Advancing Value-Based Oncology With the Predictable Cost of Care Working Group: Part 2

In this episode of Oncology Innovations, experts discuss the Predictable Cost of Care Working Group’s efforts to develop a standardized model for assessing oncology treatment costs, highlighting insights from phase I, the goals of phase II, and the importance of incorporating broader industry and patient perspectives.

Read the full transcript:

Gordon Kuntz: Welcome to Oncology Innovations, a Journal of Clinical Pathways podcast, focusing on candid discussions with innovators dedicated to enhancing quality, value, and the role of clinical pathways in the evolving cancer care ecosystem. I'm your host, Gordon Kuntz. I'm a consultant with 20 years of experience in oncology clinical pathways and the business of oncology.

I've worked with oncology practices, pharma, payers, group purchasing organizations (GPOs), and pathway developers—basically, every aspect of the oncology ecosystem. I'm really looking forward to today's conversation.

We're joined today by Carole Tremonti, senior director of customer success and provider partnerships at ConcertAI. While this is a new role for Carole, it is not a surprising move by any means. Carole is a pathways expert and longtime innovator in oncology care. I've had the pleasure of working with Carole for several years, and over the past couple of years, collaborated with her on several projects. Today, we're going to be talking about the Predictable Cost of Care Working Group that she and I comoderated in 2024 and focusing on the work we're starting in phase II in 2025.

Carole, the results of phase I are just being published in the Journal of Clinical Pathways. While we didn't accomplish everything we might have hoped, what we did was actually quite significant for the world of oncology pathways. Can you describe how the work from phase I can be used?

Carole Tremonti, RN, MBA: I completely agree with you, and I am so proud of what we were able to accomplish for the broader oncology pathway community.

We tackled a problem that people have been talking about for years but that everyone thought was too complicated to even approach. Given that there were so many variables that could or should be included in the model—which was the difficult part—one of the things we decided to use to make this process more objective was a Delphi round-robin method to define those variables in the model in a way that was unbiased. In doing so, it wasn't just the representatives in the working group making the decisions; rather, it was the voices from the broader oncology community—providers from different domains—that shaped and ultimately defined the variables we chose to use in this model.

From what was created in our phase I work, we can now begin the work of constructing the models, beginning with metastatic non-small cell lung cancer, and then test them in real-world settings to see whether they work and whether they are applicable. We do believe we've created an approach that is flexible enough to be applicable to other solid tumor disease states, but we need to test that theory.

Kuntz: Absolutely. Again, the goal of the work here in phase I and again in phase II—where, as Carole described, phase I was to find the model, and phase II is actually building the prototype and what will eventually be the model itself for use in the future—is to have an industry-wide solution.

The phase of definition was really critical, and I feel like we did a very good job with that. As you mentioned, the Delphi method broadened that set of perspectives to, I believe, 25 people voting on various aspects of the model in addition to the number of people we had in regular discussions. As we broaden the number of pharma participants that are involved, we are able to get some new voices in there as well. I think we really have something that we can absolutely say is built for the industry, not just for a particular specific use case.

Tremonti: I agree. As you recall, we presented the outcome of this work—where we were going with it—last year at the Clinical Pathways Congress and Cancer Care Business Exchange, and the response from the oncology community was wonderful, overwhelming, and a little unexpected. Everyone really embraced it. Additionally, during our question-and-answer session, so many unique perspectives and thought processes came to bear, and the opportunity to invite those different perspectives in to solve the problem is the beauty of the Delphi method. We knew we needed a broader perspective, and people really stepped up to the plate for that, which we were thrilled about.

Kuntz: I had at least half a dozen people find me after the conversation on stage and ask how they could help and get involved. It was truly wonderful. To that end, we're actually starting phase II very soon. We're recording this in early February, and I think we're going to get started in March sometime. Things are just being scheduled right now. What kind of things do we want to tackle in 2025, and how will this be different than our first time around?

Tremonti: The number 1 lesson learned, as I mentioned, is to be sure we have a broader and more comprehensive perspective in the group itself. Number 2 was to test the model. We had hoped we could get that far in phase I, but it wasn't realistic. We bit off quite a bit. In phase II, we want to test the model and do a complete analysis to see whether it is useful in the decision-making setting—because that's the intention—and specifically what decisions this information was useful for.

Again, we've got different perspectives, use cases, and needs, and we're trying to meet all of them—or at least get to 80% of their needs, if possible. The third thing we're going to accomplish is applying the model to 1 or 2 different disease states to really understand the applicability and scalability of this. We have the variables for one, but if it doesn't work when you pick it up and apply it to breast cancer or prostate cancer, then it's not very helpful in the long run.

Kuntz: We may have to adjust something along the way. The other thing I know we're going to include—we talked about this a fair amount in phase I, but we didn't have the representation to adequately land on the answers—is that we've invited an additional participant to represent the patient perspective.

