Radhika Jain on Healthcare in the Public and Private Sectors

This episode is the 12th in a miniseries of weekly short episodes featuring young scholars entering the academic job market who discuss their latest research. In this episode, Shruti talks with Radhika Jain about her job market paper, “Private Hospital Behavior Under Government Insurance: Evidence from Reimbursement Changes in India.” They discuss the Bhamashah Health Insurance Program in Rajasthan, how hospitals are reimbursed for healthcare expenses, the effects of competition on the healthcare market and much more. Jain is the Asia Health Policy Postdoctoral Fellow for 2019-2022 at the Walter H. Shorenstein Asia-Pacific Research Center at Stanford University. Her research focuses on healthcare markets, the effectiveness of public health policy and gender disparities in health.

SHRUTI RAJAGOPALAN: Hi, Radhika. Welcome to the show.

RADHIKA JAIN: Hi, thank you so much for having me. I’m really excited about this.

India’s Public Sector Health System

RAJAGOPALAN: Me too. I’m really excited about this. Today I wanted to talk to you about this huge study that you’ve run in Rajasthan. This is about the Bhamashah Health Insurance Program that the Rajasthan government has announced. There’s a huge number of beneficiaries, I believe about 45 million.

You find something quite interesting: that even though the program is set up to be free for the beneficiaries, private hospitals engage in quite interesting behavior in that they engage in a lot of coding manipulation so that they can get larger revenues from this kind of public subsidy that’s offered. Aside from doing this with the state, they also wrongfully charge patients out-of-pocket expenses when, in fact, the program is supposed to be free. Can you tell us a little bit more about what is going on with this program? And what are the institutional circumstances in which hospitals are behaving in this way?

JAIN: Sure. Actually, let me give you a little bit of the background on this. For decades, the Indian government has largely relied on public provision directly through this network of public district hospitals and health centers. Research has increasingly shown that the quality in the public sector is often quite low, and as a result, households are overwhelmingly turning to the private sector. I think about 60% of non-childbirth hospitalizations are not in the public sector, even though the public sector is much cheaper. This is exposing households to enormous financial risk.

Because reforming the public sector has been difficult, one response is that maybe you can contract the private sector and, to some extent, leverage the market forces to provide better-quality care. In fact, that’s actually a quite interesting study by Jishnu Das and coauthors that shows that, indeed, the same doctor actually provides better care in his private clinic than he does when he’s working at the public hospital.

There’s this whole thing that the market builds in these incentives to respond better to patients. Working with the private sector has its own challenges, and you have to understand the private sector’s incentives, which are quite different from public sector actors. In that context—so what India has done since about 2008 when it rolled up the RSBY program, it’s been since then scaling up state and central health insurance programs. These programs target low-income households. They entitle households to entirely free care, at least on paper.

The big thing is that in addition to—households can get care in the public sector but also at private hospitals. Then a major design aspect of these programs is that they use what we call the prospective payment systems. Essentially, it’s a flat payment. Rather than using a fee-for-service where the hospital says, “Okay, I did X, Y and Z, and this is what I’m going to charge you,” you have this flat payment for a given service. For example, you’ll have a basic vaginal delivery. The government says, “I will pay you 3,500 rupees for doing that, regardless of what specific medicines you provided or how long the patient stayed.”

The idea of this is, it’s trying to contain program costs so that the hospitals can’t rack up thousands and thousands of rupees of charges for a patient. They can increase the efficiency in the program. What should be fairly obvious when you think about it is that, in this context, these prices are extremely important. How do you set the appropriate price for a basic vaginal delivery? The way you set these prices can have substantial implications for program outcomes. There’s lots of literature on this in the U.S., but this hasn’t really been studied in the Indian context.

Changing Hospital Reimbursement Rates

JAIN: That’s what I try to do. The problem is that the implications for reimbursement rates for hospital behavior are difficult to look at in the cross-section. What I do in my paper is take advantage of this policy reform that happened in December 2017, where the government decided to revise its reimbursement rates for different services. This change was going to take place overnight. The changes to the reimbursement rates are of different magnitudes across different services.

The economic intuition of what’s going on here is essentially the reimbursement rates that the hospital receives changes overnight. It’s unlikely that the cost of doing a service change overnight. Essentially this reform is this profitability shock to the hospital, and because the size of the changes is different across different services, you have these—the relative profitability of services is also changing. Because this is all happening at a discrete point in time, I can look at hospital outcomes before and after and compare them across services that have got different levels of change.

