In the series of reports, called “Dying for Coverage,” Families USA purports to show how many people are killed by a lack of health insurance in each state. For example, they claim 6 people die every day in Florida because they are uninsured. Seven die every day in Texas, 8 in California, and 25 in New York.
How is Families USA able to tally up all this carnage with such pinpoint precision? As it turns out, these claims are based on a 15-year cascade of studies – each repeating the errors and misinterpreting or mischaracterizing the findings of the previous one and ultimately relying on data that is 37 years old.
Let’s start at the beginning. A paper by Peter Franks et al. published in Journal of the American Medical Association in 1993 estimated that there were 1.25 deaths among those without health insurance for every death among those with health insurance. To arrive at that number, the authors compared the cumulative mortality by 1987 of people who were either privately insured or uninsured at the time of a survey taken from 1971 to 1975.
Note: If the subjects studied were normal, those who were uninsured in the original survey would have remained so for only a short period of time, and over the 19 year interval members of the whole group would have experienced shifts of insurance status (from insured to uninsured and back again) several times. Being uninsured, after all, is like being unemployed. Lots of people experience the condition, but usually for short periods of time. Yet if this is true, it’s hard to draw any meaningful inference from the study about the effects of insurance on mortality.
For the study’s inference to be meaningful, one almost has to assume that the uninsured stayed uninsured for the full 19 years! But if that were so, the group would in no way be representative of the U.S. population as a whole.
To make matters worse, the authors claim their findings support “universal health insurance to reduce both financial barriers to care and the risk of premature mortality.” Yet the study itself provided no information about the effects of government coverage on mortality because it excluded “adults with Medicaid, Veterans Administration insurance, or Medicare.”
A 1994 paper by Sorlie et al. in the Archives of Internal Medicine did provide information on the effects of government coverage on mortality. It reported that people on Medicaid and Medicare had higher mortality than either the uninsured or those with private insurance.
Continuing the saga, the Institute of Medicine (IOM) uncritically used the Franks result to claim that 18,000 deaths a year in the U.S. are attributable to a lack of health insurance. (See the 2002 report Care Without Coverage: Too Little, Too Late.) The justification for this widely-publicized estimate appears in Appendix D, where the IOM adopts the Franks’ 1.25 ratio without explanation or discussion.
In January 2008, the Urban Institute published a report by Stan Dorn updating the 2002 IOM mortality estimates by combining IOM methodology with the latest Census Bureau estimates of the uninsured.
Families USA applied the Urban Institute updates and the IOM method to even more imprecisely estimated state data to produce its Dying for Coverage series. Page 2 of the Urban Institute report explains the methodology used to create the IOM estimates, the Urban Institute estimates, and the Families USA estimates. Consider 25 to 34 year olds. Deaths are allocated among the insured and uninsured for this group in the following way:
- Take the percent of insured (79%) and the percent uninsured (21%) in this age group from the Current Population Survey.
- Let X be the rate at which the insured in this age group die.
- Assume that the rate of uninsured death is 1.25 times X (from Franks).
- Add the death rate of the insured and the death rate of the uninsured together to get total deaths. Using U.S. vital statistics data this yields 40,548 = (0.79*X) + (0.21*(1.25*X)).
- Do simple algebra to get 40,548 = 1.05*X.
- Divide by 1.05 to get the number of deaths assumed to be attributable to the insured population, 38,617.
- Claim that the 1,931 difference between total deaths and the deaths attributable to the insured population are “excess deaths” due to a lack of health insurance.
Note that in the seven steps listed above, there is no point at which anyone from Families USA actually examines a medical record. There is no interview with any doctor, any patient or any family of a deceased patient. There is only algebraic mumbo jumbo in support of an unsupportable claim.
Note also that in the age group 25 to 34, more than half of all deaths are due to accidents, suicides and assaults. How, you may wonder, does insurance status affect these deaths? In their publication, Families USA says the uninsured have a higher mortality rate because they (a) are less likely to have a usual source of care outside of the emergency room, (b) often go without screening and preventive care and (c) often delay or forgo needed care. However, they do not explain how any of these factors cause death by accident, suicide or assault.