In this episode, Michael Darden, Associate Professor of Health Policy and Management at George Washington University and Ian McCarthy, Assistant Professor of Economics at Emory University discuss how hospitals adjust private insurer prices following reductions in public funding.
Mike Passanante: Hi, this is Mike Passanante. And welcome back to the Hospital Finance Podcast.
How hospitals adjust prices with private insurers following reductions in public funding is an issue that has long been debated. To shed some light on that debate, I’m joined today by Michael Darden who is an associate professor at the Milken Institute School of Public Health at George Washington University and Ian McCarthy who is an assistant professor of economics at Emory University. Gentlemen, welcome to the show!
Ian McCarthy: Hi Mike!
Michael Darden: Hi Mike. Thanks.
Mike: So, Michael and Ian, you are co-authors on a working paper published by the National Bureau of Economic Research that examine how hospitals adjust prices with private insurers following reductions in public funding. Why don’t you start off by telling us why you chose to take a closer look at this particular issue.
Michael: Yeah! This is Michael, by the way. About a year ago, I was thinking about the Hospital Readmissions Reduction Program in a different context with respect to local area characteristics and how the risk adjustment in that program actually works, thinking about disparities, especially with respect to access.
And learning a little bit more about the program, I started to think about what were some of the unintended consequences of pay per performance schemes in the ACA but more generally. And at the time, I was actually unaware of the cost shifting literature—the cost shifting literature being that public reimbursements are reduced and private payments might be different as a result.
And so, I started thinking about that, and I realize that this was kind of a historic literature, one that’s been going on for a really long time, and one that’s quite contentious actually.
And so, I just started thinking more about the question that way. So that’s how I got into it.
Ian: And this debate is important from a policy perspective. As you mentioned, we’re not in universal agreement, across economists, as to how hospitals respond to these public payment reductions. But getting the right answer to that is really important for making the right policy prescriptions because they have completely different predictions.
So, on one end, if hospitals are not cost shifting, then a reduction in public payments will tend to, theoretically at least, lower or at least not increase private insurance payments. And so, in that sense, if we want to control costs a little bit, reducing Medicare payments is kind of a good thing.
There could be the other consequences. But in terms of private prices, reducing these public payments has kind of a positive effect on the on private insurance spending as well. But if we think hospitals are cost shifting, they’re on the complete other end of the policy prediction where a reduction in public payment actually raises private insurance prices.
And so, this issue of cost shifting and identifying whether it happens or how much it happens can generate two completely different policy predictions.
And so, getting a hold on this debate is really, we think, an important thing, and we think something that we’re able to do in this paper based off of the payment reductions that were instituted as part of a handful of provisions in the Affordable Care Act.
Michael: And just to add to that, one of the things that I found fascinating about this literature when I started getting into it was that you have a situation in which private payments from insurance companies to hospitals are negotiated by private insurance companies and hospitals. And so typically, when we think about the bargaining process, we think about two sides that are opposing each other in some way. But cost shifting is something that both sides of the negotiation claim is going on.
Private insurance companies obviously want to be paying lower prices to hospitals. And so they don’t want cost shifting to be going on. And they think that—I mean, there’s this long literature of private insurance company complaining that, for example, Medicare is not paying their share. And then, hospitals claim that cost shifting is going on because, clearly, they want the government to increase public reimbursements, not decrease them. They claim, “Well, this is the only way that we can survive,” by passing these costs off the private insurance companies.
So, I thought that was really interesting that both sides are claiming that this is going on. And we wanted to study it empirically.
Mike: Yes, it’s really a fascinating topic. And I’d like to dive into your analysis. Could you briefly talk about the data set that you chose to look at and the factors in particular that you studied?
Ian: So, we’re essentially using two kinds of central data sets here. There are a handful of things that we incorporate.
But the first dataset is data from the Healthcare Cost Institute. And our co-author, Eric Barrette is there. And this gets us a unique opportunity to look at actual negotiated payments between hospitals and private insurers—not all insurers, but this is a dataset where influencers have contributed data to this data warehouse. And the Healthcare Cost Institute manages those data.
It’s one of the few places where you can actually see negotiated payments. So much of the price data, when people talk about price and healthcare, is often a hospital charge or it’s some approximation based on a hospital charge. And so we’re able to actually see that negotiated price.
