Best of the Best Gen Surg - How Far is Too Far Cost-effectiveness Analysis of Regionalized Rectal Cancer Surgery - Dr. Leeds

I'm going to announce now the winner of the SSAT poll. It was close. It went back and forth, but, uh, Doctor Ryan Morgan's paper on local and regional tumors, uh, won by 55.6%. That was the total, very close. And so we have the first two winners, and we're going on now with the second and last paper from the ASCRS. And uh then we'll have the poll of the best of the best and uh the last paper is presented by Doctor Ira Leeds and it's how far is too far cost effective analysis. Let me see if the total title. I can't read it here. Cost effective analysis of what that, uh, hold on here. It's blacked out on my. It is. Um, how far is too far? Cost effective analysis of regionalized rectal cancer surgery. Yes, thank you. I'm sorry for that. Let's go ahead, Doctor Leeds. Thank you to the organizers and the co-hosts for the opportunity to present this work that was previously presented at the American Society of Colon and Rectal Surgeons' annual scientific meeting. I have no disclosures related to this presentation. Regionalization of rectal cancer surgery has well-recognized associated advantages. It increases lymph node yield, it increases the use of neoadjuvant therapy, it decreases 30 and 90-day mortality, and then presumably through these benefits, it is also associated with a 34% decrease in 5-year mortality. In contrast, there are costs associated with regionalization. Specifically, more than 3 times the number of patients would have to travel more than 50 miles to achieve the same amount of care that they could achieve locally. Also importantly, these patients that do travel tend to be more white and more privately insured, raising issues of healthcare equity. The purpose of this study then was to assess the cost-effectiveness of regionalized rectal cancer surgery, and we hypothesized that regionalized rectal cancer surgery would not be a preferred strategy when factoring in patient costs. We performed a decision tree model. We did this by looking at a strategic decision and looking at the two options that come from that. Each option, in this case, regionalized versus local surgery, has costs and benefits associated with that. And the goal is to create an ICE or an incremental cost-effectiveness ratio, which compares the change in cost versus the change in benefits with each arm of this strategy. This conceptualization of regionalized rectal cancer surgery shown previously meets the conditions where you have costs and benefits that can be compared in a methodologically rigorous way. To do this kind of methodology, you take parameter estimates from across the literature, some examples of those are shown here. Importantly, many times this methodology does not do primary data collection, in this case, we did not as well. Here are the results from a decision tree analysis once calculated back to their starting point of whether or not, whether or not to regionalize care or not. I'll bring you to the top-level findings here, where the marginal cost or the average cost per patient as expected in the model for the regionalized care versus the local care arm were cheaper with regionalized care. Importantly, looking at the clinical outcomes also showed a mortality benefit in the regionalized arm versus the local arm. This is a cost effectiveness condition called economic dominance, or where the costs of care are cheaper in the same arm that also has better clinical benefit. Now we mentioned this data comes from pre-existing published literature. To perform then some measure of how robust your findings are, we performed a sensitivity analysis of over 10,000 different scenarios using combinations and distributions of different values from the literature. When you do this, you notice that almost all of the scenarios cluster in one quadrant of an er plot shown here with the Y axis demonstrating cost and the X axis demonstrating effectiveness measured in qualities. Because all of the, nearly all of the scenarios occur in quadrant 4, that has a particular meaning with this methodology. The quadrant 4 is where regionalization is the dominant strategy, having both lower costs, but also better effectiveness than local care. This is important because it highlights the robustness of our model to changes in the parameter estimates that we collected from the literature. There are always limitations with any study. In this case, the cost-effective methodologies are particularly helpful for population level decision making. This averages expectations and outcomes across all patients, not one particular patient. So when thinking about strategies or policies at a regionalized level or for a society, these kind of methodologies are helpful. For the individual patient selection of who should be regionalized and who should not be, this is less helpful. And I've noted already that this methodology does rely on previously published literature, and as new literature becomes available, a model like this would have to be updated. In conclusion, regionalized rectal cancer surgery is the dominant cost-effective strategy compared to local rectal cancer surgery. Local rectal cancer surgery is rarely cost-effective in this model under even the most extreme sensitivity scenarios. Therefore, supportive financial and logistical measures may be needed for reluctant patients to induce them to the most appropriate care setting. Thank you very much for listening. I'm happy to answer any questions. Thank you. Uh, that was uh very provocative, and, uh, Very provocative and informative, uh, Ira. Doctor Rose, do you have any questions about this? Yeah, I mean, I, I listen, I think, uh, there's a lot of data out there about high volume care, and I think this is an interesting way to look at it from a cost-effective perspective. I think, um, uh, you know, I have, I have one comment which I think you highlighting the limitations that a lot of this type of analysis is based on gigantic leaps and assumptions that can sway things dramatically, which is, which isn't a, a, a, a, a complete negative, it's just a limitation. And I guess my, my other thing, I've always struggled with the why. Why, what is it about regionalization? Is it that