Will the Repeal of ObamaCare Kill 22,000 People?

I’m intrigued by folks who aren’t at all hesitant to express an opinion on just about anything; no matter how uninformed they are on the issue.  Paul Krugman (who I’ve talked about before) is one of those individuals.  Nobel Laureate in economics and syndicated columnist writing in the New York Times, Krugman recently wrote an interesting piece based on junk science entitled “Death by Ideology.”  Arguing that Mitt Romney’s pledge to repeal ObamaCare will result in the deaths of “thousands, and probably tens of thousands, of excess deaths of Americans each year,”  Krugman once again shows that his ideology trumps sound economic analysis. 

Is Krugman right?  Will the repeal of ObamaCare result in needless suffering and death?  How could Romney be so callous?  Where does Krugman get his information?  In 2002 the Institute of Medicine (IOM) published a report on the consequences of being uninsured.  In it they estimated that the uninsured have a 25% higher mortality rate than the insured and the result was 18,000 excess deaths annually.  The Urban Institute updated that report in 2006 and raised that number to 22,000. 

At the time of the original IOM study, there were only two empirical studies on the mortality consequences of being uninsured (Franks et al. in the 1993 Journal of the American Medical Association and Sorlie et al. in the 1994 Archives of Internal Medicine).  The two observational studies used similar methodology and not surprisingly came up with similar results, being uninsured was associated with a 25% higher mortality rate.  Notice I said “associated with.”  Using the methodology of the two reports, it is impossible to make causal inference.  In other words, inferences about the causal relationship between lack of insurance and mortality using the observational approach are dubious.  Even more questionable is the large statistical error associated with the estimates.  Both studies found the 95% confidence interval to range from no increase in mortality to a 50% increase, or between zero excess deaths to 36,000 excess deaths.  Instead of saying there are 18,000 excess deaths from uninsurance, IOM could have easily said going without insurance may have no impact on mortality or it could result in as many as 36,000 excess deaths. 

Richard Kronik writing in Health Services Research in 2009 (“Health insurance coverage and mortality revisited”) looked at the issue again using a much larger data base than either of the previous studies.  Kronik goes through a series of adjustments to the difference in mortality between the insured and the uninsured.  When the data is adjusted by age and sex, the mortality rate for the uninsured is 70%  higher. 

 Without going into the statistical details Kronik adjusted for other differences between the two groups, including race, education, and income.   With every adjustment, the difference in mortality rates narrowed.  When smoking status, BMI, self-reported health status, and physical activity were added, the difference vanished completely.  The 95% confidence interval after taking into consideration differences between the two groups was between minus 5% and plus 12%.  In other words, there is a good chance that the uninsured have a 5% lower mortality rate than the insured. 

These results do not mean that we should not worry about improving access to medical care by expanding insurance coverage.  But let’s stop the nonsense of calling Mitt Romney and Paul Ryan cruel and callous, wanting to kill people by taking away their insurance.  We’ve got too much to do to improve the efficiency of our health care system and expand access to those vulnerable population groups, the chronically ill and the elderly.  For my money, let’s fix Medicare first to make it sustainable for those of us nearing the age of reckoning.  I for one am not interested in a real world application of the carrousel scene in the 1976 movie “Logan’s Run.”  Look it up if you are not familiar with it.

Print Friendly, PDF & Email

Leave a Reply

Your email address will not be published.