vbcc wedn x 3

 

The titles below follow the agenda. Sidebar items are paraphrases, with transcript timestamps for reference. “Brooks” means CMS official John Brooks, distinct from Friday panelist Gabriel Brooks.

CMS Innovative Initiatives: Providers on OCM, EOM, and What’s Up Next

The provider panel presented a sharp disagreement over whether EOM advances oncology value-based care or constrains it. Several participants argued that narrowing eligibility to seven cancers and concentrating on active treatment creates financial volatility, excludes valuable decisions not to treat, and limits investment across the full cancer journey. Ron Kline defended population-level accountability: practices must evaluate aggregate performance rather than fixate on individual losing cases, and excluding drugs would remove major opportunities for savings. Yet the discussion exposed how poorly claims capture clinical reality—stage, tumor biology, treatment intent, and whether someone is receiving palliation or simply completing survivorship follow-up. Biomarker results buried in incompatible PDF reports were a particularly concrete measurement obstacle. Participants praised ePROs, social-needs services, navigation, and commercial arrangements combining total-cost accountability with quality requirements, while emphasizing their infrastructure costs. The unresolved dispute was whether better voluntary economics could attract practices or whether mandatory participation would be necessary. For Friday’s panel, the central lesson is that richer clinical measurement could address important blind spots, but financial incentives, population definitions, and implementation support remain equally consequential. 20261007 Panel All About CMMI O…

CMS Leadership Discussion with Abe Sutton

Abe Sutton described CMMI as CMS’s research-and-development arm for payment, emphasizing evidence, scalability, and willingness to abandon disappointing models—including a kidney model he helped design. Home dialysis improved nationally, but mandated model participants did not improve significantly more than comparison areas: favorable trends alone did not demonstrate model success. He explained certification as evidence of lower spending without worse quality, or better quality without higher spending, while noting that certification does not automatically compel nationwide expansion. He discussed joint-replacement expansion alongside continued testing of TEAM, defended WISeR’s technology-assisted review of selected services, and promised an accelerated early evaluation rather than waiting for the usual reporting cycle. Drug-policy discussions covered GLP-1 access, GLOBE’s international-reference approach, and GENEROUS’s pooled Medicaid rebate and coverage arrangements. Sutton welcomed outside proposals that change incentives across a clinical category or market, rather than obtain special treatment for one product. For oncology, he was open to new arrangements and potentially helpful legal flexibilities, but rejected simply increasing payments without accountability. He also acknowledged that gaps between models can dismantle staffing and infrastructure that practices have already built. 20261007 Joe grogan interviews …

CMS Leadership Discussion with John Brooks

John Brooks framed value-based care as aligning treatment incentives with the care patients would want, rather than rewarding visit volume or the margin on a particular drug. His broad policy discussion covered Part D stabilization after the IRA, Medicare Advantage’s benefits and unresolved payment and quality concerns, prior-authorization standardization, and greater transparency around 340B. He stressed that CMS cannot simply compensate providers for statutory reimbursement changes enacted by Congress. For AI, his position was simultaneously optimistic and skeptical: current applications can increase spending by amplifying existing incentives, while appropriately designed outcome-based payment could make technology reduce costs and expand access. He highlighted ACCESS as a promising framework for paying for measurable clinical improvement rather than adding another billable service. Outcomes-based drug contracts, he argued, primarily allocate uncertainty about therapeutic effectiveness between manufacturers and payers. Brooks welcomed validated data and proposals jointly supported by providers, payers, and manufacturers. His memorable warning was that consultants can sell companies months of expensive work pursuing a bespoke CMMI demonstration, only for CMS to reject it: the agency wants structural solutions that can work across a market. 20261007 joe grogan intvu brook…