We had pharmaceutical representatives, we had pathway developers, all of which really focused on the decisions that the oncologist is making. One hopes that they're made in a shared decision-making environment, but ultimately, it's focused on the professional side, the clinical side and we want to incorporate that aspect of the patient impact. It's not necessarily about asking the patient what drug they should be receiving, of course, but the goal is [to find out] what are the costs? Maybe not even in dollar terms, but what are the costs to the patient?

Can you describe a little bit about where we think that might go? We'll have to wait and see a little bit, I understand, but can you describe a little bit where we think that might go?

Tremonti: This was, as you recall, a highly discussed topic. Everyone agreed they wanted the perspective of the patient brought into things. The pharmaceutical companies were lobbying the strongest for it because that is important.

To your point, with scope keep [in mind], we are upstream from the treatment decision even being made or the discussion with the patient even being made. We are far upstream—essentially fitting it into guidelines that physicians follow when they're talking to the patient.

That being said, how can we represent the patient? We are thinking about that. One of the ways [we approached this] for phase II was ensuring we were including people who would represent the patient perspective and give us ideas on what metrics we could include with a patient perspective. But we did discuss it nonetheless in phase I, and a lot of the things we were thinking about—again, in terms of what's predictable—were very difficult for us to get. For example, the out of pocket spend of patients because everyone has a different payer, there are a lot of complicated variables to that.

However, if you go back to the clinical trial, we looked at what could you consider to be the cost? Not just the drug cost, but time toxicity, an incredibly relevant topic for patients. How much time do they have to spend engaging with the system? How does it relate to work loss? Components like that.

Kuntz: It's as disruptive as the as the care itself.

Tremonti: Exactly—focusing on the care itself. That's the approach we're going to try to take in model number 2: what are the predictable patient costs that we can include so that in that space, we are really being representative?

Kuntz: I want to loop back—and maybe we should have talked about this in the prior conversation as well—but when we talk about predictable cost of care, as opposed to total cost of care, I'm going to just talk for a minute because one of the premises that I came into this with when we first were designing this and talking about it was there are elements of cost associated with a particular regimen that we can know and predict. There are others that, honestly, stuff happens, and when you look at, as you mentioned last time, the retrospective total cost of care and tally up all the costs that were associated—typically with a given time span in which a patient is receiving care for their cancer—there are all sorts of things that might have come into play. Maybe they had a broken arm. Maybe they had a heart attack. Maybe they had something completely unrelated to the cancer, but that gets tallied up in many cases.

When we talk about predictable cost of care, we're really talking about those things that we can know and predict. You and I have had a number of conversations about that over the last year. Talk about that a little bit. Let's relate that again to the patient piece because, as you said, when it comes to the financial component, maybe they met their deductible, maybe they haven't. Maybe they have a great plan, maybe they don't. Maybe they have insurance, maybe they don't. There are all sorts of out-of-pocket costs that we can't know about; they're not knowable or predictable. We have to set those aside and focus on what we can identify and measure in a trustable way. I just want to explore that a little bit further, if you don't mind.

Tremonti: It's really important that we put that in the forefront. Again, going back to the clinical trial, we had to say, "Okay, what could be done in a clinical trial?" It was the distinction between the total cost of care, because the total cost of care is a retrospective view, and you can include everything. A clinical trial doesn't represent patients with comorbidities that often result in those one-off variables that are not consistent for everyone. What the group decided was to come back to our core of predictability because, again, you have some agents where you can't look retrospectively. They haven't been used yet. We don't know what all the nuanced ancillary costs may be in the real world because we have to focus on the clinical trial to make it apples to apples.

In considering that, when we thought about the cost we wanted to use, we really focused on what adverse events we expect someone to experience. If even 60% of patients experienced it, we're going to include it. If we know an agent is going to cause severe nausea and vomiting, then we need to include the support for that. How many patients on the clinical trial had to come in for fluid replacement? Components like that—expensive antiemetics versus low-cost antiemetics. What do I need to manage the patient? That stuff is in the clinical trial, to your point.

This gets, again, to, while we are bringing in the patient perspective, how we create boundaries around it that will then translate to multiple different situations. That really was the hardest thing for us to stay true to along the way.

Kuntz: When you talk about scope creep, one of the elements of scope creep that we fought against, frankly, was that there are a lot of exceptions in cancer care. Not every patient—maybe 50%, maybe 5%—is going to have a severe reaction or adverse event related to the drug.

I don't know which 5% of patients it's going to be. Any given patient is not going to be represented perfectly. But if you take the population and apply 5% of the cost of treating or managing that side effect—or [consider that] some patients live across the street from a cancer center while others live 100 miles away—there are a lot of challenges with creating an average patient notion. However, I found that the group finally broke through the normal way of looking at things on a very individualized basis, which is advantageous if you're a cancer patient, I understand. But when we're thinking about how to compose this treatment selection, how to think about drugs on a little bit of a broader scale, being able to break out of that individualized approach is sometimes a little bit necessary.

