Fall Forum Season Report Out: What We Heard Across Our Early Fall Forums
THMA's Fall Forum season is underway! In this first installment of our Fall Forum Insider series, we recap what we heard from health system leaders across oncology, information technology, and physician executive leadership. The specifics differed by forum, but some pressures consistently surfaced: patient demand outpacing physical and physician capacity, financial and policy pressures arriving at the same time, and AI value that depends more on how tools are introduced and owned than on what they do.
We've distilled key insights from early forums into Forum Insider pieces. Here's a quick look at what each covers:
Oncology Leaders: Capacity Constraints, Margin Pressure, and Shared Governance
Cancer programs running out of space, multiple revenue streams under pressure at the same time, and paired physician-administrator leadership that depends on visible results.
Read the Oncology Forum Insider →
CIOs: An Expanding Role, Vendor Consolidation, and the True Cost of AI
A role expanding well beyond technology, tighter margins shrinking the AI vendor stack, and hard ROI that misses much of what AI costs and delivers.
CPEs: Care Redesign, AI Adoption, and Value Measurement
Care models redesigned around workforce limits, AI adoption that hinges on clinician co-creation and trust, and value measurement shifting from activity to demonstrated outcomes.
Oncology Leaders Navigate Capacity Constraints, Margin Pressure, and Shared Governance
Oncology leaders convened at THMA's Fall 2026 Oncology Forum as demand for cancer care outgrows physical capacity and the revenue streams that historically funded cancer programs come under pressure at the same time. Many of the conversations focused on practical next steps: how to segment patients across care sites, build financial plans around cost data, and sustain physician engagement as labor pressure and change fatigue grow.
Key Takeaways:
1. Demand for cancer care is outgrowing physical capacity.
One system's cancer facility, opened only three years ago, is already full, and two programs under 10 years old said they have outgrown their space.
A group of leaders agreed that they are all facing a push toward non-hospital infusion centers and collectively argued for a single oversight structure for oncology and non-oncology infusion, so the two would not compete.
Several systems described keeping complex surgery central and moving infusion and follow-up closer to patients’ homes. One system's new referral model assumes 80% of cancers can be treated locally and connects the most complex 20% to disease experts.
2. Oncology's financial model faces sustained margin pressure.
340B is the most widely shared financial concern. A reimbursement presenter noted that for cancer service lines that rely on 340B, drug losses under the 2027 proposals will likely exceed any gains in other outpatient payments, though effects may vary.
Site-of-care rules are tightening. Off-campus hospital sites grandfathered before 2017 lost on-campus payment status in 2026, so their imaging is now paid at freestanding rates.
One presenter recommended modeling 2027 reimbursement as a portfolio and keeping hospital cost reporting accurate, since future increases will likely come through cost-based outpatient rates. Leaders are also closely evaluating alternative payment model participation.
3. Team-based care depends on shared governance and physician engagement.
One system places a physician-administrator dyad or triad at every level of its cancer service line. Another gives each solid tumor program three leaders with equal authority over funds.
Physicians seem to engage when their input leads to visible results. At one system, a council complaint about overnight calls led to a systemwide change within three months.
Leaders emphasized how workforce fatigue makes effective governance structures both more necessary and harder to sustain.
4. Patient needs extend well beyond active treatment.
With an estimated 18.6 million cancer survivors in the U.S., patient challenges are extending past treatment into cardiovascular risk, fertility concerns, financial toxicity, and fear of recurrence, which one leader called the most common concern.
In a cited study, 72.5% of patients who connected with a dedicated coordinator accessed support services, versus only 44.7% of those who did not leverage a coordinator.
One system aims to reach newly diagnosed patients within 24 hours, since patients who leave the system tend to face care delays of at least two to three weeks.
5. Operating as one system takes years of cultural work.
One system's leaders visited all six divisions in their first weeks and built a shared team brand. At another system, a leader said patients adopted the unified identity before faculty did and now expect faculty to follow.
One system gives every level, from individual to executive, a scorecard covering access, navigation, research, quality, and business growth.
One academic system spent two years merging legacy groups into single medicine and surgery departments, which a leader called the biggest cultural shift of their career.
CIOs Weigh an Expanding Role, Vendor Consolidation, and the True Cost of AI
Chief Information Officers convened at THMA's Fall 2026 CIO Forum as technology now touches every part of healthcare operations and financial pressure pushes CIOs to consolidate vendors, shorten contracts, and reconsider building in-house. The conversations focused on AI: what it really costs, where it pays off, and how its governance and liability are being settled system by system.
Key Takeaways:
1. The CIO role is expanding well beyond technology.
One system noted that a quarter of its jobs were technology-focused a decade ago and essentially all are now.
Peers noted that CIOs who are perceived as technology leaders, rather than service providers, tend to get pulled into acquisition and integration conversations sooner.
To help other leaders see their own work in technology initiatives, one system organizes its technology roadmap by audience — nursing, physicians, revenue cycle.
2. Tighter margins are shrinking the AI vendor stack.
Some leaders described no longer signing multi-year or initial-year vendor contracts, writing six- to twelve-month ROI exits into most AI agreements, or requiring payback within two years. About 70% of CIOs in one session expect to end 2027 with fewer vendors as core platforms absorb more capabilities.