Sidebar: 20 Takeaways for Friday’s AI and Oncology Measurement Panel

1. A narrower model can produce noisier results. EOM’s restricted cancer mix and focus on early treatment can make practice-level performance swing with a few expensive cases, even when those fluctuations wash out nationally. Better measurement must distinguish performance from case-mix volatility. (CMMI provider panel; 8:18–10:24)
2. Some valuable care never enters the denominator. A lengthy discussion leading a patient to decline burdensome treatment can save money and improve goal-concordant care, yet fail to trigger a treatment-based episode. This directly echoes the “missing denominator” issue in your background notes. (CMMI provider panel; 10:24–11:26, 44:21–46:13)
3. What two clinicians understand immediately may be invisible in claims. Kline recalled trying to distinguish methotrexate for rheumatologic disease from cancer treatment, and palliation from survivorship follow-up. AI’s potential contribution is extracting the clinical context needed to interpret the transaction. (CMMI provider panel; 13:37–16:58)
4. The biomarker report is there—but the measurement system cannot readily use it. Kline described opening a laboratory PDF at Johns Hopkins to discover which biomarkers had been tested, then observed: “There’s no way you can do quality measures with that.” A concrete challenge for AI extraction and interoperability. (CMMI provider panel; 36:14–38:10)
5. ePROs and social-needs screening create services, not merely data. Henschel credited these requirements with timely interventions and investments in meals, transportation, social work, and palliative care. Their value depends on the response infrastructure; collecting information alone does not deliver the benefit. (CMMI provider panel; 33:01–35:36)
6. Quality can be an actual condition of payment. Florida Cancer Specialists described a longstanding total-cost arrangement in which meeting a quality gate unlocks shared savings. The practice itself proposed the gate. Friday’s question: which AI-enabled measures would be trustworthy enough to play that role? (CMMI provider panel; 20:23–21:15)
7. An oncologist cannot accept a pen—but can earn thousands in drug margin. Thurmes illustrated the incentive mismatch with Minnesota’s strict gift restrictions. His point was that conscientious physicians still work inside an economic structure that rewards drug revenue. Measurement reform must confront that structure. (CMMI provider panel; 24:08–26:05)
8. Low participation has competing explanations—and competing remedies. Kline argued that voluntary models struggle against financially attractive fee-for-service and may require mandatory participation. Mehring countered that a model benefiting all stakeholders would attract participants voluntarily. Better measures alone will not settle this dispute. (CMMI provider panel; 37:22–42:40)
9. The expensive consultant engagement ends with “no.” Brooks joked that consultants can convince companies CMMI will create a special model for them; months of paid work culminate in a meeting where CMMI says it has no interest. His requirement: a replicable, structural proposal. (Brooks; 39:27–40:33)
10. AI can make an inefficient payment system more expensive. Brooks’s near-term assessment was that AI has been cost additive: helping the system do more can amplify its existing inefficiencies. An AI tool’s productivity gains therefore do not automatically translate into payer savings. (Brooks; 41:35–43:20)
11. Pay for the clinical improvement the technology produces. Brooks highlighted ACCESS and the example of lowering HbA1c. For oncology, the corresponding challenge is to identify meaningful, auditable outcomes that could become reimbursement targets, rather than simply adding payment for using an algorithm. (Brooks; 2:31–3:44, 41:35–43:20)
12. Physician compensation should not depend on how expensive the prescribed drug is. Brooks called the existing arrangement a deeply unsatisfactory system, while acknowledging how difficult it is to change. New oncology measures will operate within those prescribing incentives unless payment design changes too. (Brooks; 33:13–35:16)
13. A drug contract reallocates risk; it does not change the molecule’s behavior. Brooks distinguished provider incentives from outcomes-based pharmaceutical agreements. The latter distribute uncertainty about effectiveness between payer and manufacturer—making reliable outcome ascertainment central to whether the contract works. (Brooks; 37:30–38:48)
14. Bring validated data and a solution the stakeholders developed together. Brooks said joint proposals from providers, payers, and manufacturers cut through competing requests for money. For Friday’s panel, a shared clinical measure with demonstrated usefulness would be more persuasive than another constituency’s wish list. (Brooks; 53:00–54:42)
15. An improving metric does not prove the payment model worked. Sutton’s kidney example was explicit: home dialysis rose nationally, but model areas did not outperform comparison areas significantly. This reinforces your speaker notes’ emphasis on a credible counterfactual when evaluating AI or payment interventions. (Sutton; 5:11–6:30)
16. CMMI’s success test has two routes. Sutton described certification as improved quality without higher costs, or lower costs without reduced quality. That distinction matters for oncology: a useful AI intervention need not demonstrate both better outcomes and lower spending simultaneously. (Sutton; 10:55–11:23)
17. Evaluation can arrive too late to guide decisions. Sutton contrasted the usual annual evaluation cycle with WISeR’s planned early snapshot using six months of experience and three months of claims runout. Friday’s discussion could distinguish rapid operational feedback from mature evidence sufficient to judge success. (Sutton; 22:47–24:47)
18. A gap between models can erase the investment the first model created. Sutton described kidney programs losing the ability to support staff during a transition gap. Oncology measurement infrastructure—including ePRO support and clinical-data capabilities—also needs a financing path that survives model transitions. (Sutton; 44:05–46:11)
19. CMMI wants an open market, not a privileged product. Sutton described ACCESS as a reimbursement structure in which multiple companies can compete, rather than a demonstration tailored to one solution. An oncology proposal should define the clinical problem and payment framework broadly enough for multiple approaches. (Sutton; 38:52–41:24)
20. The door is open for another oncology model—with conditions. Sutton welcomed proposals that improve incentives and support patient care, while rejecting a simple payment increase without downside accountability. He also expressed willingness to consider legal flexibilities for contracting, within statutory purposes. (Sutton; 42:17–43:05, 47:26–48:06)

