For years, progressives have claimed that they are the party of science. “Our government has forced what I believe is a false choice between sound science and moral values. . . . I believe the two are not inconsistent,” said President Obama in 2009. Science is the reason we need to suppress trillions of dollars of economic activity around the world, because Al Gore — and science! — tell us that if we don’t, “our planet will experience cataclysmic warming by the end of the century.” Gore has even taken to calling conservative climate-policy skeptics “climate deniers,” a clear reference to Holocaust deniers. But there is at least one area of public policy where the Left has abandoned its rhetorical allegiance to science: health care.
For years, studies have shown that patients on Medicaid — America’s government-run insurance program for the poor — do no better, and sometimes do worse, than those with no insurance at all. The largest such study, from the University of Virginia, evaluated 893,658 major surgical operations from 2003 to 2007 and found that surgical patients on Medicaid were 97 percent more likely to die before leaving the hospital than those with private insurance. Medicaid patients were 13 percent more likely to die than those with no insurance at all. The study adjusted for income, age, geography, prior health status, and other factors.
Many other studies have shown the same thing. There are others suggesting that Medicaid isn’t worse than being uninsured, but that it isn’t better, either. The main problem is that Medicaid pays primary-care doctors an average of 52 cents for every dollar a private insurer pays. This leads many doctors — and also specialists — to stop taking patients on Medicaid, making it hard for poor enrollees to get routine check-ups and needed care.
In 2011, Jonathan Cohn, writing in The New Republic, described those of us who wrote about these studies as launching a “conservative assault on Medicaid.” Liberal health economists contorted themselves into pretzels explaining why the data — the science — must be wrong.
There was a very good reason why progressive commentators needed to convince themselves — and us — that Medicaid made people healthier. Half of Obamacare’s expansion of government-subsidized health-insurance coverage was routed through the Medicaid program. If Medicaid didn’t improve health outcomes, then neither would Obamacare. When then-senator Joe Lieberman (I., Conn.) said that he wouldn’t support Obamacare if it expanded Medicare, the Washington Post’s Ezra Klein wrote, “Lieberman . . . seems willing to cause the deaths of hundreds of thousands of people in order to settle an old electoral score.”
The Oregon experiment
One legitimate criticism of the inconvenient Medicaid studies is that — like nearly all such policy research — they look back retrospectively on a set of data in order to gain insights into how Medicaid is performing. Hindsight, as we know, can be 20/20. In theory, the most objective form of research is a controlled experiment, in which you randomly assign your subjects to be treated differently and watch what happens in the future, based upon prespecified measurements.
Four years ago, a group of leading health economists, including Obamacare architect Jonathan Gruber, sought to rectify this problem. It turned out that the State of Oregon was holding a lottery, in which, within a certain income bracket, it would randomly allow some people to enroll in Medicaid while leaving others uninsured. It was the perfect opportunity to try to conduct a study on Medicaid’s health outcomes.
In 2011, the economists released their initial findings. While it was too early to measure Medicaid’s effects on objective health measures such as blood pressure or cholesterol, patients told the researchers that they felt better about their health.
This rather modest result led to a chorus of jubilation from liberal journalists. “Amazing Fact! Science Proves Health Insurance Works,” read a headline from Ezra Klein. Wrote Matthew Yglesias, “A new rigorous study from Oregon confirms that Medicaid does, indeed, save lives.” (The study did not, in fact, detect any change in mortality.) Oregon’s result “suggests that having health insurance substantially improves health,” wrote David Leonhardt of the New York Times. One of the authors of the study, Amy Finkelstein of MIT, cheered, “What we found in a nutshell is that having Medicaid makes a big difference in people’s lives.”
There were problems with the first-year Oregon data. In their research paper, the investigators noted that two-thirds of the improvement in patients’ self-reported health took place “about 1 month after [Medicaid] coverage was approved” but before the patients had seen a single doctor or consumed any health-care services. This strongly suggested that the “benefit” that patients were reporting was the insurance version of a placebo effect. But this subtlety didn’t make the front pages.
When the following July came around and it was time to publish the two-year results of the study, the Oregon investigators were strangely silent. The election came and went. The state-by-state debate on expanding Medicaid came and went. Finally, on May 1, 2013 — ten months late — The New England Journal of Medicine published the second-year findings. Did Medicaid save lives? No. It “generated no significant improvement in measured physical health outcomes,” including death, diabetes, high cholesterol, and high blood pressure.
