Rethinking the Payment Model in Oncology with Lalan Wilfong, MD: Value-Based Care, AI, and the Case for Caring Beyond the Clinic

Guest

Lalan Wilfong, MD

Senior Vice President of Value-Based Care, Thyme Care; Practicing Medical Oncologist and Hematologist, Texas Oncology; Co-Chair, Community Oncology Alliance Payer Reform Committee; Member, American Society of Clinical Oncology Drug Shortages Committee. 

Key Insights

• Fee-for-service payments don’t allow oncologists to practice medicine the way they really want. The current payment structure rewards volume and procedures, not the conversations, proactive outreach, and careful treatment decisions that define good cancer care.
• ‍Caring for patients between visits often moves the needle more than what happens in clinic. Keeping patients out of the ED and hospital requires a plan between visits: proactive nursing outreach, patient-reported outcomes, and attention to social needs.
• ‍The pace of change in cancer treatment makes value-based benchmarking uniquely hard. In oncology, drug costs and treatment paradigms shift so rapidly that there's often no stable baseline to measure total cost of care against.
• ‍Physician notes have gradually shifted from communication tools into billing tools. The information captured in clinical conversations should serve every stakeholder downstream, and not just the revenue cycle. This includes care teams, navigators, and patients.
• ‍Being a late adopter of ambient AI revealed what actually drives adoption. For clinicians who are already efficient, documentation improvement alone may not be enough to adopt ambient AI. Broader platform capabilities such as chart summarization and contextual nudges are often what close the gap.
• ‍Cancer care is a team sport. AI vendors who build only for the physician miss the full care team that makes physician excellence possible in the first place.

What Is the Status of Value-Based Care in Oncology? 

Dr. Wilfong's definition of value-based care gets right to the point: Value-based care aligns payment to patient outcomes rather than to the volume of services delivered.

In fee-for-service oncology, a physician gets paid for actions, whether that’s administering treatments, ordering tests, or seeing patients in clinic. The work that doesn't get reimbursed is what actually keeps patients healthiest, such as real conversations about treatment goals, proactive calls to catch a worsening side effect, or coordinating transportation so a patient shows up for their infusion.

As Dr. Wilfong describes it, the system simply wasn't built for what oncology has become. Over the past decade, cancer care has been transformed by targeted therapies, immunotherapy, and a proliferation of molecular subtypes that can each carry different cost structures and treatment demands. What was once a relatively stable clinical environment is now one where a single new drug can carry a monthly list price of $40,000, a number Dr. Wilfong cites from a recent conversation.

"In a fee-for-service world, payment is: you do something, you get paid for it. That's just not the best way to practice medicine, especially oncology."

The challenge for payers trying to build value-based programs is that measuring value requires benchmarking, and benchmarking requires stability. As an example of a more stable situation for value-based contracting, Dr. Wilfong notes dialysis because the treatment modality has largely stayed the same over time.

In oncology, however, trying to hold a practice accountable for lowering the total cost of care is genuinely difficult when the underlying treatment landscape is in constant motion. Add to that the macroeconomic problem of drug pricing—a problem primarily outside the control of any individual clinician or payer—and the picture gets more complex still.

How Does the Thyme Care Model Change What's Possible Between Visits?

Dr. Wilfong joined Thyme Care in 2024 because he believed the company’s model addresses a gap that traditional practice structures can't. By partnering with payers and taking on population-level risk, Thyme Care is financially incentivized to manage patients beyond the four walls of the clinic, an approach that fee-for-service ignores. 

The most common reasons that a cancer patient lands in the emergency room don’t happen in front of their oncologist. Nausea, neutropenia, infections, and the consequences of inadequate social support all unfold at home, often without a clinical touchpoint until things have gotten bad. The fee-for-service world doesn’t provide a mechanism to reach those patients earlier.

"I firmly believe that keeping patients out of the hospital when you can saves the system significant money. It's great for patients too.”

The value proposition for oncologists themselves is underappreciated, Dr. Wilfong argues. When a patient is admitted to the hospital, their physician walks across the building, logs into another system, spends time with the patient and family, writes a note, and returns to clinic. That time could have been spent seeing patients.

When patients arrive at clinic healthier, with their transportation arranged and symptom management handled, the oncologist can focus on making the right treatment decisions for that specific person sitting across from them.

