VBCC 20260424 Strategies from OCM to V2 to V3...

 

First-pass strategic take

Yes, EOM is meaningfully different from OCM, but not enough to be reassuring. It is better designed, narrower, more patient-centered, more equity-aware, and more digitally ambitious. But the early evidence still looks like OCM 2.0 rather than a true new operating system for value-based cancer care: modest payment reductions, no visible quality/utilization gain yet, and net losses once care-management and incentive payments are counted.

That is actually a strong panel premise: VBCC has not failed as an aspiration; it has failed because the measurement layer is still too weak, too claims-centric, and too poorly connected to what oncology patients and clinicians actually experience.

---

1. OCM vs EOM: what changed, and should we be reassured?

OCM: the disappointment

The final Brooks/JAMA paper is devastating in a quiet, technocratic way. OCM ran from 2016–2022, included more than 200 practices, covered about one-fourth of systemic cancer treatment in FFS Medicare, and produced a statistically significant but modest $616 reduction per episode, increasing to $1,282 in the final performance period. But there were no statistically significant differences in hospitalizations, ED visits, or quality, and after MEOS and performance payments, Medicare had an estimated $639 million net loss.

That says: OCM learned how to bend some spending, but not enough to pay for itself, and not with a measurable patient-quality signal. The Thomas/Ward critique anticipated part of this: OCM was not really a true bundle; it preserved FFS and layered on a complicated shared-savings/payment overlay, leaving practices partly paid to do more and partly rewarded for doing less.

EOM: the real improvements

EOM is not just a rename. It makes several real changes:

Design feature	OCM	EOM
Cancer scope	Broad systemic therapy episodes	Narrowed to seven cancer types
Hormonal therapy-only episodes	Included in OCM	Generally excluded; focuses on systemic chemotherapy, not hormonal-only
Risk	Shared savings; downside risk evolved	Downside risk from the start
MEOS	$160 PBPM	$110 PBPM, or $140 PBPM for dual eligibles
Equity	Less central	HRSN screening, equity plans, dual-eligible enhanced payment
Patient-reported outcomes	Not central	ePRO collection and monitoring required
Digital/data strategy	Less mature	Requires quality, clinical, and sociodemographic data; CEHRT and CQI use
Participation	>200 practices in OCM	As of March 2026, 28 practices and 1 commercial payer

CMS describes EOM as focused on seven cancers, six-month episodes, total-cost accountability, MEOS payments, required 24/7 access, navigation, evidence-based guidelines, comprehensive care plans, HRSN screening, ePROs, CQI data use, and certified EHR technology. (Centers for Medicare & Medicaid Services)

The non-reassuring part

The first EOM evaluation already has a familiar smell. CMS’s first at-a-glance report for Performance Period 1, July–December 2023, says EOM likely reduced payments but produced a net loss to Medicare after MEOS and incentive payments, and did not affect quality or utilization measures, including hospice use before death or acute care utilization. (Centers for Medicare & Medicaid Services)

That is the core answer: EOM is more sophisticated, but the early signal is still not a proof of VBCC. It is a better-engineered experiment, not yet a successful value model.

---

2. Are there new and better metrics in EOM?

Yes, but with a big caveat. EOM adds the right categories of measurement, but many are still not mature, outcome-forward, or digitally reliable enough to bear payment accountability.

The good news is that EOM formally moves beyond pure claims logic. It requires ePROs, HRSN screening, clinical data elements, sociodemographic data, and quality reporting. CMS’s participant resources now include an EOM Clinical Data Elements Guide, Sociodemographic Data Elements Guide, Quality Measures Guide, and cost/quality performance data files for Performance Periods 1–2. (Centers for Medicare & Medicaid Services)

Your prior measurement report frames this well: the plausible springboard is the convergence of EOM redesign requirements, digital quality measures, and oncology interoperability infrastructure such as USCDI+ Cancer and mCODE.

But the caveat is decisive: EOM has better measurement ingredients, not yet better measured outcomes. The model asks for ePROs and HRSN data, but the early evaluation did not show measurable improvement in acute care utilization or quality. (Centers for Medicare & Medicaid Services)

The most promising domains for real VBCC metrics are the ones you already identified:

ePRO symptom/toxicity control. This is probably the single strongest new metric domain because it connects directly to patient experience, acute care avoidance, and clinical response workflows. Your report proposes measuring both completion of standardized symptom assessments and timely response to severe symptom alerts.