What’s almost as striking as this non-result is how few Oregonians felt the need to sign up for this allegedly “life-saving” program. The authors report that, of the 35,169 individuals who “won” the lottery to enroll in Medicaid, only 60 percent actually bothered to fill out the application. In the end, only half of those who applied ended up enrolling. Remember that this is a program on which we will be spending $7.4 trillion over the next ten years, a program that Obamacare throws 11 to 17 million more Americans into, because of the hundreds of thousands of lives that Medicaid will supposedly save.
The Medicaid spin cycle
Immediately, progressive bloggers went into overdrive to explain these results away. “The sample size was too small,” they said, even though new medicines for diabetes, high cholesterol, and high blood pressure routinely show significantly improved health outcomes in much smaller trials. “Two years isn’t long enough to show a significant benefit,” they insisted, even though new drugs that failed to show any benefit in two years would be summarily rejected by the FDA and abandoned by their sponsors.
The Medicaid cohort reported that they felt better about their health and their financial security as a result of enrolling in the program, and were less depressed. We can presume that the 40 percent of Medicaid “winners” who didn’t bother to fill out the application felt differently; they, however, were not surveyed.
Nonetheless, lefties seized on this qualified bit of good news. “This is an astounding finding . . . a huge improvement in mental health,” said Gruber. To which Ben Domenech responded, “I wonder whether we’d be better off replacing the [Medicaid] expansion with a program which hands out $500 in cold hard cash and a free puppy.”
Kevin Drum of Mother Jones, in a wince-inducing post, insisted that it was “flatly not true” that Medicaid didn’t improve health; he cited positive numerical trends on some health measures that were just as likely the product of statistical noise.
Austin Frakt of Boston University is a passionate Medicaid advocate who, for years, has disputed studies showing poor Medicaid outcomes. “That insurance . . . improves health and reduces mortality risk is as close to an incontrovertible truth as one can find in social science,” Frakt averred in 2010. Frakt held up the Oregon study as the gold standard for health-policy research. Now that its results are out, he has gone in both directions: He simultaneously insists that the Oregon study was “far too small” and highlights any encouraging sign he can unearth, downplaying all the reasons the study was biased in Medicaid’s favor.
Paul Krugman, the Medicaid-denier-in-chief, recycled the lefty-blogger talking points and then issued this classic: “If health insurance is a good idea — and you are nuts if you let this study persuade you otherwise — Medicaid is cheaper than private insurance. So where is the downside?” Oh, I don’t know. Maybe $7.4 trillion?
After that gambit didn’t take, Krugman abruptly shifted gears, arguing that Medicaid’s health outcomes don’t matter. “Fire insurance is worthless!” he snarked. “After all, there’s no evidence that it prevents fires.”
The real significance of Oregon
Krugman is right. Fire insurance doesn’t prevent fires, and it isn’t meant to. The purpose of fire insurance is to protect the policyholder from the catastrophic financial loss that occurs when one’s home burns down. In every sector of the economy except health care that’s what insurance is for. We buy car insurance to protect ourselves from the financial cost of car accidents. Yet liberals expect health insurance to save lives and improve health, instead of expecting it merely to protect us from catastrophic medical bills. We could easily apply the lessons of car insurance and fire insurance to health insurance. But we don’t, because the Left has been persistently and ideologically opposed to treating health insurance this way.
Indiana, under then-governor Mitch Daniels, placed Medicaid patients on an inexpensive combination of high-deductible insurance and subsidized health-savings accounts. That combination is meant to protect beneficiaries from catastrophic medical bills while giving them control over their own health spending. The program enjoyed a 98 percent approval rating among its participants. But the Obama administration shut down the program, insisting that it be replaced by traditional Medicaid. Florida’s GOP-controlled house of representatives proposed replacing Obamacare’s Medicaid expansion with an entirely state-funded program for catastrophic insurance and health-savings accounts. It was blocked by Republicans in the state senate, who insisted on expanding Medicaid instead. (The house–senate stalemate meant that neither bill was adopted.)
Singapore, which has a universal system of catastrophic coverage and health-savings accounts, spends one-seventh of what we spend on health care, with comparable results. If we gradually replaced our $1.5 trillion–a–year health-care leviathan with this approach, we could wipe out our budget deficit and permanently solve our entitlement crisis.
Unfortunately, Obamacare goes in the other direction. It expands Medicaid and sets up subsidized insurance exchanges where many people will be forced to buy costly, comprehensive insurance products that they don’t need. A different kind of exchange, in which individuals shop for narrower, consumer-driven health plans, could form the core of a free-market health-care agenda.
This, indeed, is the bottom line from the Oregon study: that protecting people against bankruptcy from medical bills is a good thing, and that we have far less costly ways of doing so than expanding an already failed entitlement program. That finding will reverberate for years, and provide momentum to conservative efforts to offer a better solution.