Medicaid, America’s Worst Health-Care Program
Progressives can’t admit it.


Avik Roy

In July 2010, at National Review Online’s Critical Condition blog, I wrote about a University of Virginia study, published in Annals of Surgery, finding that surgical patients on Medicaid endured a 97 percent higher likelihood of in-hospital death than patients with private insurance, and a 13 percent greater chance of death than those with no insurance at all. I noted several other clinical studies that showed similar results. Little did I know that a national firestorm would ensue.

Not everyone shared my concern about Medicaid’s poor health outcomes. A number of scholars on the left expended more effort trying to debunk the UVA study, and the dozens of others that support it, than addressing Medicaid’s many flaws. Others, such as The New Republic’s Jonathan Cohn, conceded that “for certain populations and particularly in certain states, [Medicaid is] unambiguously inferior to private insurance,” all the while describing my “novel and brash” criticisms (and those of others) as a “conservative assault on Medicaid.”

Progressives recognize the real problem: If Medicaid is fundamentally flawed, it follows that Obamacare is, too. Obamacare aims to add 17 million more Americans to Medicaid’s ledgers. For this reason, some are more eager to paper over Medicaid’s flaws than to address them.

This has led to a surreal circumstance in which anything that suggests that Medicaid offers even a minimal benefit is trumpeted in the press, whereas the overwhelming literature describing Medicaid’s flaws is ignored.

Hence the avalanche of attention that has been showered upon a recent Harvard study, published in the New England Journal of Medicine, arguing that Medicaid indeed does save lives compared with having no insurance at all. The debate is over, declared the New York Times. “The new study should lay that canard [of Medicaid’s poor outcomes] to rest.”

But a look under the hood of the Harvard study reveals a different story. The authors compared three states that expanded their Medicaid programs — Maine, Arizona, and New York — with neighboring states that did not — New Hampshire, Nevada and New Mexico, and Pennsylvania. The Medicaid expansion was associated with increased mortality in Maine, and with decreased mortality in Arizona and New York. That’s hardly a definitive outcome.

Indeed, demographic differences between New York and Pennsylvania could explain the entirety of the “benefit” that the authors ascribed to New York’s Medicaid program. Yet the authors’ conclusion — that Medicaid saves lives — hinges entirely on the comparison of New York with Pennsylvania. Without it, the authors would have shown no difference in outcomes between those with Medicaid and the uninsured, because the results in Maine and Arizona would have canceled each other out. I don’t expect that major American newspapers would have trumpeted such a result.

The study lacked rigor in other ways, too. The Harvard economists looked only at county-level data about mortality and Medicaid; they had to make assumptions about which patients had enrolled in the program, and when. The extensive clinical research showing Medicaid’s poor outcomes, such as the UVA study, has reviewed millions of individual patient records to learn what happened to specific patients with specific forms of health insurance.

Medicaid’s problems are not that hard to understand. Medicaid is the largest line item in most states’ budgets, and it continues to grow at a faster pace than tax revenues. In theory, the program is jointly run by the states and by the federal government, so states have the ability to rein in costs. But in reality, Washington bureaucrats veto most reforms that states seek to make in their Medicaid plans.