte sich wieder und saß dann regungslos da, während er auf die nächste Stunde wartete.

Bis 1899 scheuchte Pawlow manchmal die Techniker fort und bestand darauf, die Futterschüsseln der Hunde selbst zu waschen, oft ein halbes Dutzend Mal oder mehr, bis seine Hände gerötet waren.



Zur Zeit des Nobelpreises im Jahr 1904, der ihm für seine Arbeit über die Verdauung verliehen wurde, konnte Pawlow das Institut nur betreten, indem er vier Schritte vorwärts, sechs zurück, drei vorwärts, fünf zurück und acht vorwärts machte. Seine Mitarbeiter wurden Experten der Tarnung. Besuchende Journalisten wurden mit Tee aufgehalten. Fotografen wurden in Winkeln aufgestellt, die sowohl Pawlow als auch den Treppenstufen schmeichelten.

Nichts davon schadete Pawlows Ruf. Wenn überhaupt, vertiefte es die Aura um ihn. Wenn er selbst ein paar zusätzliche Reflexe besaß, nun ja, dann zeigte das berufliches Engagement. Niemand erwartet von einem Chemiker, dass er frei von Gerüchen ist, oder von einem Schmied, dass er andere als kräftige Hände hat.

Die Revolution von 1917 brachte jedoch Schwierigkeiten. Pawlow hatte einen lästigen Überschuss an Speichel entwickelt, wenn Türen geöffnet und geschlossen wurden. Ein junger bolschewistischer Chirurg, der erst kürzlich allein wegen seiner Begeisterung befördert worden war, deutete dieses Symptom als Tollwut. Pawlow wurde für dreißig Tage unter Quarantäne gestellt.

Seine letzten Monate verbrachte er in einem gepolsterten Zimmer, dessen Isolierung ihn vor dem Bellen der Hunde im Hof unten und vor der entfernten Glocke einer Straßenbahnlinie schützte. Er starb 1936 und wurde bald Gegenstand endloser Lobreden. Und irgendwo läutete weiter eine Glocke.


#

#
#
Streitlustig und, laut einem Zeugnis der fünften Klasse.  
Argumentative, according to a report in the fifth grade.
#
Schwerkraft ausüben   To make gravity (to exert gravity)
die Bauchspeicheldrüse   The pancreas
##
die Verdauung vor der Verdauung  
the digestion before the digestion
##
das Klappern einer Schüssel konnten Magensaft und Speichel hervorrufen
the clatter of a dish could summon gastric juices and saliva
>> See below:  Futterschüsseln
##
Jahrzehnte mit Fehlschlägen gepflastert.
Decades filled with failures.
##
laut Blähungen machten. Made loud [farts] [gas]
überzeugen  to convince [of something]
Mittagsgewohnheit   Lunch (lunchtime) habit
mit einem Löffel klirrte  Make noise (with spoon)
saß dann regungslos da  He sat motionless there
##
scheuchte Pawlow manchmal die Techniker fort; Futterschüsseln
Pavlov sometimes shooed the technicians away; Food-bowls
#
Hände gerötet waren.  Hands became red.
#
für seine Arbeit über die Verdauung verliehen wurde
was awarded for his work on digestion
##
schmeichelten.   Favor, flatter (in a photo)
Tarnung Camoflage
##
Pawlow hatte einen lästigen Überschuss an Speichel entwickelt
Pavlov had developed a troublesome excess of saliva
##
Tollwut.   Rabies.
##
in einem gepolsterten Zimmer  In a padded room
Lobreden   Eulogies (funeral speeches)



Thursday, June 18, 2026

Essay: Pavlov's Reflexes

Ivan Pavlov and the Sound of History


Ivan Petrovich Pavlov was born in 1849 in Ryazan, a provincial city 200 kilometers southeast of Moscow, and today a regional capital of half a million people. The son of a village priest, Pavlov grew up in a household of candles and icons, soups and samovars.  As a child he suffered from several illnesses. He was intense, observant, argumentative, and, according to a fifth-grade report, “highly food-motivated.”

