Where we stand (divided) on medical bankruptcy

The following is a guest post by Daniel Liebman, a research assistant for Dr. Ashish Jha at the Harvard School of Public Health, and a part-time research assistant for The Incidental Economist. He graduated from Brandeis University in 2012 with degrees in Health Policy and American Studies, and will begin medical school in Fall 2014. He tweets about good policy and bad puns at @D_Liebman.

What proportion of US personal bankruptcies are caused by medical bills? It’s a straightforward question that belies a cantankerous academic and political debate. After all, if medical bankruptcy is infrequent, a blow is dealt to one rationale for insurance expansion. If it’s common, coverage expansion is a natural policy solution.

What’s the answer? I took a detailed look at the literature on this issue. My full annotated bibliography is here; what follows is the gist.

Let’s begin with the island of consensus: medical bankruptcies do indeed occur. Studies have documented the increased risk of bankruptcy from serious illness or injury; Ramsey et al. (2013), for example, reviewed Washington State court records from 1995-2009, and found that adult cancer patients were 2.65 times more likely to experience bankruptcy than individuals without cancer. Similar studies with similar findings include Hollingworth et al., (2007) and Relyea-Chew et al. (2009).

The fireworks begin when we turn to estimates of the overall incidence of US medical bankruptcy. Here, just about everything—definition, methodology, interpretation, and implications—is contested.

Two comprehensive, peer-reviewed studies on the question come from David Himmelstein and colleagues (including then-Professor Elizabeth Warren). They got the ball rolling in 2005 with a headline-grabbing Health Affairs study , reporting that “major medical bankruptcies” accounted for 46.2% of US bankruptcies in 2001. Using a sample of 1,711 bankruptcy filers from five federal courts,       the study painted a portrait of a middle class besieged by underinsurance. Of the bankruptcy filers, 75% had health insurance at the onset of their injury or illness. Extrapolating, Himmelstein et al. estimated that medical bankruptcy impacts 1.9-2.2 million Americans annually.

Bankruptcy laws later tightened, leading Team Himmelstein to conduct a follow-up study using a national sample and a slightly more conservative definition of medical bankruptcy. They found that medical bankruptcies rose to 62.1% of all bankruptcies in 2007, while the rate of insured filers remained ~75%. On the basis of these two studies, the authors concluded that medical bills and illness contribute to a “large and increasing share of US bankruptcies”.

Himmelstein’s findings drew a great deal of attention; President Obama referenced their work in his opening remarks at the 2009 White House health care summit. The criticism came hard and fast. The academic debate quickly focused on the metrics Himmelstein used to define medical bankruptcy—an argument encapsulated in a 2006 Health Affairs response article from David Dranove and Michael Millenson.

Dranove and Millenson strongly contested Himmelstein’s medical bankruptcy definition, which identified a “major medical bankruptcy” using any of the following criteria: (1) illness or injury cited as a specific reason for bankruptcy; (2) uncovered medical bills in the past two years exceeding $1,000; (3) 2+ weeks of income lost due to illness/injury; and/or (4) mortgaged a home to pay medical bills. Dranove and Millenson contended that only the first criterion could be used as an indicator of causality, and point out that only 28.3% of filers in the Himmelstein study satisfied it. Beginning with this critique, the authors used Himmelstein’s data to re-estimate the medical bankruptcy rate at 17%.

Dranove and Millenson’s work sparked a fiery back-and-forth. Himmelstein accused Dranove and Millenson of “ignoring most of the data and misrepresenting the rest”, while Dranove questioned whether Himmelstein’s research was being “driven by ideology”. The dueling viewpoints flesh out what are essentially the two current schools of thought on the issue— Himmelstein arguing that that a narrow interpretation ignores the complexity of medical bankruptcy, Dranove contending that a broad definition fails to establish the relative importance of medical bills compared to other debt.

A smattering of other studies sought to address this issue. An American Enterprise Institute analysis  reported that 27% of bankruptcy filings were primarily attributable to medical debt; a 2005 DOJ report commissioned by Senator Chuck Grassley found that 13% of total unsecured debt was medical in nature; a 2008 report from the Center for Studying Health System Change corroborated Himmelstein’s finding that 2.2 million individuals live in medically bankrupt families.

Support can also be found for both broad and narrow definitions of medical bankruptcy. Jacoby & Holman (2010) demonstrated that use of court filings and stated bankruptcy reasons alone can underestimate the true incidence of medical bankruptcy, since medical debt can be obscured in the court records. Zhu (2011) found “some support for both positions”, stating that adverse events like injury/illness can trigger personal bankruptcy, but that overall consumption patterns tend to matter more.  

Apart from the methodological and empirical details, the issue of medical bankruptcy serves as a battleground in the larger war over health reform. Some see medical bankruptcy as a telltale sign of a broader failure of the private insurance market; others see an attempt to shift the burden away from personal responsibility. Those of the former persuasion tend to favor a government-directed—even single-payer—policy solution, while those of the latter generally favor a market-based approach.

Will the ACA put a dent in the bankruptcy rate, whatever it may be? It’s tough to say. One study found that the medical bankruptcy rate did not decline in Massachusetts immediately following its health reform effort, though there could be some confounding due to the Great Recession. Also, the ACA differs with regard to cost support and Medicaid expansion. The Oregon Health Study did recently find that Medicaid all but eliminated catastrophic out-of-pocket expenditures. (Aaron has previously discussed this.)

Regardless of the particular image you see reflected in the literature, I would highlight Seifert and Rukavina’s 2006 Health Affairs Perspective as an appropriate lens through which to view this issue: medical bankruptcy, whatever its “true” rate, represents the tip of a larger iceberg of medical debt and heath care costs in this country.  One in five American families struggle to pay their medical bills. Without diminishing the suffering that accompanies bankruptcy, our approach to addressing the larger burden of health care costs isn’t and shouldn’t be driven by whether the medical bankruptcy rate is 17% or 50%. Either way it’s too high. Either way bankruptcy is but one symptom of a much larger disease.

Hidden information below


Email Address*