Surprises in a Randomized Trial: The Oregon Health Study

A recent article appearing on the New York Times blog The Upshot describes some unexpected findings from a randomized controlled experiment that evaluated the costs and benefits of an expansion of public health insurance. In 2008, the state of Oregon began a limited expansion of Medicaid benefits for uninsured, low-income adults. By way of a lottery and waiting list, the state chose some low-income residents to apply for Medicaid while others remained uninsured. The Oregon Health Insurance Experiment made use of this random assignment to study the effects of Medicaid on health care use, health outcomes, financial strain, and well-being among low-income adults. For 2 years, the researchers used an in-person data-collection protocol to assess a wide variety of outcomes. Among the unexpected findings:

Patients who got insurance used the emergency room more often than their uninsured peers, undermining a common argument in favor of expanded insurance coverage.

People who got Medicaid also had a much easier time finding doctors, countering views held by the programs’ critics that Medicaid can be worse than no insurance.

According to the study authors’ conclusions published in the New England Journal of Medicine, “Medicaid coverage yielded no significant improvements in measured physical health outcomes in the first 2 years, but it did increase use of health care services, raise rates of diabetes detection and management, lower rates of depression, and reduce financial strain.” Other study analyses are published in Science, among other journals.