So, that’s one central data set we’re using. Another is we can see data—every hospital that takes Medicare patients has to submit essentially some sort of cost reports to the Centers for Medicare & Medicaid services every year. And those are called the Healthcare Cost Report Information System. They’re held in that system.
And so, we use data from those reports to get a handful of hospital characteristics. But essentially, we use those to identify how much a hospital reported being penalized or rewarded under some of these new provisions in the ACA.
So, the provisions in particular are the Hospital Readmission Reduction Program and the Hospital Value-Based Purchasing Program. So these are just two things that are essentially intended to pay a hospital based somewhat on quality or punish a hospital based somewhat on an assessment of low quality in some way.
And so, in those cost reports, the hospital actually reports how much money they received in terms of a bonus from any of those programs or of net penalty. And so we calculate that. And that is our main variable that we’re using to identify whether hospitals change their private prices.
So, if they incur this penalty, do they then try to negotiate a higher price because, effectively, their public payment for Medicare has been reduced.
Michael: And to add to that, I think it’s important to recognize that the payment data that we have is actually really important in answering this question. So, as Ian mentioned, historically, people have used what we think are inferior measures of actual prices.
So, the way it works is that a hospital has what they would call a charge master. And the charge master essentially is a set of list prices for services. And what the insurance company actually pays for a given service is negotiated. And there are usually considerable discounts on that charge.
So, if you use charges as your measure of prices, you’re not actually really studying what is actually paid by the insurance company. And in our work, we’ve been able to show that the correlation between charges and payments is actually only about 0.4. So it’s positively correlated, but they’re not the same thing.
And the literature going back a long time has found mixed evidence of cost shifting. And any time you have a situation in which you’ve got measurement error (where you’re using charges, for example), you’d be concerned of biasing results essentially towards zero. We’re finding various different types of effects all over the map.
So, in our case, we think we’ve got a cleaner measure of what’s actually being paid. And we think that’s a contribution.
Mike: That’s a great set up. So let’s talk about your major findings. And you did note, perhaps not unexpectedly, that hospitals that faced net payment reductions from the HRRP or the HBBP did appear to shift some of the costs to private payers. Tell us what you found.
Michael: So, we’re finding that, on average, when a hospital faces a net reimbursement reduction from Medicare that private payments increased by about 1.5% which is significant. It’s roughly $150 per claim.
And we’re finding kind of across the board on average for hospitals that that effect is the same essentially—statistically at least—for nonprofits and for-profits, which is somewhat surprising. But it may be because we simply don’t observe many for-profit hospitals anymore.
And we find that that effect is largely due to increases in circulatory system and nervous system services.
We should note that we’re studying just acute care admissions in this study.
Mike: And were there situations where cost shifting was more prevalent?
Ian: Yeah, there were a couple settings where we were able to identify what seem to be slightly higher rates of increases on the private side. The clearest one was among hospitals that had a larger share of private insurance patients.
So, in some previous literature, the idea of that is if I’m a hospital that has lots of private insurance patients, then I might have slightly more power in a negotiation with a private insurer because I’m such a large source of care for that insurer’s patients.
So, we think that looking at private insurance share, kind of capturing that type of bargaining power—and we do actually see higher rates of private insurance payment increases among those people with those hospitals with larger shares of private insurance patients.
Michael: But what’s interesting about that too is that you could think of it going the other way. In fact, I think my initial intuition about that was that it would go the other way in the sense that you might imagine that a hospital that has a large share of public patients would obviously be affected by a net penalty or a reimbursement reduction from Medicare more and would have a greater incentive to pass costs off to private insurance companies.
Our finding is that we found the opposite like Ian said.
It’s consistent with another paper that looked at that particular issue with respect to the balanced budget amendment from 1997. But I think it does say something about hospital bargaining power and how close these negotiations are. When a hospital has a large share of private patients, they’re a bigger player at the negotiating table with private insurance firms. And we’re seeing that in the cost shifting data or evidence that we found.
Mike: So, overall, what do you think this research tells us? And what are some potential next steps?
Ian: I think, at a minimum, this tells us, even if you’re highly skeptical of—I’ll step back briefly. The skepticism of cost shifting is mainly, among economists, because it’s theoretically difficult to imagine hospitals raising the price if they could have already raised the price before. And that’s essentially what cost shifting says. Suddenly, something happens on the public insurance side, and now I’m going to raise my price in the private insurance side. And the natural question is, “Well, if you could’ve raise your prices, why didn’t you do it last year or the year before that?”
