The Blog

Bad Faith Meets Bad Science

3:16 PM, Apr 22, 2014 • By ALEX VUCKOVIC
Widget tooltip
Single Page Print Larger Text Smaller Text Alerts

The attempts of defenders of Obamacare to rouse the American people in favor of the doomed monstrosity have become more desperate and bizarre. The most recent example is taking place in Florida, where the sudden death of a young uninsured woman is being cited as an indictment of the Republican-controlled state legislature for refusing to approve the Medicaid expansion so generously being offered by the feds. If the woman in question had access to federally-mandated Medicaid, they argue, she would of course have gone in for preventative screening which would have revealed her cardiac abnormality and somehow saved her life. Once again, heartless Republicans are causing the death of innocents.


But wait, there’s more. A study is out which demonstrates that Republican legislatures across the country will kill more than 17,000 Americans annually with policies similar to Florida’s. It’s on the blog, where the article “Opting out of Medicaid Expansion: The Health and Financial Impacts,” explains it all. Real Harvard doctors, led by the estimable David Himmelstein and Steffie Woolhandler, crunch the numbers—and they could not possibly have an agenda as they look out for our welfare. Could they?

If the names sound familiar, there is a reason. The authors have made names for themselves over the last three decades as passionate proponents of a socialized, single payer health care system. It is nice of them to come to the aid of Obamacare given that it does not go nearly far enough in that direction for their tastes, but they are no doubt seeing the ACA as merely a steppingstone to a British NHS-system, which they would love to impose on the American body politic. In fact, a look back at their tactics demonstrates the power of cynicism coupled to appeals to authority, which has driven much of their so- called research. Their methods are fairly clear-cut: manipulate large databases in order to reach their preconceived conclusions, dress what comes out of the sausage machine up with dubious statistical tools, and, presto, they have demonstrated the superiority of socialized medicine.

Himmelstein perfected this method in a 2009 paper published in the American Journal of Public Health purporting to show that lack of health insurance killed almost 45,000 Americans annually. The paper was released early—and only to a sympathetic press corps!—and was instrumental in Democrats’ arguments leading to the passage of Obamacare in the Senate just before Scott Brown’s election would have prevented it. The paper is a masterpiece of manipulation and deserves to be analyzed in detail.

Himmelstein began with a statement of the problem experienced by the enlightened as they attempt to explain the workings of the world to the unwashed masses: “Despite widespread acknowledgement that enacting universal coverage would be life-saving, doing so remains politically thorny.” Not for long, though! He examined a database of 34,000 individuals followed for twelve years, and polled once during that time period for their insurance status. He then reduced the number of records to 9,000 through five retrospective steps. The patients’ insurance status was determined by self-report—almost 13,000 of the original cohort and more than 2,300 of the final group he analyzed reported not being covered by government insurance. He neglected to note that in previous surveys, 15 percent of patients who reported lack of insurance actually were covered by Medicaid, and another 10 percent were eligible for it but did not bother to sign up. This omission by itself renders the rest of his analysis meaningless, but let’s proceed to the painful end. He controlled for alcohol and tobacco use but not drug use or abuse when comparing the insured and uninsured. Given the high likelihood that a substance-abusing population would be less likely to have insurance and more likely to die early (the cumulative mortality associated with opiate dependency is more than 30 percent!), the deck was stacked. Even though he had access to death certificates of the 351 patients who died during the follow-up period, he did not compare causes of death between the insured and uninsured. Finally, after all was said and done, the death rate was 3.0 percent for the insured and 3.3 percent for the (nominally) uninsured. Put another way, 17.2 percent of the dead patients and 16.2 percent of the 8,653 living patients were uninsured. Multiply the 0.3 percent difference in death rates by the uninsured population at the time, and one gets 45,000 dead uninsured Americans!

Recent Blog Posts

The Weekly Standard Archives

Browse 20 Years of the Weekly Standard

Old covers