you have a high volume surgeon there? Is it that you have anesthesiologists who deal with things more? Is it that you have potentially the same complication rate, but interventional radiology can drain things or you have advanced GI, um, and I, I think it's, it's, it's relevant because if it's just that one group is doing better surgery than another group, then. As you said, on an individual level, if you were tracking your outcomes and you met some bar, uh, then perhaps this doesn't make sense and, and the burden of the patient to travel, it is. You know, can be dramatic for some folks. So just curious about your thoughts on that. You hit the nail on the head with a couple of the key questions with this kind of work. Um, I'll, I'll take what I think is the easier question first, the, this issue of uh the robustness of the model, right? So. It's important to recognize what the analysis that was done here, and this is pretty standard in this methodology, is that you really, really, really want to rev up your model's engine and really push that sensitivity testing as far as you possibly can. If you're not doing that, then you leave it wide open for the kind of vulnerabilities that you mentioned. The fact that we ran through 10,000 iterations with very wide confidence bars around each of those estimates is the best answer that we can say as to why we think our model is better than just chance alone. Um, you know, I, I think that there was a, there was a great paper that we did about a year ago with a totally different topic that ended up being published side by side next to another cost to analysis that showed the opposite conclusion. What was really good about that though, was that, and this is maybe my own bias, but we did a great job of reporting our methods and the other group did a great job of reporting their methods. And because the reporting was transparent, you could actually go line by line through all their assumptions and see what was different, and it was really informative because it basically showed, as a lot of healthcare issues can be, is that it was an international cost assumption versus a domestic US healthcare system assumption, and that's what made the change. But that's where these kinds of models are valuable, right, is that interaction between, well, what does one group's conclusions come to versus what is another's and To your point, when you're doing device cost effective analysis or even something like this where you're thinking about a major policy change, it, it, it's, it's a requirement that anyone that's going to take the conclusions of this model for, for real truth needs to go through and look at the assumptions and make sure that they have, they have face validity for them in their particular context. The more, go ahead, let me just ask you before you go on, Ira, we don't have that much time as uh high volume centers, etc. start to use robotics, etc. like that, and they're sophisticated centers, they're going to increase the cost per case significantly. Also, you find historically that, uh, New York State, for example, looked at individual surgeons and actually published. In the newspaper, the outcomes of individual surgeons for heart surgery years ago. So when patients choose centers, they will enable this would enable them to look at individual surgeons costs and outcomes. It might be the way to go in the future instead of looking at centers, and a guy who does robotic surgery, for example, is gonna increase the cost per case tremendously. What do you think about that? Yeah, I think, I think you have to be careful what you're calling regionalized and, and to Dr. Rosen's point and also to Dr. Ponsky, in this case we used volume as our definition of regionalization, but you could certainly. Say, is it at the surgeon regionalization or is it at the institution that does high volume? A lot of folks have argued that it's, it's the system of care, right? Not that there's one such amazing surgeon who, who has the greatest experience in the world. That's a bit controversial, um, and, and that does play into your point, which is that regionalized care per case is more expensive, and that's been shown. Um, but that's, those are the case costs, right? And so what we include in our model, what's the downstream of costs associated with cancer recurrence, early death, etc. and that's where in 97.8% of our model, uh, the regionalized case won out over the local care. So, um, this is a, a very provocative paper. I will tell you that I've been approaching this from the opposite angle. I believe in de-regionalization. I think if you look at, if you look at where technology is heading, we are moving patients away from going to the big centers. We're doing home monitoring. We're doing, we're, we're trying to get patients away. We're working on telementoring. Uh, the idea of, instead of bringing the patient to the healthcare, bring the healthcare to the patient. And improve expertise at all these centers all over the country to try to, they still have to send them out. But let's try to mitigate that. Let's try to mitigate the need for sending patients and start improving the level of expertise by having more collaboration between experts around the country. Have you talked about that? Yeah, I think the secret untested sauce here is what is a regionalized decentralized model actually look like, right? And so a lot of healthcare systems have moved in this direction, and I don't think the data is good enough to support it yet, but I think that's my hope is that yes, there are certain things that have to be done in a high volume physical room, but for the most part, a lot of this can be exported out, really lean into technology, and really use the fact that, you know, in our system, and this is common elsewhere now. Uh, you know, a community oncologist who might be an hour and a half away from us presents every single one of their rectal cancer cases to an entire board across the entire system, and I think that's where we can overcome some of these limitations, but the data is not there yet. No, it's an evolving picture. It's been looked at for other issues like pancreatic cancer surgery and uh I'm sure it's gonna evolve into hernia surgery in other areas as well. Very nice stimulating paper.