Potentially avoidable hospital readmissions in patients with advanced cancer

Dr. Pennell talks with Dr. Patrick Conner Johnson, hematology oncology fellow at the Dana Farber Cancer Institute Massachusetts General Hospital fellowship program about his and his co-authors' new study titled "Potentially avoidable hospital readmissions in patients with advanced cancer."

Hello and welcome to the ASCO Journal of Oncology Practice podcast. This is Dr. Nate Pennell, medical oncologist at the Cleveland Clinic and consultant editor for the JOP. Potentially avoidable hospital readmissions are a major target for reducing costs in the health care system. However, for cancer patients, the issue goes way beyond cost.

Many of our advanced cancer patients have a limited lifespan. And every unnecessary day they spend in the hospital is one less day they spend at home with their loved ones. The reasons behind cancer patient readmissions may differ from other types of patients. And so broad efforts to reduce hospital readmissions may not apply quite as well to this population unless we understand the specific reasons behind readmissions for our vulnerable population.

Today we're going to be talking about this topic with Dr. Patrick Conner Johnson, hematology oncology fellow at the Dana Farber Cancer Institute Massachusetts General Hospital fellowship program about his and his co-authors' new study titled "Potentially avoidable hospital readmissions in patients with advanced cancer," which was published in the May 2019 JOP. Welcome, Conner, and thank you for joining me today.

Nate, it's truly an honor to be on the podcast. I appreciate your time and [INAUDIBLE].

So first of all, can you give us a little bit of background on why hospital readmissions are a topic that people are talking about? How big of a problem is this and what's the scope of the issue?

To start with, just hospital admissions period are a major topic within cancer care and all of medicine. If you go back to the health care cost and utilization project report since 2009, more than 4.5 million cancer related hospitalizations amongst adults. So that's hospitalizations.

Some percentage of these are potentially avoidable both in the general medicine literature and in the oncology literature. and both from a cost and value standpoint and from quality of cancer patients' lives, I think these are important issues to think about in terms of addressing.

And then when we looked, particularly in our study, we focused on readmissions by which we define having a panel of patients who already had admission once and then looking at subsequent admissions after that.

There is a fair amount of literature out there looking at causes of hospital readmissions as a target for improving value based care. What do you think is different about cancer patients that makes this something we need to study uniquely in them?

I think, in general, amongst a variety of different subspecialties, there's an importance in focusing on targeting each individual population to understand the nuances of that population, whether that's a literature on COPD or heart failure. And oncology in particular is still a wealth of drugs with a wealth of potential consequences and with a sub-population within oncology of advanced cancer patients who have defined limited lifespans.

There's a whole host of factors and different unique circumstances that could potentially affect their readmission profile a little bit different than other general medicine populations. And I think the greater understanding we have of each subset of patients within a number of disciplines is probably going to target our interventions to be more likely to be successful.

And I think that makes perfect sense. One of the other things you mentioned in the background section of your paper is that many of the studies looking at potentially avoidable readmissions have not incorporated patient reported outcomes. And why do you think that would be an important thing to include in the study?

I think looking across oncology care, the study by Dr. Schrag and Dr. [? Basch ?] and colleagues comes to mind. There's been an increasing interest in incorporating patient reported outcomes in order to better pair these with our other outcomes.

And I think that our hope with this was to gain a greater understanding of what kinds of symptoms and other things patients report and trying to identify if there is any correlation with admissions. And the same thing's being done across a number of different facets of oncology care.

Yeah. I'm not sure people outside of oncology understand that there's a significant percentage of our inpatients are admitted for symptom control specifically as opposed to general medicine problems like pneumonia or blood clots.

And so definitely in that case being able to assess their symptom burden makes perfect sense when you're trying to do the kind of study that you're doing. So speaking of your study, can you walk us through the design? How did you put this together?

So this was a longitudinal cohort study of consecutive patients that were admitted to the hospital. And patients were enrolled. And as part of their enrollment, their symptoms were assessed at the time of their enrollment.

So this was a one time symptom assessment within two to five days of their hospitalization when they completed a symptom burden questionnaire essentially. We took available data that we had.

And we had two coders go back, review the medical record with a focus on the discharge summary to try and understand the reason for hospital admission. And then we had a peer review system to try and identify which readmissions were potentially avoidable.

And we used some adaptive criteria, which has been utilized in some other studies in leukemia an GI cancer. And that process was essentially two physicians doing an initial coding review using these criteria. And then anything that was considered potentially avoidable by either of those physicians went to a panel that included two board certified oncologists.

Yeah. I'm curious about this. I know that there are published methods for how they do this. But can you give us an example? How do you determine if a patient had an a potentially avoidable readmission? What's an example of something they might find.