 

LONG FORMAT SUMMARIES

 

CMS and Oncology Value Based Care Session Summaries

Prepared for Bruce Quinn | October 7, 2026

These three sessions examine oncology payment models, CMS policy priorities, and the clinical information needed to evaluate value. Their recurring concern is how to connect meaningful patient outcomes with payment incentives while preserving practices' ability to deliver care.

CMS Innovative Initiatives: OCM, EOM, and What’s Up Next

Wednesday, October 7, 2026 | Day One | Session 15 | 11:00–11:50 a.m.

Location: Olympic Suite 1, 10th Floor

Moderator: Maddi Davidson, Managing Director of Market Access, Avalere Health

Panelists:

Paul Thurmes, MD, President and Medical Oncologist, Minnesota Oncology

Gabrielle Rocque, MD, MS, Chief Medical Officer, Atlas Oncology Partners; Associate Professor, University of Alabama at Birmingham

Rhonda Henschel, MBA, SVP, Payer Lifecycle and Value Optimization, McKesson

Ron Kline, MD, Retired CMO, Quality Measurement and Value-based Incentives Group, CCSQ/CMS

Kiana Mehring, MBA, PPMC, LION, VP of Payer Strategy & Revenue Cycle, Florida Cancer Specialists & Research Institute

Participants and central disagreement

Moderated by Maddi Davidson, this session brought together Paul Thurmes, Gabrielle Rocque, Rhonda Henschel, Ron Kline, and Kiana Mehring. The discussion exposed a fundamental disagreement about EOM: several practice leaders saw its narrower design as a retreat from OCM, while Kline defended the logic of population-level financial accountability. Participants agreed that modern oncology has changed dramatically, but differed over whether existing payment models recognize those changes adequately.

Thurmes emphasized that immunotherapies and targeted treatments complicate judgments about benefit, treatment sequence, and when to stop. Kline maintained that the definition of value remains outcomes relative to cost, including survival and quality of life. Mehring cautioned that programs have nevertheless drifted toward treating reduced expenditure as the principal evidence of value.

The population and denominator problem

Rocque argued for encompassing the whole cancer population, from diagnosis through survivorship and end of life. Restricting EOM to seven cancers and early treatment concentrates drug spending and reduces the population available to absorb financial variation. Mehring said Florida Cancer Specialists' modeling predicted losses under EOM, but a favorable result when the broader OCM cancer mix was restored. Henschel reported substantial swings across practices and performance periods within the US Oncology Network.

Kline responded that practices must assess their aggregate results, accepting gains on some patients and losses on others. He recalled repeated complaints about losses on oral myeloma therapies without equivalent attention to gains elsewhere. The disagreement concerned whether EOM's overall financial design was viable, as well as practitioners' tendency to focus on unfavorable subgroups.

