Study of Oregon Medicaid Program Reveals No Significant Health Improvements

Published May 28, 2013

A landmark study of Medicaid outcomes in Oregon published in the New England Journal of Medicine shows the program to be ineffective at improving the health of enrollees.

Researchers found enrollment in Medicaid protected most recipients from financial disaster, though they did report spending more on medical expenses. There were no significant differences in health outcomes between those who remained uninsured and those on Medicaid.

“This randomized, controlled study showed that Medicaid coverage generated 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,” the study’s authors wrote.

Rare Research Opportunity

In 2008, Oregon expanded its Medicaid program with a drawing from a waiting list. Researchers used this rare opportunity to study the effects of the Medicaid program on randomly selected participants, with those not selected by the lottery used as a control group.

Two years later, the researchers gathered baseline health and other data from 6,387 adults who had been selected to be on Medicaid and 5,842 adults who were not. They tracked blood-pressure, cholesterol, and glycated hemoglobin levels, used to track diabetes, and screened for depression and some self-reported health status. Medical expenses, including catastrophic expenses, were also self-reported.

The study found Medicaid coverage did not improve health but “nearly eliminated catastrophic out-of-pocket medical expenditures.”

Lottery Winners Didn’t Sign Up

Study participants, both on Medicaid and not, had self-selected to be on Medicaid by signing up for a waiting list. They may differ as a group from people who were eligible but did not apply to be in the program, notes Katherine Baicker, professor of health economics at Harvard University and one of the study’s lead authors.

“The people who signed up for the lottery may indeed be different those who did not, such as being sicker or more interested in obtaining health care,” Baicker said. “This does not affect the estimate of the effect of insurance on those who won compared with those who did not, but it is an issue to consider when generalizing to the effects of coverage in a broader population.” 

Of the 35,169 individuals who “won” the Oregon Medicaid lottery, only 10,405 ultimately enrolled. Baicker maintains this is not a significant problem with the study.

“There is imperfect take-up of Medicaid [meaning only some people enroll in Medicaid when it is offered to them], and this is also what is forecast under the ACA expansions [not all newly eligible people will take up], so in that regard this lottery population may be similar to those covered under expansions,” Baicker said. “But it is hard to know for sure how those who signed up for the list will compare to those who take up newly expanded coverage more generally.”

‘Classic . . . Placebo Effect’

Avik Roy, a senior fellow at the Manhattan Institute, says there are several flaws in the way the Oregon study measured the one area where small improvements were noted: depression.

“In 2011 the authors noted that two-thirds of the improvement in patients’ ‘self-reported health’ took place ‘about one month after [Medicaid] coverage was approved’ but before ‘any increase in health care utilization,'” Roy said. “In other words, patients felt better once they knew they were on Medicaid, but before they had seen a doctor, undergone a test, or filled out a prescription. That is the classic definition of a placebo effect. And it’s hardly astounding to anyone with experience in clinical trial design.

“And remember, the placebo effect works on objective health outcomes too, like blood pressure, high cholesterol, and diabetes, where the Oregon study showed no meaningful difference. How [much] worse would those measures have been if the Medicaid enrollees hadn’t known they were on Medicaid?” Roy asked.

Roy also says it’s noteworthy that the results for the second year of the study came out 22 months after the first-year results, a significant delay.

“Why did it take so long for the second-year results to be published?” Roy asked.

When asked, Baicker did not comment on the timing of the report.

Hospitals Benefit From Medicaid

In terms of financial impact, Oregon Medicaid recipients were largely, but not completely, spared from catastrophe.

“We followed a commonly used definition: having out-of-pocket medical expenses that exceed 30 percent of household income,” Baicker said.

According to Baicker, this was true of 5.5 percent of the control group, and Medicaid reduced that to about 1 percent. But according to Ed Haislmaier, senior research fellow for health policy studies at the Heritage Foundation, the true financial beneficiary of Medicaid is the hospitals themselves.

“This is not about the patients,” said Haislmaier. “This is about the medical-industrial complex.”

Haislmaier cited an Urban Institute study which calculated how much hospitals stand to benefit from the Medicaid expansion, “even when lost revenue from private insurance is included in the calculation.” That study, “The Financial Benefit to Hospitals from State Expansion of Medicaid,” calculated a 2.59:1 ratio of new Medicaid income for lost private insurance income for hospitals under the program expansion.

Expensive Sugar Pill

Noting that combined federal and state Medicaid spending is about $450 billion per year, covering approximately 50 million people, for a total taxpayer cost per covered person of about $9,000 a year, Roy says given the costs of the program and the results of the study, if Medicaid were a treatment, he doesn’t think doctors would prescribe it.

“Zero,” he said.


Internet Resources:

Urban Institute: “The Financial Benefit to Hospitals from State Expansion of Medicaid,” March 2013.