Carregando...
Loading...
Please wait
Host Dr. Davide Soldato and guests Dr. Kerin Adelson and Dr. Maureen Canavan discuss JCO article "Association Between Systemic Anticancer Therapy Administration Near the End of Life with Health Care and Hospice Utilization in Older Adults: A SEER Medicare Analysis of End-of-Life Care Quality," highlighting adverse outcomes for patients who receive any type of systemic anticancer therapy(SACT) at EOL (end of life) and the need for better communication between oncologists and patients regarding expected risk and benefits of such treatments to properly align goals-of-care.
TRANSCRIPT
Dr. Davide Soldato: Hello and welcome to JCO After Hours, the podcast where we sit down with authors from some of the latest articles published in the Journal of Clinical Oncology. I am your host, Dr. Davide Soldato, medical oncologist at Ospedale San Martino in Genoa, Italy.
Today, we are joined by JCO authors Dr. Maureen Canavan, epidemiologist and associate research scientist at Yale Cancer Outcomes, Public Policy and Effectiveness Research Center; and by Dr. Kerin Adelson, Chief Quality and Value Officer, medical oncologist, and clinical researcher on health services and clinical care delivery at MD Anderson Cancer Center.
In the manuscript "Association Between Systemic Anticancer Therapy Administration Near the End of Life With Health Care and Hospice Utilization in Older Adults: A SEER-Medicare Analysis of End-of-Life Care Quality." that you recently published in the JCO, you performed an analysis that included more than 30,000 older adults in the SEER-Medicare database, and you observed that 7.6% of these patients received any systemic anticancer medication within 30 days of death. So, I wanted you to explain why you thought that this was a priority right now, and whether there was any previous data that was published in the literature, and if you think that there was any significant gap in the literature that led you to the research you just published.
Dr. Kerin Adelson: We have published a series of articles looking at real-world trends in patterns of care, particularly related to systemic anticancer therapy at the end of life. This has been gaining increasing focus in recent years because of the understanding that when patients stay on systemic anticancer therapy, that is often a surrogate for a lack of goal-concordant care.
So, patients who continue to receive systemic therapy have worse quality of life, are more likely generally to have a medicalized death, and less likely to use hospice. And what our prior work has shown is that more and more we are seeing patients using immunotherapies and targeted therapies towards the end of life. No prior work had really comprehensively examined whether these novel therapies were associated with those same patterns of care increases in acute care utilization and decreases in hospice.
Dr. Davide Soldato: So basically, the data that we had up until that point was mostly with cytotoxic chemotherapy, and the emergence of this new treatment, which frequently are thought to be less toxic and so less problematic also in the end of life, led to this research. Is that correct?
Dr. Kerin Adelson: Correct.
Dr. Maureen Canavan: I would also build on that. I think that as the landscape of cancer care changes, it is important to really understand the availability of treatments, but then also, as Kerin noted, it is important to focus on goal-concordant care. We have established literature, studies we have done and some other studies that have looked at cytotoxic chemotherapy, but with the emergence of these targeted therapies, we really did not know a few things. We did not know the rates of utilization in a large national population, and how that was associated with these elements of medicalized death like ED use, hospitalizations, acute care use.
So this was really a question that we had going into it. How can we expand the knowledge base so that both patients and providers can be more cognizant when thinking about goals of care conversations and ensuring that that is in place?
Dr. Kerin Adelson: And our work has kind of evolved to answer some critical questions. So, one of our early papers looked at different rates of systemic anticancer therapy at the end of life, and that is where we showed that we were seeing a lot more immunotherapy and targeted therapy. And then we asked the question, well, oncologists generally when they give these treatments, they are hoping that those treatments are going to work and help the patients live longer. So we did another paper where we actually looked at practices who were more aggressive near the end of life and whether they had better overall survival than practices that were less aggressive, accounting for the fact that there could be populations of patients who benefited.
And in fact, we showed there was no survival difference. So then this paper sort of answered the question: Well, if it is not having benefit, is this treatment actually doing harm? And this study gets at that question: What are the harms of continuing patients on therapy past the point of benefit?
Dr. Maureen Canavan: And I think building off of that, the use of the SEER-Medicare database is a quite robust database. So in this, we have very specific data we can track. We can track the exact type of treatment they had, you know, was it a targeted therapy? Was it immunotherapy? So looking at those subclasses of therapy. We were also able to directly link it within that time frame to the acute care utilization, a limitation that we had in some of our previous work that that data was not always available.