So, I think all of our next steps are kind of geared around trying to identify what is really happening that would drive this increase in price, what is it that a hospital is doing to be able to negotiate that higher price.
But I think right now what we do feel confident in saying is that there appears to be some amount of cost shifting. And this is an effect that looks to be driven by say a quality improvement or the other arguments that we’ve heard that try to explain how a hospital might be able to translate these things into a price increase.
All of the evidence really is pointing toward this idea that there’s a public payment reduction. And this somehow spurs a negotiation wherein a hospital can raise their price in private insurance without having done anything that we’ve identified so far to change their product. They’re not somehow providing more imaging or having people stay in the hospital longer or doing more intensive services. We’re not finding any of that. To some extent, it’s purely a transfer reduction in payments in one end, higher price in the other, and no other change in the product.
But those are the next steps, trying to really be more confident in the claim that nothing else is changing about the underlying product the hospitals offering.
Michael: And to just expand on that, that’s an important point that some people have read about our study. The Hospital Readmissions Reduction Program creates incentives for hospitals improve their quality, to improve their rate of readmissions, to have fewer readmissions.
So, naturally, economists and others would think that these incentives are real, we would see improvements in quality. That literature is kind of mixed at the moment. There has been evidence of readmissions reductions, but there’s also been evidence of mortality increases of the result.
In our data, as Ian said, we don’t find evidence of a direct link between quality changes and these reimbursement reductions. But even if we did, we wouldn’t call it cost shifting in that case because the underlying product is changed.
But there’s still a message here that prices have gone up. And we see that robustly in our data. And so I think this is kind of a policy relevant finding in that we want to think about what pay per performance schemes really do downstream for the rest of the system.
Ian: In thinking more broadly about these pay per performance models, especially the Hospital Readmission Reduction Program and Hospital Value-Based Purchasing program, you can tell a pretty easy story that would suggest that instruments have been relatively blunt in terms of their effect on hospitals.
So, for example, in the Readmission Reduction Program, I think as of fiscal year 2017, in that year, effectively 80% of hospitals are penalized under this program. And the reason they’re penalized is because you’re assessed separately based on lots of different conditions. And if any of these conditions, you fall below a national average as a hospital, you’re subjected to a penalty.
And so, what ends up happening is that a lot of hospitals are above average in some, but below average in other conditions. And the program keeps adding more and more conditions. So what’s effectively happening is that more and more hospitals are being subjected to a penalty.
So, we are sympathetic to the idea and are taking very seriously the idea that quality might go up from this Readmission Reduction Program, for example, but we have to remember, it’s not just quality say on Medicare, it’s not just reducing Medicare readmissions, it’s also that you have to reduce private insurance readmissions for that to explain what we’re seeing in our data.
And another thing going in our favor here in terms of interpreting this as cost shifting versus something else is that the program just penalizes everyone. And it’s easy to imagine a hospital talking with an insurer essentially in a setting where we’re all penalized. So this is effectively an across the board reduction in public payments, not necessarily something that’s discriminating between hospitals of slightly lower quality or higher quality.
Michael: Yeah, we imagine a situation in which the hospitals go to the bargaining table and say, “Look, we’re under an increasingly tight Medicare situation as a result of these policies.” This is hypothetical because it’s kind of black boxes, the negotiating process. But we can imagine a situation in which a hospital claims to an insurance company, “Look, we’re being docked on services X and Y. And so if you could increase our payments for Z, that would really help us out.”
And again, we can’t identify specific mechanisms like that. But we do identify a significant increase on average that looks to be attributed to these policies.
Mike: Michael and Ian, this has been a very interesting contribution to the overall discussion. If our listeners would like to get a copy of the working paper, where can they go?
Ian: They can go to the National Bureau of Economic Research website which has a link for working papers.
Mike: Well, that’s great. And we’ll certainly include a link to it from the show notes of this podcast. And again, Michael and Ian, thank you very much for joining us today and spending some time on the Hospital Finance Podcast.
Michael: Yeah, thank you, Mike. That’s great.
Ian: Thank you, Mike.