The idea was that maybe there was some suspicion that these reimbursement rates were too low, so hospitals were not participating. There was some anecdotal evidence that maybe the hospitals were charging patients. The hope was that when you increase the reimbursement rates, you’re paying hospitals more of doing something; you should see an increase in volumes. Hospitals should be more willing to provide these services, and that you might also see some benefits to patients, whether that’s reductions in out-of-pocket charges, or maybe hospitals are providing better-quality care because they’re getting paid more per visit and this kind of thing.

That’s what the expected effects are, but this assumes a competitive market where—but if you think about it, if a hospital has monopoly power, it may not actually face the incentives to pass on the benefits that it’s given, pass on these higher payments to patients. Then the other thing to consider is that there are also incentives for strategic responses, specifically coding manipulation.

If the hospital makes a lot more from a service—that now they got a 50% rate increase—it can just say that they did service without actually providing it. It can manipulate its coding. I use the program’s administrative insurance claims data, and I also conducted surveys with patients soon after their hospital visits to look at all of this. The key results that I find are, starting with upcoding, you see these really, really dramatic changes in what hospitals are filing claims for.

Using the survey data, I show that a large share of this is not actual changes in service provision, but coding manipulation. What I see is that hospitals change the way they manipulate coding to increase their revenues. They upcode more on where they’re going to make more money.

The second thing I look at is the supply response. And in fact, I find that there is—setting aside these coding manipulation responses—there is a real increase in supply. The elasticity that I estimate is about 0.7, which basically means that for 1% increase in the reimbursement rate, you get about 0.7% increase in the supply of a service.

Then the third thing I look at is out-of-pocket payments. I provide systematic evidence that hospitals are charging patients out of pocket. Almost half of patients are paying something. Basically, what I find is that when reimbursement rates are increased, hospitals charge patients less, so the government pays the hospital more. It helps cover the hospitals’ costs and hospitals pass that on to patients in the form of lower out-of-pocket charges. I also find that a substantial share of hospitals do not pass this on and do not lower their out-of-pocket charges. Specifically, this is in markets that are less competitive.

Again, standard theories of competition suggest that where the hospital has a lot of competition, it faces an incentive to reduce what it charges patients, reduce its prices, patient-facing prices. In places where the hospital has monopoly power, it doesn’t have to do that. We can simply capture what the government is paying. Patients are going to have to go there, regardless of what the hospital charges. That’s a long summary of the paper.

How Are Reimbursement Rates Determined?

RAJAGOPALAN: It seems like there is a basic design problem with the way this program has been set up. And this design problem, to my mind, is at three different levels. One is when we are thinking about reimbursement rates—or how much should each service be priced in a particular hospital of a particular size in a particular location— the government does not have enough data on how to come upon these rates. Some of this kind of predatory behavior by hospitals is simply because of a poor design setup at the pricing level.

The second design flaw is there seems to be some genuine lack of state capacity with the government of Rajasthan when it comes to monitoring private hospitals. No matter how high you set the rates, there might always be some incentive to upcode. Some of this needs to be figured out through a monitoring mechanism or some system. It doesn’t seem like the Rajasthan government has the capability to do this.

The third is, of course, market competition, which is, as you find, the more competition there is, the greater the concentration of healthcare providers. More of the benefits of the rate increases get passed on to the customers or the patients. I want to take each one at a time. The first one: How are these rates determined? I know that you study the point where the rates increased, but is there a method to the madness?

JAIN: Yes. Actually, I think I have—one of the absolute key takeaways from this research is that these prices are key, extremely important policy lever for shaping hospital incentives. I think sometimes I talk about—imagine if you were trying to pay the private sector to provide a plate of food. You would think very hard about what the price should be for that big plate of food. You would think about whether that price should be the same in a small town in Barmer versus in Jaipur City.

It seems like sometimes when we’re thinking about health, we don’t think about prices in this way, when we think about contracting the private sector to how these prices are set. In places like the U.S., for example, they have systematic cost data from the hospital, historical cost data. They also have patient risk data, and just a much more robust data setup. Costs are based on historical costs and the cost of other hospitals in the area; they’re also adjusted for patient risk.

