In this episode, we are joined by Marah Short, Associate Director of the Center for Health & Biosciences at Rice University’s Baker Institute for Public Policy, to discuss their new study that looked at the effects of hospital mergers and physician practice acquisitions on hospital quality.
Highlights of this episode include:
- Background behind the study that looked at whether patient outcomes were improved by physician practice acquisitions and hospital mergers.
- Details on the data used for study, including 29 different quality measures and US hospital-level data from over 4,000 hospitals.
- A look at some of the study’s interesting results in areas of physician-hospital integration and market concentration.
- What the study’s findings mean for healthcare providers moving forward.
- And more…
Mike Passanante: Hi, this is Mike Passanante. And welcome back to the Hospital Finance Podcast®.
As hospitals continue to merge with each other, and at the same time, acquire physician practices, the question arises as to whether or not this activity improves patient outcomes.
A new study from Rice University entitled Weighing the Effects of Vertical Integration versus Market Concentration on Hospital Quality shed some light on that dynamic.
Marah, welcome to the podcast!
Marah Short: Yeah, thank you for having me.
Mike: So, let’s start out. And why don’t you tell me what your initial hypothesis was going into the study.
Marah: Sure! So, we hypothesized that through the vertical integration of physicians and hospitals, that there would be decreased fragmentation in the system, meaning better coordination of care among a patient’s primary care physicians, their specialists, their hospital physicians, and that this would improve the patient care.
I mean, this could also be a result of things such as the physicians just being more efficiently monitored if they’re employed or, in some cases, physicians might have salary contract where there’s a pay-for-performance type of arrangement built in where they are kind of induced to meet certain quality goals.
In terms of market concentration, previous research performed by other researchers had shown the potential for increased market concentration to reduce quality. And we expected that to the case in our data as well, but we still want to account for and examine that in our analysis.
Mike: And you looked at a tremendous amount of data. Could you just tell us a little bit about the data you examined and your methods for reaching the conclusions in the study results?
Marah: Yes. So, we looked at nationwide hospital-level data for over 4000 hospitals. And this came from the Medicare’s Hospital Compare website. The years we looked at, 2009 through 2015. So, a wide branch of years as well.
We looked at 29 different quality measures. And these kind of broke down into three different categories. First, we had 16 processes of care which kind of indicate how quickly a patient with certain services—like whether they’re getting the proper antibiotics at the right time before their surgery, or if they’re being treated with aspirin within the proper time period whenever they have a possible heart attack. Those types of things were included in that.
We also had three different 30-day readmission rates. We did this for acute myocardial infarction, heart failure and pneumonia. And those were kind of made-to-use with our data set from 2009 to 2015 because, in the middle of that, around 2013, is when the Medicare Hospital Readmissions Reduction Program started. And those were the first three procedures that were looked at underneath that.
The third set that we looked at were 10 measures of patient satisfaction. And we pulled down from the Hospital Consumer Assessment of Healthcare Providers and Systems Survey—which is a mouthful, so we refer to it as the HCAHPS. And that is a random sample of adults after discharge. And it focuses on the perspective of the patient of what type of care they had including things like whether or not their doctor or nurse always or usually communicated well with them, the cleanliness and quietness of the hospital, and an overall rating of what they felt about their hospital care.
In addition to the quality measures, we have our two key measurements which, as the title of our article suggests, the level of vertical integration and market concentration.
So, for vertical integration, we used the American Hospital Association’s Annual Survey to designate those levels. We had a category just for hospitals that had no integration whatsoever. And then, we had four levels from loosest to tightest ranging from independent practice associations (where a hospital just primarily assists with physicians and contracting with managed care plans) up to the fully integrated organizations where a hospital hires physicians as salaried employees and own their practices.
To measure market concentration, we use the Herfindahl Hirschman Index (HHI). And that’s something that’s commonly used in previous literature of this type. And it’s also what the Department of Justice uses to evaluate potential murderers.
So, using all of that data, we performed regression analysis for each of the quality measures to determine whether or not the vertical integration or the market concentration had an effect on the quality measures.
