Who’s using the emergency room?

About 10 percent of Americans are uninsured, and most of them turn to the emergency room for basic care—an expensive last resort that burdens the entire health-care system. If more people were insured, would ER visits decline?

Two studies from the past decade point in different directions. A 2006 expansion of health insurance in Massachusetts led to fewer ER visits among those who gained coverage. But two years later, a policy in Oregon that awarded Medicaid by lottery led to an increase in ER visits. “The Oregon results were very surprising,” says associate professor of economics Amanda Kowalski. She helped analyze the outcomes in Massachusetts and, curious about the discrepancy, looked into Oregon’s data.

She retested the original analysis and got the same results. Emergency room visits were more prevalent among those who had enrolled in Medicaid because they won the lottery but would not have enrolled otherwise. Kowalski termed this group “compliers.” She then expanded the scope of analysis to include “always-takers,” who found a way to get medical insurance even if they lost the lottery, and “never-takers,” who declined coverage even if they won the lottery. “I wanted to see whether the experimental intervention had different impacts on different people within the experiment,” she explains.

She found that always-takers who lost the lottery but found coverage used the ER more often than compliers, and that never-takers, had they gained coverage, would have visited less. “This suggests that people in other states who gain access to Medicaid might not increase ER utilization in the same way as Oregon compliers did,” says Kowalski. “They could go more, they could go less.” For policymakers interested in whether ER visits will swell or contract given broader access to Medicaid, the challenge is understanding which population will be recruited: always-takers, compliers, or never-takers.