Friday, August 31, 2012

Effects of Health Insurance: Randomized Evidence from Oregon

How might providing health insurance affect people along various dimensions, like how much health care they consume, their financial well-being, and their actual health? As health care economists have long recognized, this question is a lot tougher to answer than one might at first think. The basic analytical problem is that you can't just compare averages for those who have health insurance and those who don't, because these groups are different in fundamental ways. For example, those with private sector health insurance in the U.S. tend to get it through their employers, so they tend to be people of prime working ages who hold jobs, or those who get government health insurance for the elderly (Medicare) or the poor (Medicaid). It's easy to imagine cases of people who have a hard time holding a job because they have poor health, and thus don't have employer-provided health insurance. For these people, their poor health leads to a lack of health insurance, but wasn't primarily caused by their lack of health insurance. When the variables are interrelated like this, it's hard to sort out cause and effect.

From a social science research perspective, the ideal experiment would be to take a large group of people and to divide them randomly, giving health to one group but not the other. Then study the results. However, in the real world, such randomized experiments are quite rare. The one classic example is the Rand Health Insurance Experiment (HIE) conducted in the 1970s: for an overview written in 2010 with some applications to the health care debate, see here

"The HIE was a large-scale, randomized experiment conducted between 1971 and 1982. For the study, RAND recruited 2,750 families encompassing more than 7,700 individuals, all of whom were under the age of 65. They were chosen from six sites across the United States to provide a regional and urban/rural balance. Participants were randomly assigned to one of five types of health insurance plans created specifically for the experiment. There were four basic types of fee-for-service plans: One type offered free care; the other three types involved varying levels of cost sharing — 25 percent, 50 percent, or 95 percent coinsurance (the percentage of medical charges that the consumer must pay). The fifth type of health insurance plan was a nonprofit, HMO-style group cooperative. Those assigned to the HMO received their care free of charge. For poorer families in plans that involved cost sharing, the amount of cost sharing was income-adjusted to one of three levels: 5, 10, or 15 percent of income. Out-of-pocket spending was capped at these percentages of income or at $1,000 annually (roughly $3,000 annually if adjusted from 1977 to 2005 levels), whichever was lower. ... Families participated in the experiment for 3–5 years."

The basic lessons of the Rand experiment, which has been the gold standard for research on this question over the last 30 years, is that cost-sharing substantially reduced the quantity of health care spending by 20-30%. Further this reduction in the quantity of health care spending had no effect on the quality of health care services received and no overall effect on health status.

Now, 30 years later, there's finally another study on the effects of health insurance built on a randomized design. The first round of results from the study are reported in "The Oregon Health Insurance Experiment: Evidence from the First Year," co-authored by an all-star lineup of health care economists: Amy Finkelstein, Sarah Taubman, Bill Wright, Mira Bernstein, Jonathan Gruber, Joseph P. Newhouse, Heidi Allen, Katherine Baicker and the Oregon Health Study Group. It appears in the
 August 2012 issue of the Quarterly Journal of Economics, which is not freely available on-line, although many in academia will have access through library subscriptions.

The story begins when Oregon, in 2008, decided to offer health insurance coverage for low-income adults who would not usually have been eligible for Medicaid. However, Oregon only had the funds to provide this insurance to 10,000 people, so the state decided to choose the 10,000 people by lottery.  The health care economists heard about this plan, and recognized a research opportunity. They began to gather financial and health data about all of those eligible for the lottery, the 90,000 people who entered the lottery, and the 10,000 who were awarded coverage.  Here are some findings:

"About one year after enrollment, we find that those selected by the lottery have substantial and statistically significantly higher health care utilization, lower out-of-pocket medical expenditures and medical debt, and better self-reported health than the control group that was not given the opportunity to apply for Medicaid. Being selected through the lottery is associated with a 25 percentage point increase in the probability of having insurance during our study period. ... [W]e find that insurance coverage is associated with a 2.1 percentage point (30%) increase in the probability of having a
hospital admission, an 8.8 percentage point (15%) increase in the probability of taking any prescription drugs, and a 21 percentage point (35%) increase in the probability of having an outpatient visit. We are unable to reject the null of no change in emergency room utilization, although the confidence intervals do not allow us to rule out substantial effects in either direction.
In addition, insurance is associated with 0.3 standard deviation increase in reported compliance with recommended preventive care such as mammograms and cholesterol monitoring. Insurance also results in decreased exposure to medical liabilities and out-of-pocket medical expenses, including a 6.4 percentage point (25%) decline in the probability of having an unpaid medical bill sent to a collections agency and a 20 percentage point (35%) decline in having any out-of-pocket medical expenditures. ... Finally, we find that insurance is associated with improvements across the board in measures of self-reported physical and mental health, averaging 0.2 standard deviation improvement."

 Under the Patient Protection and Affordable Care Act signed into law by President Obama in March 2010, the U.S. is moving toward a health care system in which millions of people who lacked health insurance coverage will now receive it. Drawing implications from the Oregon study for the national health care reform should be done with considerable caution. Still, some likely lessons are possible. 

1) One sometimes hears optimistic claims about how, if people have health insurance, they will get preventive and other care sooner, and so they will avoid more costly episodes of care and we will end up saving money. This outcome is highly unlikely. If lots more people have health insurance, they will consume more health care spending overall.  

2) The cost of the Oregon health insurance coverage was about $3,000 per person--adequate for basic health insurance, although less than half of what is spent on behalf of the average American spends for health care in a given year. The health care reform legislation of 2010 is projected to provide health insurance to an additional 28 million people (leaving about 23 million still without health insurance). At the fairly modest cost of $3,000 per person, the expansion of coverage itself would cost $84 billion per year.

3) Although health insurance will improve people's well-being and financial satisfaction, the extent to which it improves actual health is not yet clear. As the Finkelstein team reports: "Whether there are also improvements in objective, physical health is more difficult to determine with the data we now have available. More data on physical health, including biometric measures such as blood pressure and blood sugar, will be available from the in-person interviews and health exams that we conducted about six months after the time frame in this article."

Evidence from theOregon health insurance experiment will be accumulating over the next few years. Stay tuned!