In this episode, we are joined by Joseph Doyle, Professor at the MIT Sloan School of Management, to discuss the results of their study on healthcare hotspotting programs targeting superutilizers.
Highlights of this episode include:
- Background on MIT’s study of a hotspotting program that has received national attention as a promising superutilizer intervention.
- How data for the study was collected and how superutilizers were identified.
- What were the readmission rates for patients in the intervention and control groups?
- What do the results of this study mean for health care providers going forward?
- And more…
Mike Passanante: Hi, this is Mike Passanante. And welcome back to the award-winning Hospital Finance Podcast®. Targeting patients with very high use of healthcare services or superutilizers has become increasingly widespread as a strategy for reducing cost and improving the quality of care. But do these programs really reduce hospital readmissions, a common metric for measuring the impact of these programs?
One recent study looked at exactly this issue. To explain the results of this study, I’m joined by Joseph Doyle, who is the Erwin H. Schell Professor of Management and Applied Economics at the MIT Sloan School of Management. Dr. Doyle studies Public Economics in the areas of healthcare and child welfare. His healthcare research investigates sources of value and waste to inform policies aimed at improving the quality and cost-effectiveness of the US healthcare system. Joseph, welcome to the show.
Joseph: My pleasure to be here. Thanks for having me.
Mike: So why don’t you start out by telling us what your initial hypothesis was going into this study?
Joseph: Okay. Well, there’s a big idea out there about superutilizers. About 5% of the population accounts for half of all spending. And if you go down further, about 1% of the residents in Camden, New Jersey where we were studying this program accounted for about 30% of all spending. So if you could find a way to treat those patients more efficiently and hopefully improve their lives by making them healthier, you can imagine that there can be a lot of savings to be had. This a big idea that Dr. Jeff Brenner is credited with championing from the start. And he won this MacArthur Genius Award around the idea, and it’s a good idea. And so our hypothesis was let’s take the program that Dr. Brenner started in Camden and see if it reduces re-hospitalization for patients, which is, as you said, a common metric for whether a program has a chance of reducing utilization downstream. So Dr. Brenner teamed with MIT. Amy Finkelstein, Annetta Zhou, and Sarah Taubman and I, we randomized subjects to be in this program. This flagship program that was very influential or in a usual care group. That’s because they had capacity constraints. They couldn’t serve everyone who was eligible. So this was an equitable way to see who gets into the program and provides a rigorous method to discover whether the program works or not.
Mike: You talked a little bit about the Camden area and the census that you looked up. But could you tell us about the data you examined more specifically and your methods?
Joseph: Sure. So the Camden Coalition of Healthcare Providers is a data-driven organization. They’ve really helped spearhead the creation of a health information exchange in Camden, New Jersey that links the data from across multiple healthcare providers. And so that’s the data they used to discover who the hot-spotters are, who the superutilizers are, and recruit patients from that feed. And then they can also look at tracking outcomes. And so we used, essentially, that same data to recruit people not just into the program anymore, but into the study, and then to monitor their outcome. So that data is generally electronic health record data that’s from the different providers in the area.
Mike: Okay. And why don’t you tell us a little bit about the results of your research?
Joseph: Okay. Sure. So we found first, that the re-hospitalization rates of people who are eligible for this program is about – over the six months after being recruited about 63% of them will be returning to the hospital. This is much higher than other studies and really pointing out that this Camden model is really focusing in on the most complex patients in their system. And then we found that those that were randomized into the program, they saw a pretty dramatic reduction in the re-hospitalization rates in those six months after joining, but the control group saw the same decline and an explanation for that pattern is mean reversions. So if you are targeting people who have had a run-up in health care costs, it’s natural for them to return to a lower-stakes as just you get a health shock that makes you spend more and then that shock wears off and you start to spend more toward the mean. We saw that that happened both in the treatment and control groups, so we conclude by this comparison that the program itself did not have an impact on changing the re-hospitalization of the subjects.
Mike: Were you surprised by that?
Joseph: I was surprised because I thought that one of the goals of the program was to help patients use the health care system in a new way and a more efficient way to try to prevent exactly this type of outcome. That’s why it was our primary outcome because that was a stated goal of the program and so we went into the research thinking well, let’s see how effective it is. We didn’t expect it to be ineffective. We can say a few things about that, so one is I think it is important, especially for payers if they’re thinking about paying health care organizations to do this intensive program. I should say a bit more about the program. They have a team of about 10 people who help patients by having multiple home visits, attending doctor’s appointments with patients, getting them signed up for social services that they’re eligible for. Trying to get their lives back on track both in the terms of their clinical needs and their social needs. So it’s a very inspiring program in that way and that can be pretty expensive, and payers want to know if we’re going to shell out funding for that kind of preventive service. Does it actually result in lower utilization downstream? And so we didn’t find that for expensive care in the form of hospitalization. We are continuing to do research on this topic for this study to see if affected outpatient care and we did find that it increased the rate at which the subjects were signed up for social programs such as food stamps and that was about a 5% improvement in that, so it seems like they were working toward doing that, but it didn’t have this effect on utilization.
Mike: What do you think this means for health care providers moving forward?
Joseph: So this general idea that we started with that Dr. Brenner is credited with that there are these really complex patients and we have to figure out how to serve their needs in new ways in order to improve their lives and save on costs. I’m still a believer in that idea. It’s just that this type of program, which is a sophisticated navigator program, essentially. They’re trying to teach these patients how to use the health care system in a new way. We found that, at least in Camden, that it didn’t, and for the most complex patients in Camden, New Jersey that it didn’t work to reduce utilization. So what do providers learn? I think that they– we have to go back to the drawing board and figure out well, what can we do. So there are promising avenues such as housing stability is one that has some evidence backing and I think in need of more evidence-base, but it seems pretty promising that it’s among the homeless population, if you can get them housed that that can have large improvements in terms of health, health care spending, as well as criminal justice spending in ways that it could be worth the investment. If we talk with the folks in Camden, the idea that they helped them navigate a system that isn’t really designed to treat patients with such complex needs is one of their takeaways when we’ve had discussions with them that we should think about other ways to treat these really complex patients. So some ideas include ambulatory ICUs, heavy monitoring of these patients, using digital tools outside the clinic, just try to catch escalations before they get too serious. Just we have to have new ways of treating these patients instead of just navigating them better toward the system that already existed, at least in the Camden context.
Mike: Well, it’s an important topic and thank you for moving this discussion forward. It was a pleasure having you on the show today.
Joseph: My pleasure as well. Thank you.