At twenty-one, Pavlov entered the University of St. Petersburg to study physics and mathematics, but biology soon exerted the stronger gravitational pull (and avoided any actual lectures on gravity).  Switching to medicine, he trained at the Imperial Academy, and became director of a physiological laboratory. His early work concerned the pancreas and the nerves of the heart.  But soon, what most intrigued Pavlov was digestion before digestion: the way the body prepared for food before food arrived. The smell of bread, the footsteps of an attendant, the clatter of a dish could summon gastric juices and saliva.  The stomach...had a memory??

But the journey to his greatest experiments was littered with many decades of failures.  One breed of poodle proved too melancholic and would stare silently at its food like a minor Russian poet.  Spaniels were promising but theatrical. Terriers responded only by noisily passing gas.

By around 1903, when Pavlov was in his early fifties, success at last.  A buzzer could persuade the body that lunch had arrived.

This discovery made Pavlov famous, but that brought its own threats.

Few knew, but from the 1880s, Pavlov developed a strict luncheon habit. Every day he ate the same meal: one pyrogi and a small teacup of borscht, beginning only after the noon bells of St. Piotr’s had rung.  If a junior assistant clinked a spoon prematurely, Pavlov would stop the entire lunch, recompose himself, then sit motionless as he waited for the next hour.

By 1899, Pavlov sometimes shooed away the technicians and insisted on washing the dogs’ food dishes himself, often a half-dozen times or more, until his hands were reddened. 



By the time of the Nobel Prize in 1904, awarded for his work on digestion, Pavlov could enter the Institute only by taking four steps forward, six back, three forward, five back, and eight forward. His staff became expert in camouflage. Visiting journalists were delayed with tea. Photographers were positioned at angles flattering to both Pavlov and the stairsteps. 

None of this diminished Pavlov’s reputation. If anything, it deepened the aura around him. If he himself possessed a few extra reflexes, well, that showed professional commitment. No one expects a chemist to be free of smells or the blacksmith to have other than burly hands.

The Revolution of 1917, however, introduced difficulties.  Pavlov had developed a troublesome excess salivation in response to doors opening and closing. A young Bolshevik surgeon, recently promoted on the grounds of enthusiasm alone, interpreted this symptom as rabies. Pavlov was ordered quarantined for thirty days.

His last months were spent in a padded room whose insulation protected him from the barking of dogs in the courtyard below and from the distant bell of a tramline. He died in 1936, and was soon the subject of endless eulogies.  And somewhere, a bell kept ringing.



Monday, June 15, 2026

Understanding New Section SSA 1115(g) Budget Neutrality

 SSA 1115A created the CMMI, which can develop programs it imposes on Medicare (or Medicaid, duals, etc).   Older 1115 in contrast allows CMS to authorize waiver proposals that usually originate in state Medicaid programs.

1115 projects have been understood to be budget neutral, but recent legislation created 1115(g), more explicit and mandatory budget neutrality for Medicaid waivers, and inserting the CMS Actuary into the 1115 process.

Both SSA 1115(g) and SSA 1115A require budget-neutrality.  CMS has never further defined this for 1115A, but CMS will now define the phrase in great detail for 1115(g).  Whether there will be future cross-pollination in how CMS implements the same words "budget neutrality" in 1115 and in 1115A is TBD.