Physical function preservation. More meaningful than “patient satisfaction,” and closer to what cancer patients care about: can I function, work, walk, sleep, eat, and live my life?

Avoidable acute care utilization. Still useful, but claims alone are crude. ED visits and admissions become more meaningful if paired with symptom-triggered preventability review or ePRO context.

Evidence-based regimen/pathway concordance. This could matter greatly, but it requires structured stage, biomarkers, treatment intent, and exception logic. Otherwise, it becomes a documentation game.

Goal-concordant end-of-life care. Claims can measure late chemotherapy and hospice timing, but the real outcome is whether care matched patient goals. Your report rightly pairs claims-based EOL metrics with structured goals-of-care documentation.

Financial toxicity. This is a missing VBCC domain. If cancer care bankrupts or destabilizes a patient, “value” has not been achieved. Your report proposes validated financial toxicity screening plus navigation response.

So the answer is: EOM points toward better metrics, but it has not yet demonstrated better outcomes. It is more promising as a measurement platform than as a proven payment model.

---

3. What should “V3” of CMS value-based oncology look like?

Your strongest forward-looking thesis is:

OCM tested whether care-management payments plus shared savings could make oncology cheaper. EOM tests whether a narrower, risk-bearing, equity-aware model with ePROs can do better. V3 should test whether digitally computable, patient-centered oncology outcomes can finally become the basis of value-based cancer care.

In other words, V3 should not be merely EOM with different benchmarks. It should be a model where measurement is the product.

V3 should have five design principles

First, V3 should make oncology clinically legible. It cannot rely mainly on claims. It needs structured diagnosis, stage, biomarkers, line of therapy, treatment intent, progression/recurrence, performance status, and death date. EOM’s clinical data element work and mCODE/FHIR alignment are early scaffolding, but V3 should make this the required substrate. Your report puts it neatly: oncology’s real clinical context is often buried in notes and cannot be reliably inferred from claims.

Second, V3 should use AI/NLP as measurement infrastructure, not as a magic wand. The panel topic is strongest if you say: AI review of EHRs may finally allow VBCC to measure what claims cannot—stage, progression, toxicity, treatment intent, ECOG-like function, adverse events, biomarker appropriateness, and goals-of-care discussions. But AI-derived metrics must be validated, audited, version-controlled, and bias-tested. Otherwise V3 becomes a digital façade: computable, impressive, and wrong.

Third, V3 should use a small core measure set. Something like 8–12 measures, not 50. Candidate domains: symptom control, functional preservation, avoidable acute care, evidence-based regimen appropriateness, time to treatment, EOL goal-concordance, financial toxicity, and equity/whole-person supports. That aligns with your prior report’s proposed VBCC portfolio.

Fourth, V3 should separate “drug price exposure” from “care delivery performance.” OCM and EOM struggle because oncology spending is dominated by therapies whose prices and clinical indications are not fully controlled by the practice. EOM’s early report says systemic cancer treatment drug spending accounts for about 58% of EOM episode costs, and participants reported focusing on drug spending interventions. (Centers for Medicare & Medicaid Services) V3 should distinguish: Did the practice choose appropriate therapy? Did it manage toxicity? Did it avoid preventable acute care? Did it align care with patient goals? It should not simply punish a practice because the correct therapy is expensive.

Fifth, V3 should make equity measurable without making safety-net care financially toxic to providers. EOM adds HRSN screening and dual-eligible enhanced MEOS payments, which is directionally right. But V3 should require equity stratification and closed-loop resource referral metrics while protecting practices that care for medically and socially complex populations. CMS describes EOM as requiring HRSN screening, equity plans, expenditure/utilization reports to identify disparities, and higher MEOS payments for dual eligibles.

---

A sharper conference thesis

Here is the central framing I would use:

“Value-based cancer care has been stuck in an awkward middle stage: payers can measure cost, but not value; clinicians can describe value, but not compute it; patients can feel value, but it rarely appears in payment models. OCM showed that care redesign can produce modest savings without measurable quality improvement. EOM improves the model by adding downside risk, ePROs, health-related social-needs screening, and richer data requirements, but early results still look financially and clinically inconclusive. The next version of VBCC will depend less on another tweak to shared savings and more on whether AI-enabled EHR review, ePROs, mCODE/USCDI+ Cancer, and digital quality measures can produce auditable, patient-centered, oncology-specific metrics at scale.”