In India, there simply isn’t this kind of data. We don’t have data on hospital costs. We don’t have data on patient risk, and the government is in the tough position of having to estimate things to come up with these reimbursement rates. The process for setting these reimbursement rates is, I have to say, somewhat opaque. But typically, it’s a panel of government doctors who get to sit around and think about roughly what we think this costs in the public sector.

There are typically also private sector hospital representatives that give their inputs. I’m not sure of whether there’s anything more systematic than that, whether there are any rules or things like that. That’s how the rates are set, and we are in an environment where there isn’t systematic data. We know we’re in a second-best environment, and I think we just start with that. What I hope is that some of this research starts getting people to think, “Okay, now we’ve rolled out these insurance programs. We really need to think about how we design these critical design elements.”

Hospital Monitoring and Competition

RAJAGOPALAN: Absolutely. This also comes to the second point, which is, yes, there’s a lack of state capacity and there’s lack of government monitoring, but a lot of the effectiveness of the monitoring depends on what the government thinks should be done. It needs a baseline for what is optimal behavior. If the services are systematically mispriced or underpriced, then it’s never going to quite get that baseline of how a hospital would operate were the incentives set up right—that right now there are a few bad apples who are misbehaving or something like that.

That’s where I think the second part of your study is super interesting, where you find that the greater the concentration of hospitals, the more the benefits of the public subsidy actually get passed on to the patients. What is a way of disciplining the hospitals from a decentralized market mechanism? How should we think about reducing barriers to entry and encouraging more competition and having more healthcare providers and so on, so forth?

JAIN: There were lots of points there. I will say, for a second, that—what you started out saying—there’s absolutely an interaction between monitoring and pricing. One of the implications of the paper is that if you just crack down on hospitals, but you don’t pay attention to the prices that you’ve set up if the prices are too low, cracking down on hospitals will potentially have adverse effects and essentially squeeze hospitals out of the program.

These things work in tandem, and you need to think about this very carefully before just going out there and saying, “These hospitals are crooks; we need to crack down on them.” Monitoring both public and private hospitals, and social programs in India in general, is very limited. Particularly, there’s not very much done to check with the intended beneficiary what their experience with the program was, and I would love to see more of that.

Turning to think about competition, I think the effects of competition can be complicated. I think what I’m trying to say in my paper is when you work with the public sector, at least the way it’s currently set up in India, it’s immune to competition. So you don’t have to think about that. When you work with the private sector, you’re contracting with them. Part of the reason you’re doing that is because you’re trying to leverage market forces. Well, then you have to think about those market forces. Essentially, what I’m showing is that market structure affects the way the private sector performs. As you’ve pointed out, affects the extent to which public subsidies reach patients.

In terms of—I will say the literature on competition is a bit complicated. I don’t know if your audience is into the economics of this, but there’s a really great Journal of Economic Literature review of this by Martin Gaynor and Kate Ho. One of the things that they find is that when hospitals are only competing on quality and not on price, having more competition can lead to sizable increases in quality. This is in the context of where there’s administered prices in, say, the U.K., or some aspects of the U.S. Medicare programs.

There, somebody else has been doing the price, doing the payments, and hospitals are simply competing on quality. That typically leads to improvements in quality. When hospitals are competing on both price and quality, which is what de facto has ended up happening in the Indian context, then theoretically and empirically it’s a lot less clear what will happen to quality when there’s more competition. Essentially, it’s complicated, but it depends on how demand responds to price versus quality.

I don’t get into all of that. It’s just worth saying that the implications of competition are not straightforward. I’m not sure that there’s not enough competition from the private sector in the healthcare sector in India. There seems to be a very robust private sector. I suspect that there are places, there are some areas where there’s not enough demand to sustain more than a couple of hospitals, like in more remote populations and areas and things like that.

I personally think that, actually, this is where competition from the government sector could play an important role, where the government can ensure that even in remote areas, there is at least one district hospital and that it’s a good, well-functioning district hospital. I don’t know for sure, but I think that that might help discipline the private sector. In fact, insurance can be used to do this. Insurance payments follow patients.

Right now, at least in Rajasthan when I was studying this program, insurance payments that go to the government hospitals don’t get used by the hospital in any way. Basically, the hospital has no incentive to attract patients and do what it can. If you did link some of the insurance payments to benefits for staff at the government hospitals, then you’re starting to create some incentive for them to also try to attract patients. It both might improve the performance in the public sector and act as this disciplining force. Now the private sector is competing with a relatively high-quality public sector and a relatively cheap one.