##

CMS press release on new 1115(g) implementation

https://www.cms.gov/newsroom/press-releases/cms-takes-bold-new-approach-stewarding-medicaid-demonstration-project-spending

Medicaid.gov neutrality page

https://www.medicaid.gov/medicaid/section-1115-demonstrations/budget-neutrality

Essential Hospitals info

https://essentialhospitals.org/cms-publishes-section-1115-medicaid-demonstration-budget-neutrality-guidance/

Healthcare Dive

https://www.healthcaredive.com/news/cms-increases-oversight-of-state-medicaid-demonstrations/822778/

##

##

You’re right to see two different “1115” worlds colliding. The key is that SSA §1115(a) (little a) is the old state Medicaid waiver/demonstration authority, while SSA §1115A (big A) is a separate ACA-created CMMI statute. The new development is not in §1115A; it is Public Law 119-21, §71118, which added a new §1115(g) budget-neutrality rule for Medicaid §1115 demonstrations. CMS calls that law the “Working Families Tax Cut” legislation, and CMS is now implementing it for approvals on or after January 1, 2027. (GovInfo)

1. SSA §1115: state Medicaid waiver / demonstration authority

Section 1115 is a broad demonstration statute. For Medicaid, it lets HHS/CMS approve state experimental, pilot, or demonstration projects that are likely to promote Medicaid objectives. It has two classic tools: CMS may waive certain Medicaid state-plan requirements, especially §1902 requirements, and may treat otherwise non-matchable demonstration costs as federally matchable Medicaid expenditures. (Social Security)

In practical Medicaid language, §1115 is where states go when they want a Medicaid waiver: eligibility changes, benefit experiments, delivery-system changes, managed-care structures, reentry services, HRSN/SDOH services, uncompensated care pools, or other Medicaid policy experiments. CMS reviews these state-by-state and historically attaches Special Terms and Conditions, including budget-neutrality terms. (Medicaid)

2. SSA §1115A: CMMI, not a state Medicaid waiver statute

Section 1115A was added by the ACA to create the Center for Medicare and Medicaid Innovation. Its purpose is to test payment and service-delivery models that reduce expenditures under Medicare, Medicaid, or both, while preserving or enhancing quality. Its “applicable individual” definition includes Medicare beneficiaries, Medicaid beneficiaries, and dual eligibles; its “applicable title” means Title XVIII, Title XIX, or both. (Social Security)

So yes, Medicaid appears in §1115A. But that does not make CMMI the same as a state §1115 Medicaid waiver. CMMI is a CMS model-testing platform. It may test Medicaid-focused models, dual-eligible models, all-payer models, and models “primarily on title XIX” in conjunction with CMCS, but it is structurally a federal Innovation Center model authority, not the ordinary state waiver pathway. (Social Security)

A helpful shorthand:

Feature§1115 Medicaid demonstrations§1115A CMMI
Basic thingState Medicaid waiver / demonstrationFederal Innovation Center model
Main applicant / actorState Medicaid agency, negotiated with CMSCMS/CMMI model, often with providers, payers, states, or plans
Program scopeMedicaid Title XIX demonstrations; sometimes CHIP under other provisionsMedicare, Medicaid, duals, and sometimes CHIP-like activities
Waiver powerBroad Medicaid waiver and expenditure authorityModel-specific, narrower waiver authority tied to testing
Budget testNow being codified and tightened in §1115(g)Built into ACA §1115A, but differently

3. Were Medicaid §1115 waivers already budget neutral?

Yes, but with an important caveat: they were budget neutral by CMS policy and negotiated methodology, not by an explicit Medicaid §1115 statutory certification rule. CMS says it has long generally not approved §1115 Medicaid demonstrations unless expected to be budget neutral, using “without waiver” and “with waiver” calculations, spending limits, and post-period reconciliation. (Medicaid)

That is why the old statement “1115 waivers must be budget neutral” was true in practice. But it hid a lot of policy discretion. CMS and states negotiated the counterfactual: what would Medicaid have spent without the waiver? That counterfactual could depend on base years, trend rates, rebasing, savings rollover, and “hypothetical” expenditures. GAO described this as a system where CMS policy required budget neutrality, but the methods were not defined in statute or regulation and could substantially affect federal spending limits. (GAO Files)

That is the answer to the puzzled reader: 1115 projects were “budget neutral,” but “budget neutral” meant “within a CMS-negotiated counterfactual cap.” The new law tries to make it more like an independent, actuarial, pre-approval test.

