If you’ve been following the health care debate, you’ve probably heard would-be reformers express outrage that half of all bankruptcies in the US are caused by medical bills. But is it true?
Not really. Or rather, only if you accept a number of dubious premises.
Well, what about the sister statistic that that health care costs cause a bankruptcy every 30 seconds?
Nope. Simple math proves that incorrect; the Bankruptcy Data Project at Harvard University shows that from February 2008 through January 2009 there were 1,114,811 bankruptcies of all types, personal and business, or about one every 28 seconds. This is fewer bankruptcies than earlier in the decade but the conclusion remains true even for those years.
In fact, both of these claims are based on an article published in the journal Health Affairs in 2005. It was authored David Himmelstein, Steffie Woolhandler, Elizabeth Warren, and Deborah Thorne. Yesterday, with this post already finished, I found out that they have just published a new article, upping the percentage of medical bankruptcies to 62.1 percent and adding some new methodological problems to the old ones. So today I will take on the 2005 piece and on Monday I will dissect their newest effort.
To reach their conclusions, Himmelstein et al. surveyed 1,771 people who declared bankruptcy in 2001 and asked about the reasons for their bankruptcy. Those citing medical bills as a factor were 28.3 percent of the total. However, to classify even these as medical bankruptcies would ignore the fact that there were other contributors.
But then the authors go on to add to this percentage anyone who meets a series of criteria. For instance, they considered a medical bankruptcy one in which there was $1,000 or more of medical debt, regardless of the source or size of other debts. In other words, if the person has $50,000 of credit card debt and $1,000 owed to a doctor, this counts as a medically precipitated bankruptcy.
These also include debtors who had lost two weeks’ income due to medical problems, but once again, there is no consideration of other debts and factors. Only a small fraction of people would be pushed into bankruptcy by this alone, the vast majority have other large debts or other circumstances that render them vulnerable to bankruptcy if they lose income for any reason.
Finally, Himmelstein et al. say that when costs were incurred for “addiction, or uncontrolled gambling, or birth, or the death of a family member,” it counts as a medical bankruptcy. Addiction and gambling do not belong in this category, the treatment perhaps is a matter for the medical system (or not) but the money spent on these habits certainly isn’t. Why birth and death are singled out is also mysterious; they ought to be treated the same as any condition involving medical care.
In addition, the study is uncontrolled, we know nothing about the families that don’t go bankrupt but may have the same amount of medical debt, and the article tells us the average medical debt of the people who filed for bankruptcy but not the median, so we do not know if this number is skewed by just a few people with very high expenses.
Statistics have failed to show a correlation between national trends in medical costs and bankruptcy statistics or between medically related absence from work and bankruptcy. Both a Legal Aid society study and one from the Department of Justice put medical bills at about 12 percent of what people filing for bankruptcy owed. David Dranove and Michael L. Millenson concluded in a critique of Himmelstein et al. that taking the 28.3 percent who cited illness or injury as a reason for their bankruptcy and the 60 percent of this group who blamed medical bills specifically, the percentage of people for whom medical expenses were a cause of bankruptcy (and not necessarily the only or the largest) is 17 percent.
So much for the myth that medical bankruptcies are overwhelming millions of Americans every year.
Along with the critique by Dranove and Millenson here.