Thurmes identified another omission: patients who decline systemic treatment after extensive discussion may receive valuable supportive care without entering a treatment-triggered episode. An audience member similarly argued that observation or delayed treatment can save substantial spending without earning model credit. Recognizing these patients reliably became a recurring challenge.

Claims and biomarker information

Kline described the gulf between clinical understanding and claims analysis. Two clinicians can readily distinguish methotrexate for cancer from methotrexate for rheumatologic disease; a claims-based program has much less information. Likewise, a breast-cancer diagnosis can represent active advanced disease, palliation, or long-term survivorship. He criticized legacy coding's ability to specify anatomical location more readily than clinically decisive tumor biology.

Mehring argued that practices and payers could exchange stage and other clinical data to improve their models. Participants differed over whether CMS was sufficiently responsive to practice feedback: Henschel said CMMI had been accessible, but stakeholders often reached different conclusions because they examined different data.

Biomarker reporting supplied a particularly vivid example. Thurmes described inconsistent testing platforms and uncertainty about repeat testing. Kline recalled opening laboratory PDFs at Johns Hopkins to determine which biomarkers had been assessed. Incompatible reporting formats obstruct routine quality measurement even when the underlying clinical information exists.

Drugs and financial incentives

Kline opposed removing drugs from total-cost accountability because biosimilars, generics, and equivalent guideline-supported options offer important savings opportunities. Rocque agreed that drugs belong in the calculation, while warning that supportive-care interventions cannot solve broader pharmaceutical pricing problems.

Thurmes illustrated the prescribing incentive problem: Minnesota restrictions prevented him from accepting even a small industry gift, yet practices could receive substantial drug margins. Henschel warned that simultaneous reimbursement changes, IRA implementation, and overlapping reporting requirements make participation harder and obscure which policy causes which result.

Services worth preserving

Henschel strongly supported ePROs and health-related social-needs screening, citing earlier intervention and investments in transportation, meals, social work, and palliative care. These services require substantial financial support. Once established, practices may resist dismantling them, but that does not mean new participants can build them without funding.

Mehring described a longstanding commercial total-cost arrangement with a quality gate controlling access to shared savings. Rocque's organization instead assumed downside risk while insulating participating practices. These examples demonstrated different ways to support care transformation.

Participation and relevance to Friday

Kline argued that attractive fee-for-service economics limit voluntary enrollment and that mandatory models may be necessary. Mehring countered that mutually beneficial models can attract participation while saving payers money. For Friday's panel, the session supplies concrete measurement targets: treatment intent, decisions not to treat, tumor biology, symptoms, and the full disease trajectory. It also makes clear that better data must be accompanied by workable incentives and infrastructure financing.

 

 

CMS Leadership Discussion with Abe Sutton

Wednesday, October 7, 2026 | Day One | Session 17 | 1:00–1:50 p.m.

Location: Lounge, 9th Floor

Moderator: Joe Grogan, JD, Fellow, USC Schaeffer Center; Former Assistant to the President, Director of the Domestic Policy Council

Featured speaker: Abe Sutton, JD, Deputy Administrator & Director, CMS Innovation Center, Centers for Medicare & Medicaid Services

CMMI's role and the administration's approach

Interviewed by Joe Grogan, Abe Sutton described CMMI as CMS's research-and-development arm for payment. Its task is to test whether different incentives can improve care and make public spending more efficient. He contrasted initial ambivalence about an Affordable Care Act institution during the first Trump administration with the current administration's willingness to use its authority actively. The emphasis was on evidence-driven decisions rather than preserving models because an administration had created them.

Sutton said the team reviewed the existing portfolio and canceled four models, three originating in the first Trump administration. Career staff supplied recommendations. One canceled initiative was a kidney model Sutton had helped develop, making the discussion an unusually direct acknowledgment that a favored policy had not met expectations.

Improvement versus attributable improvement

The kidney example illustrated his evaluation standard. National home-dialysis rates rose from approximately 11% when the model was designed to approximately 16%. However, mandated model areas did not show a statistically significant improvement over comparison areas. Other policy changes, technology, and evolving clinical practice apparently contributed to the national trend.

