Wharton's Atul Gupta discusses health care outcomes under the Affordable Care Act.

The Affordable Care Act (ACA) was landmark legislation when it was enacted by the federal government in 2010. The changes introduced through the ACA were vast and significant, although some of them have been difficult to measure. The latest research by Wharton health care management professor Atul Gupta aims to correct that by quantifying some of the health outcomes of the bill. In a paper titled, “Estimating Effects of Public Insurance Expansion for Adults: Evidence from California Hospitals,” Gupta and colleagues Mark Duggan and Emilie Jackson, both of Stanford University, analyze data gathered from hospitals and emergency rooms in California from 2008 to 2015. He recently spoke with Knowledge at Wharton about what the researchers found.

An edited transcript of the conversation follows.

Knowledge at Wharton: Your paper looks at the effects of the Affordable Care Act through the lens of what’s been going on in California. Tell us about what you studied.

Atul Gupta: Obviously, the Affordable Care Act is this major legislation that’s been passed, and there has been a lot of work showing that the ACA has increased insurance coverage in the United States. But there has not been that much rigorous work that’s shown us the effects of this insurance coverage increase. We don’t know if people are utilizing more health care now. We don’t know if it has improved their health outcomes. And we don’t know if it’s improved efficiency in the system, which is one of my main research interests. That’s what we try to answer in this particular project.

We used this particular research design that’s known as regression discontinuity. Simply, we take people who are very close to a particular age threshold — let’s say, 21 — and we look at people on either side of the threshold. Without going into the details, people who are below 21 had access to Medicaid coverage even before the Affordable Care Act, whereas people who were just over 21 aged out of Medicaid. They lost Medicaid coverage just because of the way the Medicaid rules were set up. But with the Affordable Care Act, this false discontinuity was eliminated. This gives us a nice, neat little experiment where people who are over the age of 21 now suddenly have access to Medicaid because of the Affordable Care Act. This allows us answer this question of what happens when people suddenly get access to insurance coverage.

“Once people have choices, they tend to choose better quality hospitals.”

Knowledge at Wharton: Are these the people that are described as being covered by the Medicaid expansion?

Gupta: Yes, exactly.

Knowledge at Wharton:  What happened when they got this coverage?

Gupta: There’s a variety of things that we look at. But let me just focus on three particular aspects that I think are key. If you just focus on people who are in this age band of 21 to 23 that we looked at, the share of people in that group who are covered by Medicaid goes up by about 20 percentage points. So, it’s a very large increase. But what we also find is that the decrease in uninsurance, which was the main goal of the Affordable Care Act, is not of the same magnitude. The reason is that some of these people who are gaining Medicaid coverage were earlier covered by other forms of insurance coverage. Particularly in this case, we find that Medicaid coverage is going towards covering people who were originally covered by county indigent programs. Basically, county tax dollars are now being replaced by federal tax dollars. This is pretty large in terms of the magnitude. About 40% of the Medicaid expansion is going towards replacing county indigent insurance.

Knowledge at Wharton: It’s not people that didn’t have insurance. It’s more about shifting the onus onto the federal government as opposed to the county or local government.

Gupta: Right. This is known as crowding out of insurance coverage.

A second takeaway is that we see a fairly robust change in where people are getting their health care. We take this as a sign that once people gain insurance coverage, they’re more likely to choose private hospitals. Private hospitals tend to have higher-quality metrics. We take this as a sign that once people have choices, they tend to chose better quality hospitals. It’s not something that’s driven by distance. It’s not something that’s driven by whether these hospitals are part of a particular chain or something of that sort. It is correlated mainly with ownership and performance on different quality measures. That’s a good sign because that suggests that people are now going to move towards higher-quality hospitals.

The third thing that we tried to look at is whether this has resulted in improvement in health outcomes for patients. In many ways, that is the bottom line that we’re interested in. These people previously did not have access to health care. We’re hoping now that they do, it will improve their lives. On that front, we don’t find robust evidence of improvement in quality of health. Let me be more specific. We’re looking at mortality, for example. Or we’re looking at the share of hospitalizations that were potentially avoidable. On both of these metrics, we don’t find robust improvements. But we do find some suggestive improvements.

I think the story really over there is that we might only have short-term effects right now because we only have two years of data post the expansions. It’s possible that in the long term, some of these effects will become more visible in the data. Also, these effects are probably not large enough right now. For that reason, they’re pretty noisy.

Knowledge at Wharton: We can’t really figure out if it’s because of the expanded health care, or if there’s some other reason why we’re seeing this?

Gupta: Exactly. I think we’re fairly confident that this is the only large change that happened around this time. What we’re not sure of is whether the average effects are large enough that we can actually catch them in only two years worth of post-ACA data. We’re still continuing to collect data, and we’re hoping that as we get more years of data, we’ll be able to carefully test these hypotheses.