4. What changed in §1115?

The legal change is P.L. 119-21 §71118, enacted in 2025. It added new SSA §1115(g). Beginning January 1, 2027, the Secretary may not approve a new Medicaid §1115 demonstration, renewal, or amendment unless the CMS Chief Actuary certifies that it is not expected to increase federal Medicaid expenditures compared with expenditures absent the demonstration. (GovInfo)

CMS’s June 2026 SMD letter says this is a major change because, before §1115(g), the Chief Actuary did not review or certify Medicaid §1115 budget neutrality, and the current application data and budget-neutrality test are not sufficient for actuarial certification. (Medicaid)

The biggest operational changes are:

  1. Prospective certification. Old method: preliminary budget-neutrality estimate, then retrospective assessment against WOW/WW limits. New method: Chief Actuary certification before approval, amendment, or renewal. (Medicaid)

  2. No classic expenditure cap as the core test. CMS’s table says the new approach would not use the old budget-neutrality expenditure limit or cap; instead, demonstrations projected to increase federal Medicaid spending would not be approved. (Medicaid)

  3. Activity-by-activity analysis. States must classify and analyze individual demonstration activities, especially activities that can only be authorized under §1115. CMS expects actuarial, economic, statistical, or similarly rigorous analysis. (Medicaid)

  4. MAPS replaces the old “hypothetical expenditure” concept. CMS will classify activities as either Medicaid Authorizable Populations and Services — populations/services the state could otherwise cover under the Medicaid state plan or other Title XIX authority — or §1115-only activities. The §1115-only activities drive the budget-neutrality calculation. (Medicaid)

  5. Less room for manufactured savings. Under the new table, MAPS activities do not generate savings; savings rollover is limited to the current demonstration period plus prior-period savings limited to five years, rather than the older ten-year legacy-savings concept. (Medicaid)

  6. Administrative costs count. CMS reads §1115(g)’s “federal expenditures” language to include administrative costs caused by the demonstration, unlike the current approach. (Medicaid)

GAO’s 2025 report helps explain the policy concern behind the change. GAO recommended that CMS stop treating costs for populations or services that could not otherwise have been covered under existing Medicaid authorities as “hypothetical” when setting spending limits, and instead require those costs to be offset by other reductions in demonstration spending. (GAO)








5. How this differs from CMMI §1115A budget neutrality

Here is the really important contrast: CMMI is not required to be budget neutral at the initial testing stage. Section 1115A expressly says the Secretary shall not require, as a condition for testing a model, that the model be budget neutral initially. But after testing begins, the Secretary must terminate or modify the model unless the Secretary determines, and the CMS Chief Actuary certifies, that it is expected to improve quality without increasing spending, reduce spending without reducing quality, or both improve quality and reduce spending. (Social Security)

For expansion of a CMMI model, the statute is stricter: the Secretary may expand a model only if it is expected to reduce spending without harming quality or improve quality without increasing spending; the Chief Actuary must certify that expansion would reduce or not increase net program spending; and the Secretary must determine that expansion would not deny or limit coverage or benefits. (Social Security)

So §1115A has always had statutory spending safeguards, but not in the same form as the new §1115(g). In fact, at the initial test stage, §1115A is more permissive than the new §1115(g). CMMI may test a model that is not initially budget neutral; Medicaid §1115(g), by contrast, will block approval of a Medicaid demonstration, amendment, or renewal unless the Chief Actuary certifies no expected federal Medicaid spending increase.

6. The clean way to explain it

For a puzzled reader, I would say:

Old Medicaid §1115 budget neutrality was a CMS policy discipline built around negotiated counterfactual spending limits. New §1115(g) turns that discipline into a statutory, Chief-Actuary-certified, pre-approval gate.