So it is more focused in the sense that we were looking at older adults, so patients 66 years of age and older, but we were able to get those individual metrics. So to Kerin's point, we did not see the survival benefit. What do we see then for these medicalized death elements? So the higher rates of all of them across the board.
Dr. Davide Soldato: So coming back to the cohort and to the data that you utilized, Dr. Canavan mentioned the use of the SEER system to analyze these data. You already mentioned that you included mostly older adults, so those aged 66 and more. And also there was a little bit of restriction regarding the fact that the patient needed to be covered by Medicare in the last year of death concerning Part A and Part B, and the last 30 days from death concerning Part D. So I just wanted to ask a little bit of a question regarding these findings and whether you think that we also need additional work, especially in the younger population because I think it is something that all of us who work in oncology have seen. The aggressiveness, and this is also something that you showed in your data, tends to increase as the age of the patient tends to decrease. So we tend to be more aggressive towards younger patients. So just a comment on that on the population and generalizability of the findings.
Dr. Maureen Canavan: Yeah, I will start with the data question element. Thank you. I think there are a few things to point out for that. So in terms of the restriction to ensure that they had continuous Part D coverage, that was necessary for us to track their oral medication use during that time. So kind of an easy response. The Part A, Part B requirement, it is actually pretty widely used in studies of SEER-Medicare data, and that is you want to establish the patient population, that they are not getting treated with another insurance provider in some way that you are not able to track.
So that ensures that we can track not only their systemic anticancer therapy use but also when we are trying to make sure that we are controlling for confounders like chronic conditions and stuff, we are able to track the presence of chronic conditions. So we wanted to make sure we were not biasing the data, so I think that was an important consideration.
You do point out very wisely that there are then limitations with the generalizability, and I think we would be lacking if we did not account for that. But I think it is important to establish this baseline relationship association, and then you can step out, we will say, to more diverse populations. So I think we could potentially maybe try to relax the timeline to see if people that might have influx in and out of the Medicare system are still seeing those same rates. I think it is likely they would.
But I think to the bigger point that you bring up is that establishing this within the older adults where, you know, we do see as they get older maybe less rates of systemic therapy, extending it to the younger population. There is a challenge with that in that just that data is not available to the robust level that SEER-Medicare is. Both Kerin and I have noted that there is the possibility to look within one specific insurance provider type. Again, recognizing the limitations of the generalizability, but always slowly pushing the needle, finding out more about younger adult populations.
And I think this is maybe in an ideal world, but setting the precedent that we really do need to track this on a national scale within younger adults because they do have the need. We do see these higher rates of utilization, and really making sure again with the mindset always of the best interest of patients and the most informative to providers in how we are looking at care.
So I think generalizability is definitely a goal. However, there are limitations of the availability of data for younger populations and I think that they are a necessary restraint that all researchers should acknowledge.
Dr. Kerin Adelson: Yeah, I think it is important for our audience to understand that health services research and large database research is really limited by what databases are available and what are the characteristics of those databases. So we have done a lot of work in an electronic health record database, and there you can get certain kinds of granularity that you may not be able to get in a payer or a claims-based database. But what you do not get is that comprehensive look at, say, what happens if a patient goes to another practice.
Claims-based databases offer you that, but research on US populations is limited by our payment system. So when you look at younger patients, there are so many different insurance companies that when you are trying to get that comprehensive view, it can be hard or very expensive actually. These commercial insurers will sell their data to different databases. So for us, the largest single payer in the United States is the US government, and that is for patients who are over age 65, and that is why you see lots of US-based studies done in the Medicare population.
Interestingly, a recent paper by a Canadian group showed very, very similar patterns. It was a significantly smaller study but, right, Canada is a single-payer system and so they were able to really look at all ages, and we did see the same patterns of care in a different payment system.
Dr. Davide Soldato: Going back a little bit to the type of treatments that were observed in your manuscript, so we start from a 7.6% of patients who received any type of systemic anticancer therapy within 30 days from death. And when we split the different categories that you analyzed, which I think is a very strong aspect of your manuscript, we see that more or less 50% of the patients received chemotherapy, 20% more or less received immunotherapy, more or less 20% targeted therapy, and then there is a combination of those agents. So just wanted to have a little bit of your opinion compared also to the data that you already published and that you mentioned before. Was this in line with previous data? Was there anything surprising about this? We saw a little bit of a raise in the use of immunotherapy and targeted therapy as you were saying, but still, there is a very high proportion of chemotherapy, 50%.
