Thursday, July 30, 2009

Medical bankruptcy myth?

Diana Furchtgott-Roth says that the perception of medical bankruptcy as a widespread problem is largely a myth:
In testimony before a subcommittee of the House Judiciary Committee, Mrs. Edwards declared, "Medical debt is, of course, a symptom of larger problems in our health care system-and the solution to medical debt and medical bankruptcy is real health reform that results in affordable, reliable health coverage and affordable health care for all Americans."

Mrs. Edwards based her testimony on a study in the American Journal of Medicine conducted by Dr. David Himmelstein and other researchers from Harvard University and Ohio University. An unassuming reader might conclude that medical debts are the major cause of personal bankruptcy in America, because the study finds that 62% of bankruptcies in 2007 were "medical."
However, it seems the study was a bit flawed:
Dr. Himmelstein's study contradicts the economics literature on personal bankruptcies. Most reputable studies are based on the Survey of Consumer Finances, published by the Federal Reserve, which lists different types of consumer debt. Medical debt rose slightly from 5.5% of all debt in 2001 to 5.8% of all debt in 2007, according to the Fed.

A study by the Department of Justice examined more than 5,000 bankruptcy cases between 2000 and 2002. It found that 54% of bankruptcies involve no medical debt, and more than 90% have medical debt of less than $5,000. Even among the minority of bankruptcies that report medical debt, only a few have enough to cause personal bankruptcy.

Dr. Himmelstein gets different results because he uses a smaller sample and a different methodology than other studies. He started with a random sample of 5,251 bankruptcy petitions and wound up through a series of screenings only using 1,032. His survey assumes that when a medical problem is mentioned that associated medical costs are automatically associated with bankruptcy. In addition, anyone is counted as medically bankrupt if they cite illness or medical bills as a reason for bankruptcy, even if other debts, such as foreclosure and credit card debt, are a primary reason.

Furthermore, if respondents lost two weeks of work due to illness or injury they were counted as medically bankrupt, even if they had no medical debt. Hypothetically, someone could go into bankruptcy while on Medicare or Medicaid, even if they owed no medical bills at all.
Given the frequency with which we hear medical bankruptcies cited as a key reason for health care reform this deserves some more attention.


Zach said...

Oh, more posts on statistics!!! Thanks for the blog material for later tonight, Colin. A very hearty HT will be heading out to you.

Let me just say, as a lay statistician, that such cherry picking of statistics is very interesting and wrong. I will describe how I would have approached trying to answer such a question later tonight, and it would have been the exact opposite of dear Dr. Himmelstein's study. The worst part is that he is likely not a statistician himself, yet his gross misuse of the tools smears people like myself who use the tools correctly to understand phenomena that are not otherwise obvious.

Harrison said...

Excellent thank you!

Paradigm Shifter said...

Colin, it took a little longer than I thought, but a post to my analysis of the two studies is below. Part of the delay came in the fact I went back and read both of the original studies. As suspected, the statistical analysis is highly flawed and doesn't support the author's conclusions. I hope you don't mind me posting a cross thread on your site. Read the link for more details.