In 2005, Senator Elizabeth Warren – then a professor – published an academic paper along with David Himmelstein, Deborah Thorne and Steffie Woolhandler, in the journal Health Affairs that made a very bold claim: Over 40 percent of all U.S. bankruptcies came as the result of medical problems suffered by the bankrupt.
Four years later, in 2009, they published an update to that figure – and it looked much, much worse. Medical bills were actually responsible for more than 62 percent of all American bankruptcy filings.
Needless to say, it was a very eye-catching statistic. Arguably, that was the point: Dr. Woolhandler and Dr. Himmelstein are physicians, public health researchers at Hunter College and advocates of a single-payer healthcare system. Ms. Thorne is a sociologist who studies bankruptcy. At the time, Warren was a Harvard law professor with a longtime issue in social justice and a professed desire to throw data at questions in the social sciences instead of speculation.
The work made a big impact. The statistic was often cited, to notable effect by presidential candidate Barack Obama, and later President Obama as he made his case for reforming healthcare. On the strength of her efforts and advocacy, Ms. Warren eventually went on to successfully win a seat in the United States Senate.
It’s also been influential on academics.
“I started becoming interested in medical bankruptcy partly after reading this paper,” Neale Mahoney, a health economist at the University of Chicago Booth School of Business, told The New York Times.
There’s just one small issue.
The data may be dead wrong – or at least terribly overstated.
It’s a dispute that began when the first paper was put out in 2005, and is recurring yet again 14 years later. The setting of the most recent spat over the numbers is on the pages of The New England Journal of Medicine. Warren and her co-authors critiqued a recent paper that argued medical problems cause a much smaller share of personal bankruptcies that most people believe (thanks in part to Warren’s 2005 numbers).
Points of Contention
From early days, the main point of contention was the study’s sample group: 1,800 Americans who had declared bankruptcy in 2001, half of whom were interviewed on their views about the causes of their financial woes.
Many experts pointed out that this is a potentially problematic method, since the entire interviewed population was bankrupt, and human beings have a tendency to emphasize sources of debt that make them seem sympathetic, while downplaying others.
Moreover, Amy Finkelstein, an MIT professor and co-author of the recent research that Warren and company are disputing in the journal, said that by focusing only on bankrupt people, the old research is badly skewing the sample.
Finkelstein and group argue that the approach of Warren’s paper is akin to trying to find out how to be successful in business by interviewing big technology entrepreneurs about how they got their start. Such a research bent, she noted, would lead one to the conclusion that dropping out of college is a great way to become a tech billionaire, despite the fact that the course of action far from correlates with great business success in the general population.
“The original paper was extremely problematic, precisely because it limited its analysis to the set of people who went bankrupt, and said how common really large medical bills are in that sample of people,” she said. “It says nothing about how common really large medical bills are in the non-bankrupt population.”
Moreover, critics have noted, researchers of the original paper were fairly expansive in their definition of bankruptcy by medical debt. Their criteria counted people who had medical debt of more than $1,000, who said that illness or injury caused a bankruptcy, who missed more than two weeks of work because of illness or who mortgaged a home to pay medical bills.
Neale Mahoney, the University of Chicago economist who was inspired into his work by the paper, noted that the team “wrote the paper in a way that was deliberately provocative, and they got out ahead of their skis.”
Notably, other research teams have come to very different conclusions about the data. Last year, a team of economists from MIT, Northwestern University and the University of California, Santa Cruz (Dr. Finkelstein’s team) drew their figures from the databases of every hospital in the state of California.
Looking at the credit reports of all the hospitalized people who filed for bankruptcy afterward, that team did not find that medical costs accounted for 62 percent of all bankruptcies – rather, they found that figure clocked in at around 4 percent.
Among other contemporary critiques: In 2006, David Dranove and Michael Millenson, then both on the faculty at Northwestern University’s Kellogg School of Management, looked at the same data set as the Warren team, and found that medical problems were probably responsible for less than 20 percent of all American bankruptcies, as opposed to more than half of them.
Craig Garthwaite, a health economist at Kellogg, said that at this point, politicians might like Warren and company’s figures – but economists don’t. “There are no reputable economists who I deal with who believe the number in the paper or the methods in the paper are appropriate in trying to get at the true underlying question.”
Pushback on Pushback
When Dranove and Millenson first published their research in 2006, the Warren team hit back, and pretty hard. In a letter defending their interpretation of the data, they noted that both Dranove and Millenson had received financial support from the insurance industry.
Dranove told The Times he was taken aback by the aggressiveness of their response, which he described as “disparaging our motives” rather than engaging in a scholarly discussion.
This time around, several outside researchers are praising the new work out of MIT, Northwestern and UC Santa Cruz as solving for many of the problems of the survey group in the original report, and getting closer to isolating the degree to which medical problems caused bankruptcies.
Once again, Senator Warren responded to the new research, noting that it might be undercounting medical debt because it is only looking at patients who received treatment in a hospital setting. That is problematic, experts note, because patients (particularly those with chronic conditions) will seek medical care, often very costly care, outside of a hospital setting. That care could contribute to bankruptcy, but would not appear in the new data.
David Cutler, a health economist at Harvard, echoed that critique. “If you look at the reasons why people are really sick today, they don’t have a lot to do with hospitalizations,” he said.
The study authors agree that this is a limit with their data, and that there is reason to believe they are undercounting when it comes to medical debt and bankruptcy. What they were probably less likely to agree with is the bit of an elbow Senator Warren threw in her response to the research.
“Characterizing debtors’ self-reports as ‘myth’ is demeaning to people struggling with healthcare costs, and artificially narrowing the definition of medical bankruptcy does not improve understanding of its causes,” she noted.
In an email, Dr. Himmelstein said the “Myth and Measurement” article “both misrepresented our research, and presented a highly slanted statistical analysis.”
Ms. Warren issued a milder statement later, saying the March study “significantly adds to our understanding of the links between illness and financial hardship.”
Ray Kluender, a co-author of the “Myth and Measurement” article, said he was disappointed by the new letter, but was delighted to see a United States senator engaging in research that could influence public policy.
“As long as she does not view us antagonistically, which I hope she doesn’t, I’m personally very excited about it,” he said. “This is why I decided to pursue economics as a career. I’m excited to have her attention.”
What both the new research and Ms. Warren’s original paper agree on is that it’s not just medical bills that push sick people toward bankruptcy, but the lost income due to illness that disrupts consumers’ financial lives to the point that they are pushed to complete insolvency by a medical emergency.
The majority of Americans, for example, do not have $400 in cash or credit to throw at an unexpected financial emergency. Both studies, disparate though they are, agree that the main problem might not be the bill alone, but the bill plus the lost income. It’s not just one factor, but a constellation of combined factors that pushes consumers into the red.