The between-visit model also addresses a dimension of care that standard clinical documentation regularly misses. Dr. Wilfong describes noticing a family member's expression shift during a visit, a look of worry that led him to ask a follow-up question. He quickly learned the patient's son was an hourly employee who couldn't afford to take time off work. Without that observation and question, a caregiver’s financial crisis would have unfolded without anyone in the care system knowing to intervene.

Why Did a Tech-Curious Oncologist Wait Two Years to Try Ambient AI?

One of the more candid threads of the conversation involves Dr. Wilfong’s comparatively late adoption of ambient AI despite being a self-described technology enthusiast (and close colleague of DeepScribe CEO and show host, Matthew Ko.) In fact, Dr. Wilfong began adopting DeepScribe ambient AI two years after his colleagues at Texas Oncology. 

His explanation is simple and worth understanding. Dr. Wilfong had already developed a fast workflow, especially as someone who types over 120 words per minute. For him, the time cost of learning a new system didn’t feel worth the results when his baseline was already considered efficient.

“You came out with this new chart summary feature, which I'm super interested in. That is one of the biggest pains, having to click through thousands of documents trying to find the thing that you need."

What shifted for Dr. Wilfong wasn't a matter of better documentation. It was a chart summarization product feature. Trying to locate a relevant finding amid thousands of documents doesn’t get easier with faster typing.

When DeepScribe released the capability to surface a coherent patient snapshot from a complex record, Dr. Wilfong described it as potentially "getting me over the edge.”

It’s a familiar pattern that Matthew has heard from clinicians across the country. Ambient documentation is a starting point, but a system’s breadth is what gets clinicians interested and drives retention. Some physicians arrive through the documentation use case but those who are already efficient come in through features that address the problem of keeping up with an exploding volume of clinical information.

What Would Useful AI Nudges Look Like in Oncology?

The conversation moved toward the concept that information captured during a clinical encounter should do more than just feed the note. It should also flow to every stakeholder who needs it, in real time, without requiring the physician to document it differently.

For Dr. Wilfong, that vision maps directly to the between-visit work Thyme Care is taking on. If a clinician prescribes a medication to manage a side effect, Thyme Care's nursing team should know about it immediately so they can follow up with the patient the next day. If a patient says something during a visit that implies a social need, that signal should reach the care navigator.

"One type of nudge to start? It would be around toxicity management – many of these new drugs are coming out with unique toxicities that I may forget to ask about or dig into."

When it comes to surfacing information in his own clinic, Dr. Wilfong believes toxicity management nudges would be most useful. Many of the newest targeted therapies carry side effect profiles that are hard to retain in working memory, especially for an oncologist in the community setting who sees a wide range of cancer types.

A prompt in the right moment, asking whether a specific test has been ordered or whether a patient on a particular drug has been asked about a particular symptom, would catch exactly the kind of thing that falls through the gaps from a heavy cognitive load.

Dr. Wilfong also suggests that the oncologist is not always the right recipient of a nudge. Much of what surfaces as useful information during a clinical encounter might be better routed to a nurse, a care manager, or a medical assistant. This frees the physician to stay in the conversation with the patient rather than fielding signals.

What Do Technology Vendors Get Wrong About Oncology?

Dr. Wilfong closes with a direct message for anyone building tools for cancer care: The oncologist is not the whole care team. Treating them as if they are produces tools that underdeliver.

Cancer care, by nature, is a collaborative endeavor. It runs on a support infrastructure that includes nurses, MAs, navigators, social workers, and care coordinators. When any one of those functions breaks down, the oncologist's work gets harder. A patient who misses an appointment because transportation wasn't arranged, or who shows up sick because a side effect wasn't caught early, creates more work at the point of care.

"Cancer care is a team sport. Focus too much on the physician and you're missing all of the support team that is required to help that physician work to the top of their license."

The story Matthew Ko tells in response symbolically captures the risk of only focusing on doctors. Years ago, when DeepScribe first went to Texas Oncology to implement ambient AI, the team spent most of the first day trying to get through the front door. The gatekeepers were doing their job. But the experience was a signal that any technology built for physicians has to earn the trust of everyone else in the building.


You may also like:
Ambient AI and the Oncology Value Chain: Ways to Reduce the Cost of Complexity
Building Trust and the Future of Ambient Data in Oncology with Dr. Ravi Parikh

‍

Subscribe to “Beyond the Chart” to be notified of future episodes.

Learn more about DeepScribe's industry-leading ambient AI for oncology.

‍