Dr. Kerin Adelson: So I think that really, really reflects the time period in which we studied where immunotherapies were gaining ground. There was tons of excitement and we were seeing this shift. I bet if we do the same study in five years that chemotherapy percent may even go down to half, and we are going to see more and more targeted and immunotherapies, and that is just reflecting the pattern of drug discovery that we are seeing.
Dr. Davide Soldato: Coming to the real question that you wanted to answer with this manuscript, so is systemic anticancer therapy associated with worse outcomes in terms of healthcare utilization and use of hospice resources? Was there any hint that for example immunotherapy was related to less of these adverse outcomes?
Dr. Kerin Adelson: So I will be honest, I was a little bit surprised that the combination of chemotherapy and immunotherapy was that much more strongly correlated with acute care use at the end of life. You know, I had really thought most likely that what we would see were similar rates. And we did. Each different type of systemic anticancer therapy was associated with significantly higher odds of ending up in the hospital, going to the ICU, dying in the hospital, going to the ED. But that group that got dual therapy was that much higher, you know, over three times the risk.
And that surprised me because what it suggested is that there is likely a component of treatment toxicity that is leading to some of the acute care use. It is not simply just a constellation of patients who have not yet transitioned towards hospice or palliative care or end-of-life care who are then more likely to end up in the hospital. But the fact that we see a difference between, say, single-agent immunotherapy and dual combination with chemotherapy does suggest that the treatments are actually contributing to some of what we are seeing.
Dr. Davide Soldato: But still, all of the treatments that you evaluated were still associated with higher healthcare utilization. Like there was no signal that, for example, giving immunotherapy at the end of life was not associated with these adverse outcomes. Correct?
Dr. Kerin Adelson: Correct. And you will find oncologists out there who will say, actually, these treatments are so good that they might actually lower rates of hospitalization because they keep patients healthy. And certainly, that may be true upstream or earlier in the course of disease, but at the end of life, any form of systemic anticancer therapy is really a surrogate marker for lack of transition towards what is likely appropriate end-of-life therapy.
And I just want to point out that time spent in the hospital, going back and forth to invasive procedures, going to the intensive care unit, even going back and forth to an infusion center, that is time that is not spent at home with loved ones for people who have very little time left to live.
Dr. Davide Soldato: Thank you very much. That was exactly the point that I wanted you to stress because I think it is really the most important message that we can get as oncologists from this manuscript. Like there is no treatment that is not associated with potentially harming our patient and, as you were saying, taking off time with loved ones in a critical period of the life of these individuals who have been diagnosed and treated for cancer.
So, basically what we saw in the paper was a 7.65% utilization of systemic anticancer therapy. And I might imagine that for some oncologists or for some hematologists that might not actually be that much. Like they could potentially say, "Okay, but it is like 7%, it is not that high. I would have expected something higher." So I just wanted a little bit of perspective regarding also quality metrics that we have available for these types of indicators at end-of-life care. What would be the appropriate percentage of people receiving any type of treatment within 30 days from death?
Dr. Maureen Canavan: A couple caveats, as a data person I always like to give those. This was among all cancer patients, so not necessarily patients that had been on active treatment. So I think that number was actually quite lower than when we looked in another study about patients that had chemo within the last year, so on, you know, active treatment. So I think that is an element to take into consideration is that those numbers will vary based on who your denominator population is. So that is important to consider.
Additionally, the National Quality Forum, they call for reducing rates of systemic therapy at end of life. But I think they, similar to how I would be, are cautious to point out this is the exact number, or it should be zero. Because there are cases where you have to go in line with patient preferences. And if a patient is very adamant that they want to continue treatment, that needs to be a decision that comes between them and their provider.
So, you know, the zero, though sounding ideal to us who want to encourage transitions and encourage goals of care conversation is a nice number, it is not a realistic. So, to evade your question completely, I do not think there is a set number. But the goal is to make sure that both patients, providers, everyone is informed and is making the best holistic decision. So there is this natural tendency, I think, to keep fighting both for the patient and the provider to try to beat something, but recognizing the point at which we are beyond a benefit of treatment and what would be most beneficial to the patient in terms of getting back to that idea of, you know, the time with their families and whatnot.