Improving Quality

RAJAGOPALAN: I think your point about quality is very well taken, because that is the hidden aspect in all of this that gets missed. We also don’t have very good post-care surveys that are done, say, a month later, three months later, six months later—risk of infection. There are just so many aspects to quality of healthcare, which are hidden at the time when the healthcare service is used and come up much later. Another interesting finding in your paper is that increasing these reimbursements rates also increases overall service volumes.

This, I think, is really incredible because India as a country, and Indians as a subgroup, are one of the most underserved groups in terms of healthcare provided per capita and healthcare expenditure per capita. Because of this historic trend in India of not having invested enough in healthcare, that infrastructure is still weak. If you look at doctor-patient ratios, nurse-patient ratios, the number of beds available per 100,000—on each and every one of these, India, for a country of its income level, or as part of the emerging economy group with, let’s say, Brazil and Russia and so on—it’s just nowhere at the same level as these other countries.

This increase in service volumes, I think, might be one of the most interesting findings in the paper. What is driving this? Is it just simply greater availability of funding, in a way that out-of-pocket expenses were never something private healthcare providers could have relied on, certainly to increase the scale of their operations? That’s really part of it. You want certainty before you expand.

JAIN: Essentially, all I’m showing, on some level, is that it’s a standard upward-sloping supply curve. You pay more and hospitals are willing to do more. I say that like it’s straightforward, and yet it seems not to be in policymaking. I think, to your point, what do we do to increase supply? Honestly, I think we just don’t spend enough on health. Now, the future of the country is very poor and simply cannot pay for the healthcare that they do need. Many of them forgo care.

The idea of this largely free public sector was supposed to rectify that, but we don’t spend very much on the public sector. It’s understaffed and under-resourced and is not able to provide the quantity or quality of care that we need. Then you have these insurance programs that are trying to rectify that and bring in the private sector. Even there, we constrain the supply by either paying too little, as I suggest in my paper, or there are all sorts of other barriers that are erected to utilization that effectively ended up rationing the supply of care in these programs.

I suspect that if everybody got the hospitalization that they need under these insurance programs with no overutilization, these programs would go bankrupt. They’re not funded to actually provide the levels of care that people need. What’s happening is that low-quality private care is filling the gaps because it can provide these services at the prices that people can afford to pay. I think the long story is we just don’t pay enough to get more care. Now, I understand that there are other policy priorities that the government has to think about, but personally, I think we need to spend more on healthcare.

The Gender Gap in Government Health Insurance

RAJAGOPALAN: I would agree, especially given what’s happened in the last couple of years. It’s exposed the enormous weaknesses in the healthcare system and frankly, the missing healthcare system, in some sense. It just doesn’t exist yet.

I want to talk to you about another paper of yours, which is also based on this particular program. This is your work with Pascaline Dupas. This is about the beneficiaries more specifically. In this case, you’re looking at why there is such a wide gap between men and women who are utilizing this government healthcare insurance program. The way the program is set up is, it benefits both groups equally. It’s supposed to provide this care for free. You still find that there’s this massive gap between men utilizing these services and women utilizing these services.

What is going on in Rajasthan? Is it a question of the missing women? I would be surprised if that can explain such a wide gap. Is it some other institutional reason? Is the problem on the supply side, that the hospitals don’t cater to women as much? Or what’s going on?

JAIN: As you pointed out, the gaps are very large. And we do a bunch of work to try to show that this is not due to differences in the population, sex ratio or illness incidence. It’s not that women are just sick less frequently. What we actually try to show in the paper is that there are all sorts of things that are barriers to female utilization. One of the factors is gender bias within the household in how many resources households are willing to allocate to males versus females.

Essentially, as I discussed earlier, we find that these programs are not free. Hospitals are charging patients out of households out of pocket. That’s one cost of using the program, but there are also other costs of using the program, like travel costs, forgone wages from the time off from work, things like that. Essentially, households face some cost to seeking care for an individual. If the household is biased and is less willing to spend on the female than it is on the male, that even with a program that heavily subsidizes healthcare, you end up seeing these gender disparities.

As you pointed out just a few minutes ago, the subsidy applies to everybody equally. You might think that the cost is similar for everybody, but households’ willingness to bear that cost differs by the gender of the patient. I think the two things to take away from that paper: One is that it goes against the standard intuition that if women aren’t using it, we should subsidize it more. What we show is that, well, subsidizing it is great. It does increase the number of women, but it doesn’t reduce disparities because men also benefit at least as much and sometimes more.

