A study finds limited changes in hospice utilization, highlighting challenges in real-world implementation.
In an interview at the 2024 American Society of Clinical Oncology annual meeting, Ravi Parikh, MD, MPP, assistant professor of medicine and health policy, Perelman School of Medicine, University of Pennsylvania, discussed the outcomes of a palliative care study at Tennessee Oncology, providing insights into the challenges and limitations of evaluating hospice utilization and quality-of-life improvements in the real-world setting.
This transcript was lightly edited.
The American Journal of Managed Care® (AJMC®): You found no differences in quality of life or hospice use. Tennessee Oncology was already a longtime Oncology Care Model (OCM) participant; would you expect to see differences at a practice with a less robust commitment to value-based care?
Parikh: I'll make 2 comments. First, we found reasonable numeric changes in hospice utilization. I think the reason they didn't reach statistical significance is largely due to the fact that we powered the trial on palliative care referral, not on end-of-life care. So, we didn’t observe enough deaths to have the hospice outcomes be adequately powered. That's my thinking. So, we would need a larger, better powered study for end-of-life care to be able to comment on that.
The rates of late hospice referral, the rates of the bad outcome decreased from 14% to 10%, which is a small, but I think, still a meaningful increase if this were integrated as part of a large value-based care model. The fact that they were relatively low to begin with, and that we didn't see a large change could be related to the fact that this is a relatively high performing practice. I would suspect that across ranges of practices, this would still have an effect given what we know about palliative care in patient-level efficacy randomized trials.
The second part that you mentioned about quality of life is really important, because quality of life and patient-reported outcomes is the metric that I care about as an oncologist, more than whether someone seeks palliative care, more than what the care [is] that they can receive at the end of life. I want to make sure my patient's quality of life is better through either controlling their symptoms, managing their psychosocial distress, or making sure that they have robust plans for what happens next, even including the end of life. And so, the fact that we didn't see changes there, I think is far more due to the fact that we just had a lot of variable levels of follow-up and survey completion that resulted in foregoing our study, rather than being reflective of the overall intervention.
We know that when palliative care is delivered in a structured, standardized way, in patient-level randomized trials, that it improves quality of life. So that's not up for debate here. I think the question is in what we call an effectiveness study, which is in the real world, in large numbers of patients, where patients have the opportunity to refuse or accept palliative care, and more clinicians have the opportunity to opt in versus opt out of palliative care, do we see those quality-of-life benefits hold up?
I would argue that we were largely unable to see that because patients didn't fill out their quality-of-life surveys enough, because this is a real-world population, and patients could choose whether to do it or not. It's not like we had a dedicated research coordinator to be handing out these surveys every time they come into clinic, like they do in a standardized clinical trial. So, this is largely just a function of our trial design, and not due to any meaningful conclusion that we can draw about the impacts of palliative care on quality of life.
AJMC: Can you discuss the reduction in end-of-life chemotherapy in the intervention arm? Are you able to quantify how this translates into savings per patient?
Parikh: I couldn’t directly do that in this study, because we haven't yet aggregated our end-of-life chemo findings into claims data that would allow us to calculate savings. I'll note that the reasons for that are not because we don't want to, it's because in this real-world practice, there's a variety of different payers that they serve, and trying to get payer data for every single one of those patients that might have Medicare or Medicaid, or commercial insurance is just a really tricky task. So, while we can’t do that, we have run separate studies in the advanced care planning space, where we've been able to estimate savings associated with avoiding chemotherapy near the end of life.
In a separate trial, we ran a default machine learning-based nudge to clinicians to discuss advanced care planning. It wasn't palliative care referral, but it was an advanced care planning referral. This was work that was recently published in the New England Journal of Medicine AI (NCT03984773).