CMMI §1115A is different. It is an ACA Innovation Center statute for testing payment and delivery models in Medicare, Medicaid, or both. Its initial models need not be budget neutral at launch, but they must be evaluated, terminated or modified if they fail the statutory spending/quality tests, and actuarially certified before expansion.

The shared word “Medicaid” is the source of confusion. §1115 is Medicaid waiver law. §1115A is Innovation Center model law that can include Medicaid. They overlap in subject matter, but they are not the same pathway, and the new 2026 CMS activity is about §1115(g), not §1115A.

##

##

1115A CMMI demonstrations are imposed by CMS management on the Medicare program, and I suppose when relevant, on the Medicaid (or duals) program. But 1115 demonstrations tend to originate IN the states and are brought to CMS for endorsement. Reasonable summary?

##

Yes — that is a reasonable high-level summary, with one word softened.

I would say:

§1115A / CMMI demonstrations are generally CMS-originated payment or delivery models, selected and designed by the Innovation Center, and then applied to Medicare, Medicaid, dual-eligible, or multi-payer settings depending on the model. Some are voluntary; some Medicare models have been mandatory for selected providers or geographies.

§1115 Medicaid demonstrations, by contrast, usually originate with a state Medicaid agency. The state asks CMS for waiver and/or expenditure authority, and CMS approves, denies, negotiates, or conditions the project through Special Terms and Conditions.

The statutory language tracks that distinction. §1115A creates a Center inside CMS whose purpose is to “test innovative payment and service delivery models” under Medicare, Medicaid, or both; it defines applicable individuals to include Medicare beneficiaries, Medicaid beneficiaries, and duals.

By contrast, Medicaid §1115 is framed around a project “in a State or States,” under which the Secretary may waive state-plan requirements and treat otherwise non-matchable costs as expenditures under the state plan. CMS’s 2026 SMD letter also describes §1115 as giving states flexibility to design and improve their Medicaid programs, with CMS doing case-by-case review of each state proposal.

The caveat is that CMMI is not always simply “imposed.” Many CMMI models are voluntary or contractual. But the center of gravity is federal: CMS designs the model, chooses the model parameters, selects participants or regions, evaluates the model, and may waive certain Medicare/Medicaid rules solely to carry it out. For Medicaid-heavy or duals models, states often have to be partners, because Medicaid is jointly federal-state and operationally state-run.

So the clean formulation is:

CMMI §1115A is top-down federal model testing, sometimes with state or provider participation. Medicaid §1115 is bottom-up state waiver innovation, subject to CMS approval and federal guardrails.


##

It sounds like the CMS Actuary for SSA 1115 Medicaid waivers, as enhanced by recent SSA 1115(g), makes the CMS Actuary a bit like the Senate Parliamentarian, standing at the gate with a red light and a green light. He's not always Mr. Popular.

##

Yes — that’s a good analogy, with two refinements.

The CMS Chief Actuary becomes a gatekeeper, but only for one gate: federal Medicaid budget neutrality. Under new §1115(g), CMS says that beginning January 1, 2027, the Secretary may not approve a Medicaid §1115 application, amendment, or renewal unless the Chief Actuary certifies that it is not expected to increase federal Medicaid expenditures versus the no-waiver world. CMS also says the certification must be prospective and based on actuarial methods, not merely the older negotiated “with waiver/without waiver” bookkeeping.

So the “Senate Parliamentarian” analogy works in this sense: the actuary is not the politician, not the policy entrepreneur, and not the advocate. He or she stands at the procedural/technical gate and says, in effect, this proposal passes the required test or this proposal does not pass the required test.

But the analogy is not perfect. The Senate Parliamentarian is mainly interpreting chamber rules and reconciliation constraints. The CMS Chief Actuary is making a professional actuarial certification: are the assumptions, data, counterfactuals, activity classifications, administrative costs, and claimed savings credible enough to support “no expected federal spending increase”? CMS describes this as an independent actuarial determination and attestation based on sound methods and reasonable assumptions.