So is the number zero? No. Could it probably be lower than we have? I think yes, definitely.
Dr. Kerin Adelson: I completely agree with everything Dr. Canavan said. I think one of the other challenges is that this data isn't being tracked and publicly reported across the world. And so what that optimal rate is, is a little unclear. We see different rates also depending on the population included. So one of the things Dr. Canavan said is our database included patients who were likely treated long ago for cancer and cured of their cancer. So they were less likely to die on systemic therapy. But until everybody starts tracking and reporting, it is really hard to know where we are as a country or really as a global population, and then what are the bars that we want to achieve in driving down the rates.
I think some data shows that probably something in the range of 10% or below, you know, for patients who have more active cancer is probably where we should be going and driving towards. But until we have more public reporting of these metrics and consistency in how we measure them, it is really hard to come up with a single number.
Dr. Davide Soldato: I have the impression that sometimes there is also a little bit of difficulty for the oncologist or the hematologist to really understand who are the patients who are approaching end of life. So there has been some data and you also report some of them in the discussion of the manuscript regarding, for example, prompts inside of the electronic health records or the use of artificial intelligence to try to predict what is the disease course. So just wanted a little bit of perspective if you think that these tools could potentially be helpful and if you think that we will be able at a certain point to implement them in routine clinical care.
Dr. Kerin Adelson: I have been working on trying to do this actually at MD Anderson and coming up with a really reliable data tool that will tell us who are the patients who are going to die in short order after receiving systemic anticancer therapy. And it is not that easy, I will say. So, you know, I think we all want this amazing machine learning model that is incredibly reliable. But like any statistical test, there are problems, right? So a very sensitive test that is going to identify high, high risk of dying at the end of life is going to be compromised by false positives. And when an oncologist knows that the test might be a false positive, it becomes very hard for them to take action on it.
Similarly, you know, a very, very specific test is going to be compromised by false negatives. So in that case, you could end up having patients who are at risk for dying and still treating them with chemotherapy. And so, you know, I think in the end we need some tools. It will be great if machine learning becomes very reliable and we have the right structured data elements in our electronic health records to give these reliable prediction tools.
But I think there are some basic things that we all know, and those are the markers of chronicity of cancer. So patients who have had multiple lines of therapy already, right? Past the point of clinical trial benefit. Patients who have lost significant amounts of weight. Patients who are not getting out of bed and have worse performance status. Patients who are increasingly confused, right? And not mentally engaging the way they did previously. Those markers have been shown in numerous publications by a colleague of mine, David Hui and others, to really be pretty strong predictors, and they resonate with clinicians more than a machine learning score might. You know, I think when clinicians do not understand what the elements in a machine learning tool are, they are less likely to trust it and more likely to say, "Oh, it is a false positive or a false negative." But very few clinicians can argue against the fact that the patient who hasn't gotten out of bed in two weeks is somebody who is less likely to benefit.
Dr. Davide Soldato: Dr. Adelson, I would like to close this podcast and I would like to thank you again for joining us today.
Dr. Maureen Canavan: Thank you so much.
Dr. Kerin Adelson: Thank you so much for having us.
Dr. Davide Soldato: Dr. Canavan, Dr. Adelson, we appreciate you sharing more on your JCO article titled "Association Between Systemic Anticancer Therapy Administration Near the End of Life With Health Care and Hospice Utilization in Older Adults: A SEER-Medicare Analysis of End-of-Life Care Quality."
If you enjoy our show, please leave us a rating and review and be sure to come back for another episode. You can f ind all ASCO shows at asco.org/podcast.
The purpose of this podcast is to educate and to inform. This is not a substitute for professional medical care and is not intended for use in the diagnosis or treatment of individual conditions.
Guests on this podcast express their own opinions, experience, and conclusions. Guest statements on the podcast do not express the opinions of ASCO. The mention of any product, service, organization, activity, or therapy should not be construed as an ASCO endorsement.
Disclosures
Kerin AdelsonStock and Other Ownership Interests: Carrum Health Consulting or Advisory Role: Abbvie, Quantum Health, Gilead SciencesPatents, Royalties, Other Intellectual Property: Genentech Other Relationship: Genentech/Roche
Employment: Emilio Health/Brightline Health(An Immediate Family Member) Stock and Other Ownership Interests: Emilio Health/Brightline Health, Lyra Health (An Immediate Family Member)