Knowledge at Wharton: Were these people covered by Medicaid when they were under 21 because they had medical conditions that made them eligible, or were they covered because of income? Or both?

“[When] you read the popular media, it’s almost taken as a given that expanding insurance coverage is good.”

Gupta: Disability was definitely one reason why some people were covered by Medicaid. But that’s not only for people below 21. On either side of that age threshold, if you were disabled, you would get Medicaid coverage. The main difference is between people who have kids and people who do not have kids. If you were older than 21 and childless, you couldn’t have Medicaid just because of income. What the ACA did was make all of these rules much more simple and logical. Now, regardless of your age, as long as you are below a certain income threshold, you’re eligible for Medicaid. Frankly, that makes a lot more sense. It’s just that it gives us a nice, neat little experiment that we can use to look at the effects of insurance coverage.

Knowledge at Wharton: The ACA continues to be a topic of discussion. What can policymakers and health care providers take away from this research?

Gupta: I think there are several implications of this. One, we always knew that the ACA is a massive redistribution exercise — redistributing from healthy towards the sick, redistributing from the rich to the poor. But what we are finding here is that it’s also a redistribution in terms of who is paying for these costs. We’re finding that local tax dollars are now being replaced by federal tax dollars. It will be interesting to see what’s happening with those local tax dollars. Are they going and doing something else? For example, you might find improvements in amenities or services such as education or other kinds of programs that local governments generally provide.

Two, I think we’re asking somewhat of a provocative question because when you read the popular media, it’s almost taken as a given that expanding insurance coverage is good. While that makes sense intuitively in the economics literature, it hasn’t really shown up in the data in very carefully designed studies. We’re not finding robust evidence in our study as well. So, I think it’s a little bit of a provocative question because we’re trying to ask, “Is it really the case that expanding insurance coverage for these young adults did anything to improve their quality of care and health outcomes?” Or was it the case that these people in general are pretty healthy, so when you expand insurance coverage for them, they tend to use it for things that don’t show up in mortality, avoidable care, those kind of quality metrics that we have?

I hope that this research spurs other people to look at this because we’ve found in the past that when Medicaid was expanded for kids, it was greatly beneficial. And there’s a lot of robust evidence that when you expand Medicaid for poor kids, it leads to improvement in education outcomes. It leads to improvements in labor outcomes over the long term as well. But we’re finding that for young adults, that’s not the case. Like I said, it’s still early days and you might find that over the next five years, things can change. But I think it’s reasonable to at least examine the evidence and see what kind of bang we’re getting for this buck.

The third thing is that when people talk about the value of the Medicaid program for consumers, we generally think of it only in terms of “Oh, people who were poor now have access to health care.” But the fact that we find people changing their providers once they have access to insurance coverage is a very powerful endorsement of the Medicaid program. That suggests that people know which providers are better quality. It’s only that they did not have access to those providers because they lacked insurance coverage. Even if we don’t find improvements in health outcomes, the fact that people are now able to choose better providers, better hospitals, better physicians — that’s a source of value. When people try to quantify the benefits of Medicaid, this is a source of value that’s not really talked about at all.

“I think it’s reasonable to at least examine the evidence and see what kind of bang we’re getting for this buck.”

In our follow-up work, we’re hoping to put a dollar value on this. How much do people value being able to go to a Stanford University hospital versus some local county hospital, for example?

Knowledge at Wharton: What else is next for this research?

Gupta: There are a couple of things. One is not specific to this particular project, but it’s a pet peeve of mine. We don’t really have great health outcomes that we can look at in the economics literature. In general, we focus on very, very rare outcomes like mortality or readmission. And that’s what we’ve done in this study as well. But these things are typically not very sensitive. For example, when you look at young people, they don’t die very often. They also don’t get readmitted very often. When you want to look at improvements in quality of care and how that shows up in health outcomes, one thing that we’re hoping to do now is move away from these super-rare outcomes like mortality and readmission, toward something that’s a bit more sensitive. For example, spending relative to their sickness level and their distribution. Then revisit these results using our new and, hopefully, improved metrics.

The second thing we want to do is move beyond just health outcomes. The value of insurance could be not just in terms of health, but in terms of happiness. It could be in terms of financial well-being of these beneficiaries. For example, now they might have more disposable income. Or they might have less variance in income, or less variance in wealth because they have insurance coverage. That should be factored in when you look at the benefits of the Affordable Care Act because that improves consumer utility. We have a follow-up project where we have some survey evidence. People were asked about happiness and well-being at a fairly disaggregated level. We want to understand whether people are now feeling happier and whether they have self-stated improvements in consumer well-being, because that could also be an important value of the Affordable Care Act.