That is why the role can become unpopular. A state may have a politically attractive waiver; providers may love it; CMS leadership may favor the policy direction; advocates may say it improves access. But if the actuarial case does not work, the actuary’s answer can be: no certification, no approval.

In one respect, the §1115(g) actuary may be even more of a hard gate than the Parliamentarian. A Parliamentarian’s advice is institutionally powerful but formally advisory to the Senate’s presiding officer. Here, CMS is reading the statute as a legal prerequisite: without the Chief Actuary’s certification, the Secretary may not approve the Medicaid demonstration.

The crucial nuance is that the actuary does not decide whether the waiver is good Medicaid policy. CMS still says the Secretary must separately determine that the demonstration promotes Medicaid objectives. The actuary’s role is narrower but newly decisive: the fiscal green light.

For CMMI §1115A, the Chief Actuary also matters, but differently. CMMI models do not have to be budget neutral at initial testing; after testing begins, and especially before expansion, the statute brings in spending/quality findings and actuarial certification. For Medicaid §1115 after §1115(g), the actuary is moved much farther forward in the process — from a background fiscal discipline to a front-end statutory gate.

##

##

SSA §1115 is much older than the ACA. It comes from the Public Welfare Amendments of 1962, Public Law 87-543, signed by President Kennedy on July 25, 1962. It was created as a general demonstration authority for federal-state public welfare programs — not originally as a Medicaid-specific authority, because Medicaid did not exist yet. (NCBI)

Then, in 1965, Congress created Medicare and Medicaid in the Social Security Amendments of 1965, Public Law 89-97. Medicaid became Title XIX of the Social Security Act, and §1115 was amended to include Title XIX demonstration authority. (GovInfo)

So the chronology is:

1935 — Social Security Act enacted.
1962 — §1115 demonstration authority added for public welfare programs.
1965 — Medicare and Medicaid enacted; Medicaid added as Title XIX; §1115 becomes available for Medicaid demonstrations.
2010 — ACA adds §1115A, creating CMMI.

The conceptual lineage is also different. §1115 was a Kennedy-era welfare-program flexibility tool, meant to let states test experimental or pilot approaches in federal-state assistance programs. §1115A was an ACA-era payment and delivery reform tool, meant to let CMS test Medicare/Medicaid innovation models through CMMI.

##

##

Here is all of SSA §1115(g) as currently shown in the preliminary U.S. Code version of 42 U.S.C. §1315(g). It was added by Pub. L. 119-21, §71118(a), July 4, 2025, 139 Stat. 306. (U.S. Code)

(g) Requirement of budget neutrality for Medicaid demonstration projects

(1) In general
Beginning January 1 2027, the Secretary may not approve an application for (or renewal or amendment of) an experimental, pilot, or demonstration project undertaken under subsection (a) to promote the objectives of subchapter XIX in a State (in this subsection referred to as a “Medicaid demonstration project”) unless the Chief Actuary for the Centers for Medicare & Medicaid Services certifies that such project, or, in the case of a renewal, the duration of the preceding waiver, is not expected to result in an increase in the amount of Federal expenditures compared to the amount that such expenditures would otherwise be in the absence of such project.

For purposes of this subsection, expenditures for the coverage of populations and services that the State could have otherwise provided through its Medicaid State plan or other authority under subchapter XIX, including expenditures that could be made under such authority but for the provision of such services at a different site of service than authorized under such State plan or other authority, shall be considered expenditures in the absence of such a project.

(2) Treatment of savings
In the event that expenditures with respect to a State under a Medicaid demonstration project are, during an approval period for such project, less than the amount of such expenditures that would have otherwise been made in the absence of such project, the Secretary shall specify the methodology to be used with respect to the subsequent approval period for such project for purposes of taking the difference between such expenditures into account.

Note the odd-looking “January 1 2027” without a comma. That is how it appears in both the enacted Public Law text and the preliminary U.S. Code display.