Tremonti: It is, because I'll channel my favorite phrase by Voltaire, "Do not let perfect be the enemy of good." When you think about the way providers make decisions, 50% is based on evidence, 25% is based on what they are trained on, and then the other 25% is based on their nuanced experience that they've had throughout their career, no matter how short or long it is. While we like to be precise, even individual decision-making isn't precise. What elements can we include that we can fundamentally agree are good enough?

It isn't going to be perfect. There is no perfect, especially in clinical care, because we're talking about humans, and humans are inevitably different from each other. We have different scenarios. You can't apply the same scenario to every individual because then you're not providing individualized care. Pathways isn't necessarily about providing individual care; it's about what would work for 80% of the population. What's good enough? We have to think about patients in the same context.

Kuntz: Absolutely. By the way, I'm very proud of you for having a favorite quote from Voltaire.

Tremonti: Thank you.

Kuntz: What kind of additional participants do you think we're going to be looking for in phase II? I know we're starting to issue invitations, but who are we looking for in terms of participation?

Tremonti: Interestingly, a broader range of pharmaceutical companies and the large players. We had 2 of the large players in the first round, and we want to include more specifically because they are going to use this model. We want to make sure that, as much as we can, everyone understands the work we're doing, how it relates to them, and how it will be used. We want to be sure to have those ancillary or orthogonal perspectives not only listened to but also responded to because every single thing that a pharmaceutical company wants—or even an academic medical center wants—doesn't work for everyone.

The second thing, and we mentioned this a little when we talked in our first podcast about this work, is bringing in members of groups that advise on the finances of oncology. An example could be Milliman.

Thirdly, what we wanted to do in phase I—and as we were just talking about, that was out of scope—was really to bring in that patient perspective. That's a tough variable to account for because who is the right perspective?

We're getting it. We're working on it. When we presented this work initially at the Clinical Pathways Congress Cancer Business Exchange, there were a lot of patient advocates there. In the comment section, during the question-and-answers, they really stood up and said, "You need to include the patient perspective." Even providers said that. That's the priority this round.

Kuntz: Absolutely. Getting a broader range of pharma companies involved—I’ve had conversations with my clients and others about this—creates buy-in on their part. It also ensures that it's not the perspective of just a couple of pharma companies, but that we have this broader industry perspective so that those that didn't participate at all, when we get to the end of this and we publish the model and we say, "Here's the formula. Here's the recipe for how you do all of this," they can look at it and go, "Yeah, that makes total sense."

You don't want to get to the end of it and have them go, "That doesn't make sense for us at all. We're not going to use that. We're going to do our own thing again." That just makes life more difficult for everybody involved.

We did phase I. We're into phase II. Do we see a phase III and IV?

Tremonti: I would certainly hope so. This is hard. This is really hard work and it's complicated. It might take 3, 4, or 5 iterations. We're not going to know until we do this work along the way. Truthfully, the interest this work has received has been beyond what I expected and, I believe we've talked about this, beyond what you've expected. We were all pleasantly surprised at how we were embraced by the community. Given the critical need to understand and help people to understand and curb the costs in oncology, it makes this work even more important and more relevant.

I think something that we understood and gathered from phase I is that—and this has happened in oncology in the last 5 years—think about big data: we all have to come together. It's a common problem, and it's going to take all of us to solve it. We know it's real. So why not work together more collaboratively?

Pending model performance from phase II, I think this group will expand their work. Importantly, we have solid tumors now, but really, what does a hematologic malignancy model look like? Because there are distinct nuances in how solid tumors vs the hematologic malignancies are treated. We want to be sure to be cognizant of that. When you're thinking about multiple phases, yes, we might get it to a point where it works for solid tumors. Then what about everybody else? You can't leave out the rest of oncology care.

Kuntz: Of course. Excellent. That wraps up this episode of Oncology Innovations. Carole, thank you for a great conversation, and thank you for joining us today. Thank you to the Journal of Clinical Pathways for producing this episode. As always, our goal is to bring you candid conversations with industry leaders who are driving innovation and shaping the future of oncology care.

If you enjoyed this discussion, be sure to subscribe, leave a review, and share this episode with your colleagues. For more insights on oncology clinical pathways, value-based care, and payer dynamics, visit the Journal of Clinical Pathways and follow Oncology Innovations for future episodes.

I'm Gordon Kuntz, and thank you for listening. Until next time, stay informed, stay engaged, and keep driving innovation in oncology care. Please download, rate, review, and subscribe to the podcast. For more episodes, you can visit www.journalofclinicalpathways.com. Also, be sure to share Oncology Innovations with a friend or colleague. See you next time.