That is a strong panel because it is neither naïvely optimistic nor drearily cynical.

---

Draft paragraph-long panel proposal

Are We Finally Ready for Real Value-Based Cancer Care? From OCM and EOM to AI-Enabled Measurement.
For more than a decade, value-based cancer care has promised to reward better outcomes rather than higher volume, yet progress on the ground has remained limited. The Oncology Care Model produced modest reductions in Medicare episode payments but no significant improvements in utilization or quality, and net losses to Medicare after model payments. Its successor, the Enhancing Oncology Model, adds important improvements—downside risk, narrower cancer scope, electronic patient-reported outcomes, health-related social-needs screening, enhanced services, and richer clinical data requirements—but early results remain inconclusive. This panel will ask whether the missing ingredient has been measurement itself: the ability to capture oncology stage, biomarkers, treatment intent, toxicity, function, goals of care, financial toxicity, and equity outcomes in computable, auditable form. We will explore whether AI-enabled EHR review, ePROs, mCODE/USCDI+ Cancer, and digital quality measures can support a true next-generation “V3” oncology value model—one that moves beyond claims-based cost control toward patient-centered, clinically meaningful cancer care performance.

---

Possible panel title options

Best straightforward title:
From OCM to EOM to V3: Can Better Measurement Finally Make Cancer Care Value-Based?

More provocative:
Value-Based Cancer Care’s Missing Operating System: Metrics, AI, and the Road Beyond EOM

Most “conference program” friendly:
The Next Generation of Value-Based Cancer Care: Lessons from OCM, Early EOM, and AI-Enabled Outcome Measurement

Most Bruce-style:
After OCM and EOM: Is Value-Based Cancer Care Still Waiting for Its Measurement System?

---

Suggested panel architecture

You as moderator/host should set up the tension: “We have had a decade of VBCC conferences, but the field still often means payer coverage plus ASP drug pricing. What would make it real?”

Ideal panelists:

1. OCM evaluation author
   Gabriel Brooks would be excellent, because the JAMA paper is now the cleanest empirical anchor. Nancy Keating would also be very strong, especially for evaluation design and interpretation. The Brooks paper lists Brooks, Trombley, Landrum, Liu, Simon, and Keating among key authors, with Brooks, Trombley, and Keating drafting the manuscript.
2. CMMI/EOM representative
   Someone who can speak to EOM’s design choices, first evaluation, clinical data elements, ePROs, HRSN screening, and what CMS hopes to learn by 2030. Even if they cannot speculate about “V3,” they can discuss what EOM is designed to test.
3. Community oncology practice leader
   Preferably someone actually implementing EOM, ePROs, navigation, urgent care access, and “call us first” workflows. This prevents the panel from becoming a policy seminar in the clouds.
4. Measurement/interoperability person
   Someone from ONC/ASTP, NCI’s USCDI+ Cancer effort, HL7/mCODE, NCQA, or a serious EHR/FHIR implementer. Their role: what needs to be standardized before oncology outcomes can be measured without heroic chart abstraction?
5. Patient-centered outcomes / financial toxicity voice
   Ideally someone working on ePROs, symptom monitoring, financial toxicity, or patient advocacy. This keeps the panel honest: VBCC should not become “cost-based cancer care with nicer adjectives.”

---

My bottom-line answer to your three questions

OCM vs EOM: EOM is a real improvement in design, but not yet reassuring in results. It narrows the model, adds downside risk, requires ePROs and HRSN screening, and builds a richer data strategy. But early EOM still shows payment reductions offset by participant payments and no clear quality/utilization improvement.

Metrics in EOM: EOM has better metric ingredients—especially ePROs, HRSN screening, clinical data elements, sociodemographic data, and CQI reporting—but the decisive step from “data collection” to “valid outcome measurement” is still ahead.

V3 forecast: V3 should be an AI/digital-measurement-enabled oncology value model: clinically rich, patient-centered, equity-stratified, auditable, and built on structured oncology data plus validated AI extraction from EHR text. The goal is not simply to save money on chemotherapy episodes, but to measure whether cancer care actually improves symptoms, function, appropriateness, goal-concordance, financial well-being, and avoidable acute care—without drowning practices in manual reporting.