We found that for every patient, there was an average of $13,000 in end-of-life savings in the last 6 months of life. That was almost entirely driven by decreased chemotherapy received. And so, if we extrapolate those savings to this particular study, just to walk through how we might be arriving at those calculations, we found that the rates of end-of-life chemotherapy decreased from 16.1% to 6.5%. That is an 8.6 percentage point difference. What that means is that for every 100 patients who get this intervention, 8.6% avoid chemo near the end of life. And so, if we're talking about for that given cohort of 100 patients, 8.6 instances of avoided chemotherapy translate from our previous study translates to approximately $111,800 saved in end-of-life savings per 100 patients. Extrapolating this to your population of 100,000 patients that might be served by a typical community or a large community oncology practice, we're talking about several $100 millions of dollars of savings, even up to $1 billion dollars of savings that could be encompassed through just avoidance of end-of-life chemotherapy alone. So those are very rough back of the envelope, modeling statistics. We're going to need to validate this with actual numbers in this trial. I don't want to overstate these findings too much, but I would say that even $111,000 of savings per 100 patients targeted in this intervention is still a very meaningful difference.
AJMC: What unanswered questions do you have following these results? What is next?
Parikh: I would say there are probably 3 lines of work that this work directly helps to inform. So the first is, what happens if we make the algorithm better? If we target this algorithm, so that it's more correctly identifying individuals who are going to be in trouble either from a symptom management standpoint, or from a death standpoint, then we can target the referrals in a more robust framework to people who may need them and concentrate savings and hospice referrals and things like that to individuals who are most likely to pass. And so, making the algorithms better is a key area of interest here, and in particular, incorporating variables into those algorithms, like patient and caregiver reported metrics that are usually not included, but that can improve the ability to detect individuals with high levels of palliative care need much better.
The second line of work is tailoring the intervention to make it a lot better. And so, [there was] some work presented in the same session by Jennifer Temel, MD, FASCO, medical oncologist, Massachusetts General Hospital, that [focused] on a stepped framework towards palliative care (NCT03337399). So, a more routine framework of delivering palliative care at structured time periods than we used in our trial, where we've largely left it up to the palliative care clinician, as to whether a patient got longitudinal palliative care follow-up or more ad hoc follow-up. And so, thinking about a more structured delivery mechanism of the palliative care intervention, I think would allow us to still adhere to this more precision framework that we introduced in our trial, but have the effects of the palliative care be a little bit stronger.
And then the third area is just running more and more trials of this type of precision palliative care framework across different care delivery settings. While we targeted a community oncology practice that was relatively resource constrained, this is a practice that are a series of practices that had a lot of interest in value-based care and where there was a lot of incentives aligned. And so, we need to target this towards settings that are more diverse, for example, populations in inner cities, or that serve of a higher prevalence of Black and Brown individuals and practices that aren't necessarily bought in to devalue base calculus, because we've shown that this can work in the value-based community oncology setting, there's been a ton of work that suggested highly resourced academic settings can benefit from this, but most patients are still getting their care from Mom and Pop oncology practices in the community. And so, we need to show that this framework works in settings that might not even have an onsite palliative care provider and need a more telephonic-based intervention. Because until we do that, we're not going to realize the potential of what this has to offer.
Exploring Racial, Ethnic Disparities in Cancer Care Prior Authorization Decisions
October 24th 2024On this episode of Managed Care Cast, we're talking with the author of a study published in the October 2024 issue of The American Journal of Managed Care® that explored prior authorization decisions in cancer care by race and ethnicity for commercially insured patients.
Listen
Community Investment, Engagement Are Essential to Fully Address Cardiovascular Health Disparities
November 19th 2024Community-based researchers can teach clinicians a lot about how to best approach underserved populations disproportionately impacted by cardiovascular health complications.
Read More
Examining Low-Value Cancer Care Trends Amidst the COVID-19 Pandemic
April 25th 2024On this episode of Managed Care Cast, we're talking with the authors of a study published in the April 2024 issue of The American Journal of Managed Care® about their findings on the rates of low-value cancer care services throughout the COVID-19 pandemic.
Listen