Study on ER visits by people with Medicaid challenges theory they would go less

A new study published in "Science" found that low-income people who have Medicaid insurance go to the emergency room for care 40 percent more than their counterpoints without any insurance. Katherine Baicker of Harvard School of Public Health, a co-author of the study, joins Hari Sreenivasan

Read the Full Transcript


    A study published in the journal "Science" this week found that low-income people with Medicaid insurance go to the emergency room about 40 percent more often than their uninsured counterparts do. The study raised questions about the long-held theory that extending Medicaid coverage to more Americans would cut down on emergency room use.

    It involved thousands of poor people in the Portland, Oregon, area, and it was part of the Oregon health insurance experiment.

    Hari Sreenivasan takes the story from there.


    Joining me to discuss this now report is Katherine Baicker. She is the professor of health economics in the Department of Health Policy and Management at the Harvard School of Public Health and was one of the authors of the report.

    So, Ms. Baicker, Judy went over one of the headlines. Let's dig a little deeper into this. Forty percent, that seems like a very large spike in E.R. usage. What does that actually translate to for the people in the study? How much more did they use it?

    KATHERINE BAICKER, Harvard School of Public Health: Well, on average people in the control group went to the emergency department about once during the 18 months that we looked at, and people who got Medicaid went 1.4 times.

    So that's about a 40 percent increase. It's helpful to think about why people were covered by Medicaid vs. not in our study. Oregon had a lottery for a limited number of spots in its Medicaid program, and they drew names from a waiting list by chance. So we were able to compare people who didn't gain access to the program through the lottery to people who did. And that gave us a really good control group to figure out what happened when people gained access to Medicaid.


    So is there a possibility that this is a surge? Is 18 months long enough to suss out long-term patterns?


    It's a good question.

    Longer-term effects might certainly differ from you what see over an 18-month period, although we didn't see an increase in the beginning that petered out. It looked fairly steady over the 18 months we examined. But certainly longer-run effects might differ. And you also might expect to see something different when insuring many more people at the same time.

    Our study had about 10,000 newly insured people. That's a lot of people, but it's small relative to the population of Oregon or even relative to the uninsured population in Oregon.


    And speaking of the population in Oregon, how representative is it of the rest of the country? Some of the critics of the study have said, well, maybe the poor in Oregon won't model the same way that the poor are in the rest of the country.


    Of course, one always wants to be careful in generalizing from any one state to others or from any one study to the nation overall.

    That said, Oregon's Medicaid program looks similar to programs in other states in many ways. It covers the same types of services, including not just the emergency department, but doctor's office visits, prescription drugs, hospitalizations. It has no co-pays for enrollees. The population in Oregon looks similar to that of the U.S., with one important exception, that there are fewer minority residents in Oregon than there are in many other states.

    So Oregon might not be very representative of racial or ethnic differences that would you expect to see elsewhere.


    Were there any subgroups in the demographics that were using the emergency room more often, perhaps the elderly or one gender over the other?



    We studied people who were 19 to 64. That's the group that was eligible for this program, where older people would already be eligible for Medicare. And we saw increases in lots of different types of utilization. We saw people going to the emergency department more, for example, for conditions that might most readily be treatable in other settings, in primary care settings, or emergencies that might have been averted by earlier primary care.

    We also saw a bigger increase for men in the probability of going to the emergency department at all.


    So what accounts for that?

    I mean, if people are going to the emergency room for care that they could have gotten at the doctor's office elsewhere, is it because they don't know the difference? Or is there a spike in the number of primary care physicians that might be referring people to the emergency room?


    There are lots of open questions about the underlying reasons that we see the patterns of care that we do. But we know a couple of things.

    First, people also went to the doctor more and they also went to the hospital more and used more prescription drugs. And all of that is consistent with the idea that insurance makes health care more affordable for patients. And when health care is more affordable, when the price to them is lower, they use more of a lot of different kinds of services.

    It's possible that patients sitting at home with a certain set of symptoms, like a sprained ankle that might be broken or pain that's been around for a day or so, but might get better, choose not to go to the emergency department when they worry that they will be faced with a large bill, whereas those who are insured decide to go rather than wait.

    Similarly, if they call their doctor and their doctor says, I'm not sure about those symptoms, you better go to the emergency department, the uninsured may be reluctant to do so because they fear the large bills that they would incur.


    And so you have published two previous sets of findings from this study. Anything more that is coming out?


    We certainly have much more data that we're eager to explore.

    Our previous studies were able to look at utilization across a range of settings, like the doctor's office visits and hospitalizations I mentioned, but also to assess physical health and mental health. And we saw a big reduction in depression. We didn't detect any other changes in physical health measures like high blood pressure or cholesterol.

    That depression result is one that we really want to dig into with more data on modes of care and treatment for that condition. We also found big improvements in financial security and well-being that seem very important for the well-being of the individuals who are less likely to have a bill sent to collection, and important for the health care system, where those bills sent to collection were never paid to providers.


    All right, Katherine Baicker from the Harvard School of Public Health, thanks so much for joining us.


    Thank you for having me.