Does Health Coverage Make People Healthier?
A new study provides a compelling answer to the vexing question underlying the health care debate.
The passage of last year's Affordable Care Act promised to bring health coverage to all Americans, including the country's poorest and least likely to be covered under the current system: Among the act's many provisions was a guarantee of Medicaid coverage for uninsured low-income adults.
As with many aspects of the ACA, the effect of insuring millions of poor Americans represents a leap into the unknown. How much more health care would they use, at what cost, and to what benefit? Would Medicaid actually produce healthier, more able-bodied Americans? Or simply increase health care use with no discernable payoff? Would it even increase medical access at all?
A landmark study released this morning on a major health policy experiment in Oregon provides some hint at what the answer to these questions might be if the ACA becomes a reality. A consortium of researchers from MIT, Harvard, and the state of Oregon has studied the impact of randomly assigning Medicaid insurance to poor Oregonians in 2008 as part of an expansion of the state's health coverage. They found that Medicaid's impact on health, happiness, and general well-being is enormous, and delivered at relatively low cost: Low-income Oregonians whose names were selected by lottery to apply for Medicaid availed themselves of more treatment and preventive care than those who remained excluded from government health insurance. After a year with insurance, the Medicaid lottery winners were happier, healthier, and under less financial strain.
While the financial and health effects of medical insurance may seem self-evident, the costs and benefits of insurance can actually go in either direction. For example, one oft-cited rationale for making insurance universal is to discourage people from falling back on costly (to society) emergency room visits for problems that could be dealt with in a doctor's office. The insured may also seek treatment earlier than the uninsured, saving the state from covering costly hospital stays that could have been prevented by pre-emptive care. (But then again, they might not.) At the other extreme, the financial costs of expanding health coverage are potentially enormous if it turns out that the newly insured rush out to get all manner of unneeded procedures and medical attention that they no longer have to pay for.
The health benefits are also hard to gauge. Given the limitations of Medicaid, some have argued that the quality of care that comes with it is scarcely better than nothing. So why expect any health improvements at all from Medicaid enrollment? Insurance is also the source of what economists call "moral hazard," where those who are protected against the consequences of their actions take greater risks than they otherwise would. Insurance, even limited insurance, might encourage people to drive faster, eat more French fries, and keep smoking cigarettes, knowing that they at least won't be on the hook for the financial consequences of their self-destructive behavior.
One reason that the health care reform debate is so dominated by ideology rather than fact is that it is extremely hard to measure the impact of providing insurance coverage. You can't just compare people who buy insurance with those who don't because the two groups differ in all sorts of unobservable, unmeasurable ways. The kind of person who buys insurance is probably less of a risk taker to begin with, for example, and also more concerned about his health. And since most Americans get health insurance through their employers, it's only the unemployed and people with lousy jobs who are shut out of insurance—i.e., the types of people who are typically in worse health to begin with. Bad health itself may result in job loss, financial calamity, and loss of insurance, reversing the chain of causation.
Oregon provided a window of opportunity for identifying the impact of insurance when it opened a waiting list for a limited number of new spots in the Oregon Health Plan Standard that provided Medicaid insurance to poor adults who didn't quite meet the usual conditions for Medicaid (for example, poor but able-bodied adults). Nearly 90,000 Oregonians applied for the 10,000 slots, and the state picked the lucky Medicaid recipients by lottery. This gave the researchers a group of insurance applicants who were identical in every way except whether they got lucky in the OHP draw. If the lottery winners ended up spending more time at doctors' offices or told researchers they were happier and healthier, it wouldn't be because of some pre-existing condition or circumstance. Their insurance status was quite literally the luck of the draw.
The state conducted eight lottery drawings during 2008. The winners received the right to apply for Medicaid, and about 30 percent of these ultimately were enrolled in the program. (Many failed to submit their applications, and about half of those who applied were found to be ineligible, mostly because their incomes were above the poverty line.) The researchers followed both lottery winners and losers over the following year by matching their sample of lottery applicants to hospital and other administration data. They also conducted surveys on the health, general well-being, and financial circumstance of both groups.
The study turns up a number of stark differences between the "treatment" (those who received Medicaid) and "control" (those who didn't) groups, some more surprising than others. Those selected by the lottery were about 7.5 percent more likely to be admitted to the hospital for treatment. But when you take into account the fact that most lottery winners never actually got insurance (and some lottery losers found other insurance), the numbers suggest that insurance increased the chances of hospital admission by 30 percent.
Relative to the control group, the newly insured used more of just about every kind of health care service the researchers considered. This included big jumps in preventive care, ranging from a 60 percent increase in mammograms to a more than 50 percent increase in having a regular doctor and clinic for primary care needs.
The only type of care with no statistically discernible increase—or decrease—was emergency room use. Given that usage goes up or is at best flat in the one area you might expect to see a decline, adding to Oregon's Medicaid rolls naturally added to state health care expenses. A rough calculation based on the increased service use is around $800 per year per enrollee. (The full cost to the state of Oregon is much higher, since the government is now on the hook for all the services that the new enrollees had been using previously, many of which had presumably been covered by charity care and other nongovernment service providers that care for the uninsured.)
Given the added expense, did the Medicaid expansion prove to be cost-effective? That is, did the treatment group actually have better health outcomes? It's too early to really see how much the increase in early interventions pay off in terms of heart disease, cancer, or life expectancy. But to get a sense of whether there were health improvements even within the first year, the researchers surveyed the lottery participants on the state of their health in the summer of 2009, a year or so after Medicaid enrollment. At least based on their survey responses, the health impact of medical access was already kicking in. Insurance increased the chances that a respondent reported himself in "good," "very good," or "excellent" health (as opposed to "poor" or "fair") by 25 percent, and reduced the number of "sick days"—days lost to work and/or recreation because of poor health—by about 15 percent.
The survey also asked whether respondents were "happy, pretty happy, or not too happy." The insured were 32 percent more likely to answer "happy" or "pretty happy." This is an enormous effect. To get a sense of how large, it's helpful to use estimates from studies that link income to happiness as a benchmark. You'd have to double a participant's income to see an effect like this, making Medicaid ridiculously good value for money. Better health is surely part of the reason for the more upbeat view of life. But insurance also relieves the financial anxieties that come with being uninsured in America—the fear that, quite literally, one wrong step could spell financial calamity. The study found that Medicaid recipients were 40 percent less likely to skimp on other financial obligations as a result of medical bills, and presumably even those that never had health problems worried less about such possibilities.
There are limits to what you can extrapolate from one, single-year study of 10,000 Medicaid recipients in Oregon to health care reform more generally. If millions of poor Americans were enrolled in Medicaid tomorrow, it might overwhelm the system's capacity. And while the program might have longer-term effects not seen in a one-year study, as preventive care starts to have an impact, it's also possible that the benefits of Medicaid may lessen with time. For example, once people adjust to their newly insured circumstances, their sense of contentment and well-being may quickly dissipate. We'll have more information on these long-term effects as researchers survey participants in the Oregon Medicaid lottery in future years. They're also collecting data on physiological measures like cholesterol levels and blood pressure to measure more objectively participants' well-being.
For now, though, the study represents the best evidence we've got. And based on its findings, Medicaid seems like a very cheap way of making Americans better off, and the goals of the Affordable Care Act well worth fighting to put into practice.
Ray Fisman is the Lambert Family professor of social enterprise and director of the Social Enterprise Program at the Columbia Business School. He is the co-author, with Tim Sullivan, of The Org: The Underlying Logic of the Office. Follow him on Twitter.
Photograph of doctor checking patient's blood pressure by Darrin Klimek.