There's no question that it's a challenge. And it's rife with some subjectivity at times. In order to try and minimize that, we have defined criteria. For an example, one of those is premature hospital discharge, which in this study was defined as being readmitted within seven days of discharge with identical symptoms to the prior admission. And that's by a review of the hospital discharge summary.

OK. That makes perfect sense. And you would think that that would be a significant risk for patients since the length of stay is such a big target for hospitals to try to reduce costs. And also our patients typically want to go home.

You know, we have to sometimes convince them to stay when we think they need to stay. OK. So if they were readmitted within seven days with the same symptoms that's how you determine that. So tell us some of what you found.

Thanks, again. Major points from our paper are, first, similar to some other studies that try and look at potentially avoidable hospitalizations at large. More than 30% of the admissions were qualified as potentially avoidable readmissions, which I don't think is anything that has ever been described.

Again, that's fairly consistent with actually the general medicine and with some other oncology literature, but it speaks to the importance of the topic and the sizable possibility of interventions down the road. And the two major important risks that we identified in a multivariable model were marital status, and which we think is a proxy for social support and was protective against potentially avoidable readmissions, and higher physical symptom burden.

And those with higher physical symptom burden were more likely to have potentially avoidable readmissions. And finally, when we looked at the most common reasons as using our criteria for potentially avoidable readmissions, those were premature discharge from a prior hospitalization and also not having what's called a timely follow up, which was a seven day follow up. And so that speaks in our mind to the hazardous time period that is the discharge is fraught with a lot of possibilities of difficulty in terms of making that transition.

Yeah. I think that makes perfect sense. I know our institution in particular has instituted a mandatory call from the outpatient team to a patient the day after they're discharged to just check and see how they're doing. And then we try to get everybody an appointment within seven days although that's not always possible. Are there any other interventions that you think would come out of what you found that might help reduce potentially avoidable readmissions?

I think from the social support side of things, given that social support can be challenging, there is an idea that if we identify patients with that limited social support that that might be the patients that we target for patient navigation programs, more intensive social work involvement programs, or a specially designed care transition programs at hospital discharge as well as potentially patients who have a higher physical symptoms.

And the association with higher physical symptoms also makes perfect sense, although that's always a challenge to address appropriately. I know that there's a lot of focus certainly in solid tumors about integrating palliative medicine and [INAUDIBLE] of medicine support for patients with solid tumors to control that. Is that something that you think could be helpful in this setting as well to help reduce readmissions?

Absolutely, Nate. I think that the hope would be that this also raises a possibility of identifying a patient population that may already be plugged in with palliative care. But if they're not, this could help identify another group of patients that can benefit from integrated palliative care with the hope being that we can identify interventions that can reduce their hospitalization burden.

And where do you think we're going to go here in terms of research. So you've identified some nice potential associations. And there's some low hanging fruit in terms of arranging fast follow up. But what's the next steps in terms of trying to reduce potentially avoidable readmissions for our patients?

I think an integrative palliative care interventions for those with high physical symptom burden. And I think that targeted interventions such as more intensive social work involvement or care transition programs for those with limited social support would be potentially good intervention based studies to start with.

I also think that, as you mentioned, you raised good points about there's still a good bit of research to having a greater understanding within the world of oncology. What is the ideal follow up after discharge for each sub-population even within oncology? And there's probably a great deal more research into understanding that as well as more about the physical symptom burden of hospitalized patients in oncology.

I don't know what your opinion is. Do you think we'll ever be able to avoid almost or all potentially avoidable readmissions?

No. I don't think so. I think that it's a patient population that has a high symptom burden and has a high complexity of care. But I do think that any interventions to reduce the burden of hospitalizations could potentially have far reaching consequences.

I know. I agree with you. I mean there's no way we'll ever be able to avoid this completely. And we all have experience with patients who we can tell when we're getting ready to discharge them that they're at high risk of not successfully transitioning home and yet they want to try.

Perhaps they might be better off in a facility where they could have a higher level of care. But they really want to try to get home. And we want to give them the chance to succeed. And it's just not always successful. So giving them every resource that we can sounds like the right thing to do.

[INAUDIBLE] I agree with you totally. I think the other point to mention is just that the care transition time is a very fragile one. And other interventions to try and improve that transition period as well are something that would be of interest for us or other folks to explore around this topic.

Yeah. It sounds like that would be ripe for a quality of care study to look and see if really intensive interventions in that first few days or a week after discharge can reduce this. I know that we've moved forward with doing that, but I'm not sure if we have any data that it's effective. But it certainly makes sense that it would be helpful. Connor, thank you so much for talking with me today.

Thank you so much for having me.

And I also want to thank all the listeners out there who joined us for this podcast. The full text of Dr. Johnson's paper is available online at ASCOpubs.org/journal/JOP in the May 2019 issue of the JOP. This is Dr. Nate Pennell for the Journal of Oncology Practice signing off.