During the recent Democrat Party Convention in Denver, a local nonprofit ran an advertisement in USA Today and many of Colorado’s major newspapers urging elected officials to solve the nation’s health care crisis. The ad, purchased by the liberal Colorado Health Foundation, claimed that “Every year, 18,000 Americans die because they don’t have health insurance.”
That statistic of deaths due to lack of health insurance has been frequently cited in health care debates since the Institute of Medicine (IOM) reported it in 2002. A recent update from the liberal Urban Institute showed that the number of deaths due to lack of health insurance had subsequently increased to 22,000.
Yet the statistic has come under increasing fire by critics who questions its accuracy. Linda Gorman, a senior fellow with the conservative Independence Institute in Denver, said, “It’s not a robust result. You don’t make good public policy on such results.”
Chris Power-Bain, senior communications officer at the Colorado Health Foundation, defends the statistic. “The Institute of Medicine is a credible source of information. Further, there is a growing body of evidence showing a strong relationship between health outcomes and insurance.”
One concern cited by Gorman is the wide degree of uncertainty surrounding the statistic. “Care Without Coverage,” the 2002 IOM report, based the statistic on an earlier study in the Journal of the American Medical Association (JAMA). That study found that lacking health insurance increased the risk of mortality by 25%.
One can follow the IOM’s methodology in determining death due to uninsurance by reading the Urban Institute report. In short, the IOM used the statistic from the JAMA study to calculate how many people would have died if everyone in the U.S. was insured. It then subtracted that number from the number of people who did die to determine the number of people who died due to lack of insurance.
The problem lies in the statistic in the original JAMA study. Specifically, the data in that study could not rule out the possibility that lacking insurance has no effect on mortality. Nor could it rule out that the effect might be larger, up to 55%. Thus, it is possible that the number of deaths annually due to uninsurance could be as low as zero or as high as 47,000.
Wilhelmine Miller, an associate research professor in the Department of Health Policy at George Washington University and the director of IOM report, defends the report. “We made some judgments about what studies to use,” she said. “The ones we used were methodologically strong.”
Helen Levy, research assistant professor at the University of Michigan’s Institute for Social Research, disagrees. “The basis for the (IOM) number is ridiculous. Lots of other things affect those deaths.”
Health outcomes are affected by many factors, including income, education level, smoking, and diet. Those factors also affect insurance status, making it difficult to discern the exact effect lack of insurance has on mortality.
An additional problem is that some people are uninsured for brief periods while others are uninsured for longer period of time, something which studies linking mortality to insurance often don’t account for. Richard Kronick, professor of family and preventive medicine at the University of California at San Diego, said, “That likely understates the problem. If you could control for insurance status over a longer period, the real effect might be bigger.”
Jack Hadley, a principal research associate at the Urban Institute, concurs. Studies he has conducted that follow the insurance status of the older Americans over time show that uninsurance “tends to have a somewhat large effect on mortality,” he said. “More recent studies show that the relation between insurance and morality is pretty robust.” However, those studies do not form the basis for the 18,000 figure.
There is also considerable debate over whether increasing the number of people with insurance is the best way to improve health outcomes. Michael Cannon, director of health policy studies at the libertarian Cato Institute, said, “There is no evidence that health insurance will deliver better outcomes than spending money on other things. Money spent on community health centers, nutrition programs or improving education may have a stronger impact.”
Hadley, however, points to problems with that approach, noting that some approaches are easier than others. “You can give everyone insurance tomorrow. You can’t give everyone a college education that quickly.”
Of course, giving everyone government-funded insurance would prove extremely costly. Before we, as a society, decide to pay that cost, we must know with some certainty the cost of not providing insurance. And right now, the statistic on deaths from uninsurance is anything but certain.