Relax, take a deep breath and let the government worry about the bill. Photographer: George Frey/Bloomberg
Relax, take a deep breath and let the government worry about the bill. Photographer: George Frey/Bloomberg

Does giving people health insurance help control costs?

Conventional wisdom holds that it should, by diverting them from expensive emergency room use to less-expensive visits to doctors and nurse practitioners. This argument was very popular with advocates for health reform in 2009, and it remains a sort of folk wisdom among educated people; I’ve heard some version of this argument in virtually every discussion I’ve had about health care in the last decade.

To health-care economists, though, the question is more complicated. Health insurance does theoretically let you go to a primary care physician rather than relying on ER docs who are legally required to treat you. On the other hand, it also reduces the cost of going to the ER. And as the basic laws of economics tell us, when you reduce the price of something, people usually want to consume more of it.

Which of these effects is stronger? There’s only one way to find out: test it. And thanks to the state of Oregon, we finally have a good test.

Longtime readers of this blog will be familiar with the Oregon Medicaid study, one of the most exciting studies in decades of the effect health insurance has on our lives. For those who aren’t, here’s the executive summary: A few years back, Oregon found enough money in its budget to add 10,000 people to Medicaid. Because more than 10,000 people wanted to go on Medicaid, it used a lottery to hand out slots ... and thus gave researchers an excellent setup for a randomized controlled trial of the effect that health insurance had on peoples’ lives.

Last year, the study made big headlines when deep analysis of the data showed that people who had won the lottery had lower rates of financial hardship and scored lower on tests of depression than people who hadn’t won the lottery ... but didn’t show statistically significant improvement on the three objective clinical health markers the researchers had chosen: blood pressure, cholesterol and diabetes control. As I wrote at the time, this was a big, big deal. Only two large-scale random tests have ever been done on health insurance, and both have come back with the same surprising result: giving people Medicaid, or more generous health insurance, doesn’t seem to significantly improve clinical measures of good health.

The headlines eventually faded, but work on the Oregon data set has not. In the fall, the study’s researchers came out with another interesting result: Adding people to Medicaid made no difference in their incomes; they did not drop out of the labor force because they had someone to pay for their health care; nor did better access to medical treatment enable them to boost their incomes. And today they’re releasing another paper, this time on the effect that becoming eligible for Medicaid had on ER usage. “It’s especially important to have that kind of gold standard,” in cases like this, says Amy Finkelstein, an MIT professor on the research team. “Where it’s not just a question of what the magnitude of the effect is, but a priori we don’t know what the sign of the effect could be.” Which is to say: Does giving people Medicaid drive ER usage up, or down?

The answer, it turns out, is “up.” People who got access to Medicaid used doctors more than people who didn’t. But they also used the ER more. And contrary to the theory that Medicaid might divert people from ERs to primary care physicians, the big increases actually came from the somewhat-less urgent problems that we were hoping to divert from ERs:

Medicaid increases use for visits classified as “non-emergent,” “primary care treatable,” and “emergent, preventable.” We find no statistically significant change in use of visits classified as “emergent, not preventable.”

Medicaid increases outpatient emergency department visits (visits that did not result in a hospital admission). We find no statistically significant increase in emergency department visits that did result in a hospital admission.

It’s not surprising that this is where the increase comes from -- if you’re in a car wreck or your arm has been cut off, you go to the ER whether or not you have insurance. But if it’s a bad case of strep, you might decide to tough it out. What is surprising ... or at least interesting ... is that the visits increased.

As health-care researcher Jonathan Gruber told the Washington Post’s Sarah Kliff, this helps put to rest one of the fonder hopes of health-care expansion, that it would save money through preventative care and diverting people away from expensive emergency room visits.

"I would view it as part of a broader set of evidence that covering people with health insurance doesn't save money," says Jonathan Gruber, a health economist at the Massachusetts Institute of Technology, who has also studied Oregon's Medicaid expansion but is not affiliated with this study. "That was sometimes a misleading motivator for the Affordable Care Act. The law isn't designed to save money. It's designed to improve health, and that's going to cost money."

There’s a broader point to be made as well: Obamacare probably doesn’t solve the problems that most people were expecting it to solve.

When I talk to smart, educated, well-read people about Obamacare, I generally hear three major points advanced in its favor: first, that it means that people with pre-existing conditions will finally be able to buy insurance for the first time; second, that it prevents young people from free-riding by going without insurance and then sticking the rest of us with the bill for uncompensated emergency room care; and third, that it will allow poor people to go to a regular doctor rather than relying on expensive, unnecessary ER visits. All three rely on a folk conception of how the health-care system works that isn’t quite right.

Take pre-existing conditions. It’s not that this isn’t a problem -- it is. But it’s not nearly as large as so many people I meet seem to imagine. In fact, it doesn’t even seem to be as large as health-care wonks thought, prior to 2010. The health-care law created high-risk pools to take care of this problem, allowing this group of people to buy insurance at subsidized rates. It was projected to cover almost 400,000 people, but ended up covering about a quarter of that number -- and in order to get that many, administrators had to engage in aggressive promotion and relax the qualification requirements.

In part, that’s because we already passed a law to deal with pre-existing conditions, the 1996 Health Insurance Portability and Accountability Act. The law covers a lot of ground, but one of the things it does is ensure that you’ll be able to keep buying insurance if you already have it. Even before that, individual policies often had a feature called “guaranteed renewal,” which allowed you to keep renewing even if you got sick.

And in many cases the problem for people with pre-existing conditions was not necessarily that no one would sell them insurance, but that the insurance people would sell them cost more than they wanted to pay. That’s a very different problem from “can’t buy it at any price” -- but the latter is the problem that most people thought we were solving with Obamacare.

Or look at uncompensated hospital care, another area where Obamacare was supposed to make a big difference. Like pre-existing conditions, uncompensated care is a real problem. But it too, is much smaller than people imagine. The American Hospital Association estimates that uncompensated care cost $41 billion in 2011. This in a nation that spends more than $2 trillion a year on health care. Reducing this cost would make hospitals happy, but it is not going to noticeably lower health-care costs, or transform American health care.

And then there are the uninsured people who use the ER, from whom we hoped to glean fabulous health-care savings. Oregon suggests that insurance status isn’t the main reason that people are using the ER for less severe conditions. And if you think about your own life, is that such a surprise? I have resorted to the emergency room more than once for something that I theoretically could have gotten treated by a primary care physician, and insurance was never the driving factor; I went to the ER because I was traveling far from home, because it was outside of normal business hours and I didn’t have a regular doctor I could call, or because I tried to nurse along an infection that turned severe. The cure for this was not to change my insurance; it was the myriad urgent care options that are flowering across the U.S.

This is true of most of the middle-class people I know. Why did we assume that the poor were any different?

Obamacare mostly solved a quite different problem: the fact that health insurance and health care are expensive. It probably isn’t going to lower costs, or dramatically alter mortality rates, or turn obese diabetics into marathoners. On the other hand, it probably will improve the financial stability of a lot of low-income households, while raising costs (and taxes) on the more affluent. As Finkelstein said to me: For an economist, insurance is a financial product, health insurance as much as life or auto insurance. Auto insurance probably doesn’t improve your driving much, but it does protect your assets if you’re in an accident. It may not be what we expected Obamacare to do ... but it’s probably what we should have expected.