andrew shulman, health care for the uninsured, health insurance, health services research, left on longwood, Medicaid, National Bureau of Economic Research health insurance, Oregon health insurance, Oregon Medicaid
It turns out that having health insurance not only improves access to health care but also helps people to feel better and avoid financial catastrophe. Although these findings are entirely intuitive, well-controlled data comparing similar groups of patients with and without insurance have not been available until now. A study published today by the National Bureau of Economic Research is the first randomized control trial on the effect of inclusion in Medicaid on individuals’ assessment of physical and economic well-being.
The circumstances of the study, led by Katherine Baicker of the Harvard School of Public Health and Amy Finkelstein of MIT, are remarkable. In 2008, Oregon wanted to include more uninsured individuals in Medicaid but only had sufficient funding for 10,000 new enrollees. When almost 90,000 people applied, Oregon decided to select the lucky 10,000 by lottery. This Medicaid roll of the dice offered a dream opportunity for health services researchers and economists to design a randomized controlled trial (RCT) of the effect of insurance on a variety of health and financial outcomes. The RCT is the gold-standard test of interventions in medicine, ranging from new medicines to new policies. In order to meet the standard, the groups being compared have to be suitably alike in all regards with the exception of the intervention being tested. Since it would be unethical to kick people out of insurance programs to serve as the negative (no insurance) control group, a true RCT of health insurance has been hard to come by.
One year into data collection, the effect of inclusion in Medicaid on health care utilization and self-reported assessment of health are impressive. The insured were 35% more likely to see a doctor, 15% more likely to take prescription drugs, and 30% more likely to be hospitalized. They were 55% more likely to be able to identify a particular physician as their doctor. Those with insurance were 25% more likely to say that their health was excellent or good and 40% less likely to say that their health had worsened in the previous year.
The financial effect of insurance was also dramatic. People with insurance were 40% less likely to have to borrow money or not pay other bills in order to pay medical bills. The Medicaid group had a 25% decrease in having unpaid medical bills sent to a collection agency. The increased utilization did have a short-term cost; medical expenditures were 25% more for the insured.
We can keep an eye on the Oregon experience for some very interesting follow-up data on objective health measures in those with and without insurance. For example, the insured were 20% more likely to have their cholesterol levels checked. Did this translate into a higher percentage of individuals on cholesterol lowering therapy, lower measured cholesterol levels, or a reduction in heart attacks in those with insurance? We’ll see. The timeframe of this accidental RCT may be too short to answer these really interesting questions. Oregon obtained funding for the 80,000 lottery losers and expanded coverage to all who applied in 2010.
With Congress battling over the future of Medicaid and states and medical institutions facing dramatic cuts in Medicaid reimbursement, the timing of the new study could not be more spot-on. This unique data set should make it more difficult for the health care privateers to argue that public health insurance does not benefit those in need. We can expect the Oregon experiment to strengthen the view that in the long run, providing basic health care to the most vulnerable benefits us all.