One system now builds roughly a third of its AI in-house, delivering solutions in weeks for a fraction of the roughly half-million dollars most vendors charge.
3. Hard ROI misses much of what AI costs and what it delivers.
A model developed by one system puts the commercial price of a technology at roughly 35 cents of each dollar, with enablement, governance, and post-go-live operation making up the rest.
At one system, ambient documentation was treated as a perk until leaders saw that low adoption drove the cost per encounter to nearly three times the target. Once visible, the business chose to own that risk, and every provider now carries an AI adoption metric.
One leader explained their strategy is to evaluate AI on hard metrics where possible, but not to discount how AI can provide a return on experience when a technology saves time, reduces burnout, and maximizes efficiency.
4. AI pays off when narrowly scoped and operationally owned.
Limited to routing, cancellations, and reschedules, one call center AI tool returned 8.8x ROI across 300,000 calls a year at one system. However, stretched to end-to-end scheduling at a second system, the same vendor’s product booked only 9% of appointments without a human.
One system now names an operational sponsor before any pilot begins. A peer system goes a step further, requiring an accountable owner with the authority to end a pilot that isn’t meeting its goals.
Several systems screen AI requests by first asking whether the underlying process is standardized, documented, and repeatable. One noted that many workflows trace back to a decade-old memo no one has revisited.
5. Governance and liability are being settled system by system.
A twice-monthly review committee vetted every AI tool as one system’s portfolio grew from 300 to more than 450 deployments this year. No tool cleared review without some form of mitigation, yet leaders credit governance with enabling that pace.
Proposed state legislation has slowed AI and data collection efforts at one system. Facing the same regulatory uncertainty, a peer system said its governance structure has helped implement deployments faster.
The “human in the loop” concept raises more questions than it answers. One CIO questioned whether clinicians making thousands of judgment calls a day are genuinely reviewing AI output or absorbing liability.
CPEs Tackle Care Redesign, AI Adoption, and Value Measurement
Chief Physician Executives convened at THMA's Fall 2026 CPE Forum as patient demand continues to outpace physician supply and federal policy changes draw clinical executives into planning discussions alongside finance and government relations. A shared tension ran through the conversations: faster tasks and more appointments do not, by themselves, establish that care improved.
Key Takeaways:
1. Care models are being redesigned around workforce limits.
Physician replacement math reinforces the case for redesign. By one leader’s estimate, backfilling a single physician departure can require two-plus FTEs once training time and reduced initial productivity are factored in.
One system embeds virtual clinicians within its medical group to manage chronic disease panels and leverages remote patient monitoring to offload routine patient visits. While these efforts allow in-person physicians to increase productivity, adoption across hundreds of sites required compensation redesign to align incentives.
Advanced Practice Providers (APPs) are becoming a larger component of care delivery. One organization employs almost 2,000 APPs alongside 4,800 physicians and expects to reach parity within a couple years.
2. AI adoption hinges on clinician co-creation and trust.
Systems report stronger uptake when physicians help define the problem, design workflows, and see measurable results, such as reduced after-hours EHR time, rather than simply receive finished tools.
Several leaders deliberately recruit critical but constructive physicians as champions. Peers trust these clinicians' judgment because they identify weaknesses rather than evangelize, though this profile can be harder to find.
Multiple systems have AI governance committees that vet products for safety but lack authority over strategy, prioritization, or cross-service-line coordination.
3. Value measurement is shifting from activity to demonstrated outcomes in AI, access, and pharmacy.
AI-supported appeal letters reduced drafting time at one system, but leaders had not decided how to deploy the recovered staff capacity, making return on investment difficult to quantify.
Several systems described shifting measurement from slots created to how quickly patients with specific conditions reach care, though standardizing these metrics across service lines remains a work in progress.
One system evaluates its pharmacy program across clinical outcomes, medication access, health plan savings, and margin — rather than isolating a single department’s financials — though building the model required significant upfront capital.
4. Health equity work is moving into clinical quality structures.
A “stealth equity” posture is common. Leaders reported sustaining equity outcomes while adjusting communications in response to scrutiny, characterizing their systems’ stances as “guarded,” “committed,” and “constrained.”
By applying an equity lens to quality measures from cancer screening, hypertension, and diabetes management programs, one system finds that strong aggregate performance can mask variation across populations.
Folding equity leadership into the clinical quality structure, rather than running it as a separate function, could make the work harder to dismantle. Early adopters of this model are still in the first stages of the reorganization.
5. Policy and financial pressures are pulling clinical leaders into enterprise strategy.
Medicaid eligibility changes risk interrupting treatment for patients who lose coverage, with downstream effects on chronic disease management that clinical leaders are being asked to model and communicate.
340B program changes put significant revenue at risk, with leaders describing the need to model multiple policy scenarios simultaneously rather than wait for regulatory clarity.
Policy navigation, scenario planning, and advocacy increasingly consume leadership bandwidth, with clinical leaders expected to bring patient consequences into enterprise decisions earlier than they have historically, though the scope varies by system.