Sutton argued that static performance targets could consequently reward improvement that the model itself had not generated. The relevant question was whether the payment intervention changed behavior beyond what would otherwise have occurred. This is especially pertinent to evaluating AI in settings where clinical practice and technology are already improving.

Certification and nationwide expansion

Sutton described certification as demonstrating better quality without higher spending, or lower spending without worse quality. CMS's actuarial and clinical assessments inform the secretary's decision. Certification permits consideration of expansion but does not automatically require it.

He used joint replacement to explain why CMS might expand a successful approach while continuing to test an alternative. CJR and TEAM use different episode durations and designs; future evidence could favor the newer framework. He also noted that a model's relationship to other payment programs matters. Apparent savings may be less attractive if inconsistent benchmarks create opportunities for arbitrage between programs.

WISeR and technology-assisted review

A substantial portion of the interview defended WISeR, which applies technology-assisted review to selected services in six states. Sutton emphasized that the model enforces existing coverage standards rather than introducing new ones. Providers can submit for prior authorization or face prepayment review; he reported that most chose authorization.

He described participant incentives intended to discourage inappropriate denials: repeat denials do not generate repeated rewards, eventual approval can eliminate a claimed saving, and excessive reversals can trigger penalties. He said denials require physician review and argued that rapid approvals and accessible clinical discussion distinguish the process from familiar authorization frustrations.

Grogan pressed him about congressional opposition and reported implementation friction. Sutton acknowledged early problems, particularly in Washington State, but defended the model as protecting patients from inappropriate care as well as taxpayers from waste. He promised an early evaluation snapshot using six months of experience and three months of runout. His favorable assessment remained an account of implementation, ahead of that evaluation.

Drug access and purchasing arrangements

Sutton described balancing broader GLP-1 access against substantial near-term costs, referring to a bridge demonstration and an announced price of $245 per month. On GLOBE, he explained using the IRA's inflation-rebate structure to test international price references. He contrasted its manufacturer-focused design with earlier approaches that also contemplated changing physician payment.

GENEROUS, in Medicaid, offers standardized supplemental rebates alongside standardized coverage terms. States can compare that offer with their existing arrangements. Sutton said manufacturer participation exceeded initial expectations and presented pooled contracting as a way to reduce the burden of negotiating separately with many states. He drew a parallel with the cell-and-gene-therapy model.

What oncology proposals could succeed

Sutton welcomed outside ideas backed by persuasive data, including proposals from academic researchers. He wanted approaches that could reshape a clinical category or create a competitive market. A request to bypass ordinary coverage review for one device would not fit that purpose.

For oncology, he was open to new arrangements but rejected simply increasing payments without meaningful accountability. He also indicated willingness to consider contracting flexibilities within statutory limits. Finally, he warned that gaps between models can destroy staffing investments and make subsequent recruitment harder. For Friday, his remarks connect clinical measurement to causal evaluation, scalable payment design, and continuity of the infrastructure needed to act on the data.

 

 

CMS Leadership Discussion with John Brooks

Wednesday, October 7, 2026 | Day One | Session 21 | 2:00–2:50 p.m.

Location: Lounge, 9th Floor

Moderator: Joe Grogan, JD, Fellow, USC Schaeffer Center; Former Assistant to the President, Director of the Domestic Policy Council

Featured speaker: John Brooks, JD, MBA, Chief Policy and Regulatory Officer, Deputy Administrator, Centers for Medicare & Medicaid Services

Patient interests and payment incentives

Interviewed by Joe Grogan, John Brooks framed value-based care around the treatment incentives patients would want their clinicians to face. He questioned a system in which drug margins influence prescribing and visit-based payment can reward additional utilization. His goal was to align payment with patient benefit while recognizing that changing an established reimbursement structure is difficult.

Brooks saw new opportunities in digital therapeutics and AI that were less developed during his previous government service. Such tools could extend access at low marginal cost, but their economic effect depends on payment design. He described the administration as having established a policy direction toward outcomes, with substantial implementation and evaluation still ahead.