Then the corollary of that is if you want to address gender disparities, you’re going to need to do things that specifically target women, whether that means offsetting their costs in various ways.

RAJAGOPALAN: Some of the findings are really startling. Some of the largest disparities are in young women, young girls under 10 years, and elderly women. If you remove childbirth and care related to childbirth and health services related to childbirth, the gap is just staggering for almost every other kind of illness and care related to that. It just seems like women go to the hospital largely for childbirth. That’s a massive part of how the subsidy is used, and for everything else, they are just completely marginalized.

JAIN: I was shocked that for children under 10, I think females comprise about 33% of all hospital visits. Sometimes people would say, “Well, there are travel barriers. Women have a harder time reaching the hospital.” These are all legitimate and real things, but actually what’s interesting about seeing the disparity amongst children is that those barriers don’t apply there. These are children. Whether it’s a girl child or a boy child, it will be taken to the hospital by a parent. I think that’s one of the places where it’s proof that really this is about households making different decisions for boys and girls.

RAJAGOPALAN: I just want to quickly go into one other paper. This is your paper which is looking at the effects of the COVID lockdown on non-COVID care. The thing I want to particularly talk about in the context of the rest of our discussion is, of course, we’ve all lived through the pandemic. We know that the lockdown affected non-pandemic-related healthcare—cancer treatments, dialysis patients, these are some of the most acutely affected. In this particular study, again with Pascaline Dupas, you talk about how there’s, once again, a bigger impact on women.

Are the reasons the same as in the case of Rajasthan, which is a question of household resource allocation? Or is that exacerbated by something else, which is very lockdown and pandemic-specific?

JAIN: I think the underlying mechanisms are the same. It is exactly as you said, that I think it boils down to households are willing to allocate more to the health of males than they are to females. This manifests in routine care over time, but during the pandemic, essentially households’ budget constraints are tightened. Essentially, it puts households in more of a bind, and maybe they have to make more of a decision about what they’re going to spend their scarce resources on and this kind of thing.

I think what the pandemic did was just push households harder on the same underlying mechanisms of bias and resource allocation.

Pandemic Activities

RAJAGOPALAN: Speaking of the pandemic, what have you been up to during the pandemic? Other than writing lots of papers, because you have a lot of research that’s coming out.

JAIN: Actually, this was one of the things I did. I was trying so much to try and find some way to be helpful with what was going on in India. I did actually do that a lot with work on various aspects of either research or pandemic response efforts.

Besides that, I’ve been reading a fair amount, some odd books. I’m really into animals and plants. I read this book, “Other Minds,” that’s about octopuses and octopus intelligence and linking that to the philosophy of mind and consciousness. I refused to eat octopus for the last decade, since I first read about this stuff.

I read another book called “The Hidden Life of Trees,” which is all about this guy who gets really excited about old-growth forests. I have this theory that there’s a lot more going on in trees, in terms of communication and all of that, that we just don’t quite understand because it’s happening on a much slower time scale.

RAJAGOPALAN: Have you read “H Is for Hawk”?

JAIN: No, what is this?

RAJAGOPALAN: This is a book by Helen Macdonald. It’s really hard to describe. It’s about Macdonald. It’s like a memoir, and it’s about her story of how she spent a year training this northern goshawk after her father’s death. It’s really quite a lovely book. I think you’ll enjoy it.

JAIN: That sounds super cool. Thank you.

RAJAGOPALAN: Now for the most important question: During the pandemic, what have you been binge-watching?

JAIN: I watched “Tiger King,” which everybody watched. Then there were a bunch of—I’m even forgetting whether these were during the pandemic. There were a few series that came out early on. “Watchmen” was one.

Then the other was this crazy sci-fi “Lovecraft Country” also, which I thought was very weird initially, but then it totally sucked me in. Again, both of these were—I’d never seen anything quite like this, where you have sci-fi with prominent African American actors and tapping into actual history of the Tulsa massacre and things like that in the U.S. That was pretty interesting.

RAJAGOPALAN: Very cool. Thank you for these recommendations. It was such a pleasure to speak with you, and I’m looking forward to reading more of your research.

JAIN: Thank you so much. This was really lovely. It was nice chatting with you.