For immediate release:
Tuesday, February 28, 2006
12:01 a.m. EDT


Christopher Fleming

Medical Costs Contribute To Fewer Than One In Five Bankruptcies, Say Kellogg Management School Researchers

Dranove and Millenson Fault Analysis In Earlier Study,
But Himmelstein And Harvard Coauthors Defend One-In-Two Estimate

Bethesda, MD -- Data from a much-cited 2005 study show that medical expenses contribute to less than 20 percent of all bankruptcies, far fewer than the 50 percent estimate offered by the original study’s authors, according to an article by David Dranove of the Kellogg School of Management and Michael L. Millenson, a consultant and visiting scholar at the Kellogg School, published today as a Health Affairs Web Exclusive.

In “Medical Bankruptcy: Myth Versus Fact,” Dranove and Millenson also characterize most Americans who declare bankruptcy because of medical expenses as “marginally middle class” at best, and they say that a single-payer national health insurance system would not be an effective weapon against such bankruptcies.

In their paper, Dranove and Millenson critique “Illness and Injury as Contributors to Bankruptcy,” published 2 February 2005 as a Health Affairs Web Exclusive and written by David Himmelstein, Elizabeth Warren, Deborah Thorne, and Steffie Woolhander. Himmelstein and Woolhandler are members of the Harvard Medical School faculty; Warren teaches law at Harvard; and Thorne, formerly of Harvard, is currently an assistant professor of sociology and anthropology at Ohio University.

Two Perspectives on Dranove and Millenson’s paper also appear on the Health Affairs site. The first is Himmelstein and colleagues’ response to the Dranove and Millenson; the second is an overview offered by Robert Seifert and Mark Rukavina, the policy director and executive director of the Access Project. Dranove and Millenson also present a short response to Himmelstein and colleagues’ Perspective.

“It is insufficient to show that medical problems are associated with bankruptcy; one must determine whether, and to what extent, medical spending causes bankruptcies,” Dranove and Millenson write. To answer that question, “we must identify those people who stated that illness or injury was a cause of bankruptcy and also stated that medical bills contributed to bankruptcy.” In reanalyzing the data from Himmelstein and colleagues, Dranove and Millenson conclude that “medical expenditure bankruptcies” actually constituted only 17 percent of the bankruptcy filers those authors surveyed.

They add: “Even for that 17 percent, we cannot state with any degree of certainty whether medical expenses were the most important cause of bankruptcy. To move from causation to magnitude, it is necessary to perform multivariate statistical analysis,” using a control group of solvent individuals and including variables such as employment and marital status. While Himmelstein and colleagues did not perform this type of statistical analysis, other studies that did so concluded that “medical debt is like any other debt -- a cause but not the most important cause of bankruptcy,” Dranove and Millenson write.

Indeed, while Himmelstein and colleagues assert that “solidly middle-class Americans are at risk,” of being bankrupted by health care costs, Dranove and Millenson note that the median income of the debtors surveyed by Himmelstein and colleagues was only $25,000. This is a group for whom “health care spending amounting to a few thousand dollars in the two years prior to bankruptcy would represent just the tip of the iceberg threatening to sink their creditworthiness,” the two researchers say.

While Himmelstein and colleagues’ original article suggested that a Canadian-style national health insurance plan could be an effective weapon against medical expenditure bankruptcies, Dranove and Millenson cite research showing that direct medical costs are just one part of the large financial burden associated with serious illness. Any universal health insurance plan comprehensive enough to guarantee financial protection against all of those costs would have to be more comprehensive than any existing single-payer system, they write, making it so expensive that paying for it “might actually poke holes in the safety net for other vulnerable citizens.”

In their Perspective response, Himmelstein and colleagues charge that Dranove and Millenson “manipulate the data [from our original paper] far beyond legitimate reinterpretation.” The Harvard researchers defend the accuracy of their original conclusion that “medical problems contribute to about half of all bankruptcies.” They note that “Dranove and Millenson’s definition misclassifies costs for medications and home care as ‘nonmedical’” and excludes, for instance, “bankrupt families who took out second mortgages to pay medical bills [but] described their reasons for filing as ‘to save our home.’”

Himmelstein and colleagues write that debtors’ $25,000 median income “in the year prior to filing reflects the setbacks that led to bankruptcy (for example, illness that caused job loss), not socioeconomic status before the onset of disaster.” They point to the fact that most debtors had owned homes, attended college, and held good jobs, as proof of medical debtors’ middle class status. In addition, they criticize Dranove and Millenson for “misrepresent[ing] the previous literature on medical bankruptcy” -- for example, ignoring one study’s conclusion that “medical debts are often buried in credit card balances and second mortgages,” as well as a Commonwealth Fund study’s finding that “37.2 million adults annually are ‘contacted by a collection agency about owing money for medical bills.’”

The Harvard researchers also defend their view that national health insurance could prevent most medical bankruptcies, writing that “many other countries have implemented comprehensive health and disability coverage that does the job [of protecting against illness-related financial ruin] quite well.” Himmelstein and Woolhandler are co-founders of the group Physicians for a National Health Program, which advocates for single-payer national health insurance in the United States.

Himmelstein and colleagues note that “most [U.S.] medical debtors had taken the prescribed steps to protect themselves. Three-quarters had health insurance when they got sick. But for many, private coverage lapsed when they lost work because of illness. In other cases, the insurance they bought in good faith had so many holes that it left them unprotected.” Himmelstein and colleagues write that Dranove and Millenson “should take a hard look at the defective products sold by their funders in the insurance industry.”

In their Perspective overview of the controversy, Seifert and Rukavina do not address the specific incidence of medically induced bankruptcies. Whatever that number may be, they say, medical debt is “surprisingly common,” affecting the insured as well as the uninsured, the middle class as well as the poor. Seifert and Rukavina state that medical debt both undermines families’ financial security and functions as a barrier to health care. Indeed, “the care-seeking behavior of privately insured adults with medical debt is much more like that of uninsured people,” the two authors state.

Seifert and Rukavina assert that simply counting the number of uninsured people is not an accurate measure of the success of efforts to improve access to health care. “It is time to expand our policy vocabulary beyond insurance coverage and begin to examine the adequacy of that coverage in ensuring health access and improving health status and financial security,” the authors declare.

You can read the Dranove and Millenson paper at,

Himmelstein and colleagues’ Perspective at,

Seifert and Rukavina’s Perspective at,

and Dranove and Millenson’s final response at


Health Affairs, published by Project HOPE, is the leading journal of health policy. The peer-reviewed journal appears bimonthly in print with additional online-only papers published weekly as Health Affairs Web Exclusives at


©2006 Project HOPE–The People-to-People Health Foundation, Inc.