Part D and Medicare Advantage

Brooks characterized Part D as adapting to the IRA's redesigned benefit. His assessment combined lower beneficiary financial exposure, higher taxpayer costs, fewer plan choices, and a more stable market. He defended reducing additional premium-stabilization support after CMS examined plan bids, rather than making that decision without market evidence. He also identified a future premium-cap issue that would require congressional attention.

On Medicare Advantage, he acknowledged a crisis of confidence involving quality measurement, coding intensity, and selection. Nevertheless, he emphasized its additional benefits and appeal to beneficiaries seeking affordable coverage. He described collaboration between CMS and MedPAC to reconcile coding-related payment estimates, noting that different data years and assumptions explained some apparent disagreement. The comparison remained complicated by differences between the enrolled populations.

Prior authorization and data exchange

Brooks described insurer commitments to reduce and harmonize services subject to authorization, standardize documentation, and improve electronic processing. He reported an initial reduction in the number of procedures requiring authorization, while emphasizing continued work toward faster decisions.

Large health systems and insurers can sometimes achieve nearly instantaneous processing because they possess the necessary resources. Smaller practices need access to comparable capabilities. Interoperability, investment, and longstanding distrust between parties remain obstacles. Brooks's objective was timely exchange of the information necessary to establish that a patient should receive the requested care.

340B and provider reimbursement

Brooks defended examining hospital acquisition costs for 340B drugs and reconsidering reimbursement accordingly. He emphasized the beneficiary as well as the taxpayer: coinsurance calculated from reimbursement can sometimes exceed the institution's own drug acquisition cost. He described transparency and avoidance of duplicate discounts as important goals and credited CMS's implementation of the Medicare transaction facilitator.

He was equally direct about oncology compensation. Physician reimbursement should not depend on how expensive a prescribed drug is, but the agency's ability to repair that structure is constrained by statute. CMS cannot simply replace revenue reductions Congress has embedded in law. He hoped pressure from forthcoming changes would create legislative interest in a better arrangement. In audience discussion, he also acknowledged that reimbursement disparities and administrative burdens can encourage consolidation.

AI and outcomes-based reimbursement

Brooks's AI position combined optimism with skepticism. He said experience so far had largely been cost additive: improving the efficiency of individual tasks can amplify the existing system's spending incentives. Technology's potential to reduce costs and expand access would require an appropriate reimbursement framework.

He highlighted ACCESS as a promising example of paying for measurable clinical improvement, using reduced HbA1c as an illustration. CMS and CMMI could then work together on a payment structure tied to those outcomes. He also discussed reducing the gap between FDA approval and Medicare coverage, while stressing the need for evidence relevant to Medicare's coverage standard.

For pharmaceutical outcomes contracts, Brooks drew a useful distinction. Provider payment can change behavior; a drug contract principally allocates uncertainty about effectiveness between payer and manufacturer. He supported facilitating such arrangements while resisting designs that merely redistribute best-price discounts.

Supply chains and proposals for CMS

Asked about China, Brooks separated dependence on pharmaceutical starting materials from competition in innovation. He was concerned about supply-chain resilience and the US clinical-trial environment, while expecting the country to remain an attractive pharmaceutical market. His warning was that decades of optimizing procurement for cost had reduced resilience.

Brooks's wry warning was that consultants can sell companies months of expensive preparation for a bespoke CMMI model, only to arrive at CMS and be told the agency has no interest. He wanted structural approaches that could serve a broader market.

He urged stakeholders to bring data CMS can validate and proposals developed jointly by providers, payers, and manufacturers. Shared acknowledgment of a problem and evidence that a solution works are more persuasive than competing requests for money. For Friday, his challenge is to show which clinical outcomes AI improves, how those improvements can be measured reliably, and why the resulting payment arrangement would improve patient care rather than simply add spending.

Sources

Agenda: AVBCC 2026 Agenda 10.2.docx, Wednesday October 7 sessions 15, 17 and 21.

Provider panel: 20261007 Panel All About CMMI OCM EOM etc etc.docx.

Sutton interview: 20261007 Joe grogan interviews CMMI leader Abe Sutton.docx.

Brooks interview: 20261007 joe grogan intvu brooks Ctr Medicare CM.docx.