In that analysis, we used hospital fixed effects to control for the fixed, unobservable differences across hospitals such as their geography or their teaching status. And we also used year fixed effects to control for any changes in quality that might have coincided with these changes that we have an integration as well.
Mike: Very thorough body of information there. And it produced some very interesting results. Why don’t you tell us about those?
Marah: Sure! So, physician-hospital integration had a limited effect. It did not improve the quality of care for an overwhelming majority of the measures we looked at. So, for only 2 of the 29 do we see a statistically significant effect. And that would be for the fully integrated. Hospitals had better adherence to the continuation of beta-blocker than non-integrated hospitals. And for pneumonia, the two tightest integrated had slightly lower readmission rates, but not integrated hospitals.
You know, a large part of that, the fact that we’re not seeing what’s going on with those integration is because we suspect, you know, physicians really want the best outcome for their patients and are going to adhere to these common processes with or without hospital oversight.
I mean, for example, we looked at basic summary statistics for process adherence. They tended to be 90% or higher. And that was across all levels of integration, all levels of market concentration.
So, physicians already have the best interest for their patients in mind.
But on the other hand, what we found that was kind of interesting is that increased market concentration—which lowers our market competition—was strongly associated with reduced quality in terms of patient satisfaction.
Now, with peer competitors, we kind of suspect that it seems maybe there’s less incentives to keep patients happy.
And not I think as important is given the nature of some of our satisfaction measures, they could reflect clinical quality as well. Just take, for example, whether your medications are explained or how well you’re communicating with your patients.
There could be arguments made that the overall clinical quality could suffer if patients don’t understand their care or the recommendations that were made by their physicians either during the hospital stay or post-discharge.
Mike: So, given your hypothesis going into this, did you find these results surprising?
Marah: Yes and no. We expected that the increase in market concentration could be associated with lower quality based on a previous research. And that is what we found in terms of patient satisfaction.
But we also expected that vertical integration could be associated with changes in quality as well—which we found little evidence of. So although increased quality is often cited as an argument or a justification for integration, we don’t find that vertical integration has a broad, positive effect—or really any effect, for that matter—on quality of care in our data.
Mike: So, based on what you found, what do you think this means for healthcare providers moving forward?
Marah: There’s a couple of takeaways that I would like them to kind of hone in on. First of all, I think providers need to think about their overarching goals of integration. So if they’re trying to use integration to increase quality, it may not be the best means to do so.
So, like I said, most physicians are already following a lot of the standards of care. We found that the physician-hospital integration only reduced one of the three readmission rates. And that wasn’t statistically significant. It was not a very large magnitude of change.
In the two tightest forms of integration, they only had readmission rates that were, on average, 0.19 and 0.12 percentage points lower than non-integrated hospitals.
Also, just to keep in mind that when you put in the structural integration through physician employment—so you have the human resource management and the financial management, for example—that’s not necessarily leading to clinical integration. You’re not necessarily going to get those coordinated patient services or the monitoring of the best practices.
And final, I would like to just consider how our finding that patients’ experience can be lower in more highly concentrated markets where there is correspondingly less competition and how that affects them. Although better patient experience may not always correlate with higher clinical quality—which we kind of what we as researchers and I guess providers tend to focus on—like it or not, measuring quality based on patient perception is increasing.
And so, this is becoming more important as consumers continued to use and increasingly use online physician ratings and reviews to choose their providers.
So, I mentioned earlier that could be the case that it matches with clinical quality. But we need to do further research on that to see does it. And if it doesn’t, then we need to think about how we’re presenting data to consumers and possibly develop and provide patients better measures in terms that they can understand and use to select their providers.
Mike: That was some very enlightening research, Marah. If someone in our audience wanted to take a deeper dive into this study, where they can go?
Marah: So, the paper is actually available for free online. You can look through the Baker Institute website. And you can also check out—the journal is published in Medical Care Research & Review. That is free to the public in both of those places.
Mike: Excellent! Well, Marah, we appreciate you coming by the Hospital Finance Podcast today and presenting this very interesting body of information to our audience. Thank you.
Marah: Thank you.