Donate Today

Harvard Study Finds Medical Bills Push Many Into Bankruptcy

Nearly half of all Americans who file for bankruptcy do so because of medical expenses, according to a new study released jointly by researchers at Harvard Law School and Harvard Medical School. The study, which is based on surveys of 1,771 individuals filing for bankruptcy, is the first of its kind to gather extensive information on the correlation between medical conditions and expenses and bankruptcy.

"We discovered that in 2004 about two million men, women and children were swept through the bankruptcy system in the fallout of a medical problem," explained Elizabeth Warren, professor of law and co-author of the report. "Good educations, decent jobs, and health insurance were no guarantee that a person wouldn't be wiped out by an illness or accident. We believe the current policy debates are overlooking a critical problem: A broken health care finance system is bankrupting middle class America."

The report is available below and with exhibits on the website of Health Affairs. By clicking here you will leave

"Illness And Injury As Contributors To Bankruptcy"


By David U. Himmelstein, Elizabeth Warren, Deborah Thorne, and Steffie Woolhandler
* footnotes appear in brackets ([1])


In 2001, 1.458 million American families filed for bankruptcy. To investigate medical contributors to bankruptcy, we surveyed 1,771 personal bankruptcy filers in five federal courts and subsequently completed in-depth interviews with 931 of them. About half cited medical causes, which indicates that 1.9¿2.2 million Americans (filers plus dependents) experienced medical bankruptcy. Among those whose illnesses led to bankruptcy, out-of-pocket costs averaged $11,854 since the start of illness; 75.7 percent had insurance at the onset of illness. Medical debtors were 42 percent more likely than other debtors to experience lapses in coverage. Even middle-class insured families often fall prey to financial catastrophe when sick.

    If the debtor be insolvent to serve creditors, let his body be cut in pieces on the third market day. It may be cut into more or fewer pieces with impunity. Or, if his creditors consent to it, let him be sold to foreigners beyond the Tiber. ¿Twelve Tables, Table III, 6 (ca. 450 B.C.)

Our bankruptcy system works differently from that of ancient Rome; creditors carve up the debtor¿s assets, not the debtor. Even so, bankruptcy leaves painful problems in its wake. It remains on credit reports for a decade, making everything from car insurance to house payments more expensive.[1] Debtors¿ names are often published in the newspaper, and the fact of their bankruptcy may show up whenever someone tries to find them via the Internet. Potential employers who run routine credit checks (a common screening practice) will discover the bankruptcy, which can lead to embarrassment or, worse, the lost chance for a much-needed job.[2]

Personal bankruptcy is common. Nearly 1.5 million couples or individuals filed bankruptcy petitions in 2001, a 360 percent increase since 1980.[3] Fragmentary data from the legal literature suggest that illness and medical bills contribute to bankruptcy. Most previous studies of medical bankruptcy, however, have relied on court records¿where medical debts may be subsumed under credit card or mortgage debt¿or on responses to a single survey question.[4] None has collected detailed information on medical expenses, diagnoses, access to care, work loss, or insurance coverage. Research has been impeded both by the absence of a national repository for bankruptcy filings and by debtors¿ reticence to discuss their bankruptcy; in population-based surveys, only half of those who have undergone bankruptcy admit to it.[5]

The health policy literature is virtually silent on bankruptcy, although a few studies have looked at impoverishment attributable to illness. In his 1972 book, Sen. Edward Kennedy (D-MA) gave an impressionistic account of ¿sickness and bankruptcy.¿[6] The likelihood of incurring high out-of-pocket costs was incorporated into older estimates of the number of underinsured Americans: twenty-nine million in 1987.[7] About 16 percent of families now spend more than one-twentieth of their income on health care.[8] Among terminally ill patients (most of them insured), 39 percent reported that health care costs caused moderate or severe financial problems.[9] Medical debt is common among the poor, even those with insurance, and interferes with access to care.[10] At least 8 percent, and perhaps as many as 21 percent, of American families are contacted by collection agencies about medical bills annually.[11]

Our study provides the first extensive data on the medical concomitants of bankruptcy, based on a survey of debtors in bankruptcy courts. We address the following questions: (1) Who files for bankruptcy? (2) How frequently do illness and medical bills contribute to bankruptcy? (3) When medical bills contribute, how large are they and for what services? (4) Does inadequate health insurance play a role in bankruptcy? (5) Does bankruptcy compromise access to care?

A Brief Primer On Bankruptcy

¿Bankrupt¿ is not synonymous with ¿broke.¿ ¿Bankrupt¿ means filing a petition in a federal court asking for protection from creditors via the bankruptcy laws. A single petition may cover an individual or married couple. The instant a debtor files for bankruptcy, the court assumes legal control of the debtor¿s assets and halts all collection efforts.

Shortly after the filing, a court-appointed trustee convenes a meeting to inventory the debtor¿s assets and debts and to determine which assets are exempt from seizure. States may regulate these exemptions, which often include work tools, clothes, Bibles, and some equity in a home.

About 70 percent of all consumer debtors file under Chapter 7 of the Bankruptcy Code; most others file under Chapter 13. In Chapter 7 the trustee liquidates all nonexempt assets¿although 96 percent of debtors have so little unencumbered property that there is nothing left to liquidate. At the conclusion of the bankruptcy, the debtor is freed from many debts. In Chapter 13 the debtor proposes a repayment plan, which extends for up to five years. Chapter 13 debtors may retain their property so long as they stay current with their repayments.

Under both chapters, taxes, student loans, alimony, and child support remain payable in full, and debtors must make payments on all secured loans (such as home mortgages and car loans) or forfeit the collateral.

Study Data And Methods

study is based on a cohort of 1,771 bankruptcy filings in 2001. For each filing, a debtor completed a written questionnaire at the mandatory meeting with the trustee, and we abstracted financial data from public court records. In addition, we conducted follow-up telephone interviews with about half (931) of these debtors.

Sampling strategy. We used cluster sampling to assemble a cohort of households filing for personal bankruptcy in five (of the seventy-seven total) federal judicial districts.[12] We collected 250 questionnaires in each district, representative of the proportion of Chapters 7 and 13 filings in that district. These 1,250 cases constitute our ¿core sample.¿ For planned studies on housing, we collected identical data from an additional 521 homeowners filing for bankruptcy. We based our analyses on all 1,771 bankruptcies with responses weighted to maintain the representativeness of the sample.[13]

Data collection. With the cooperation of the judges in each district, we contacted the trustees who officiate at meetings with debtors. The trustees agreed to distribute, or to allow a research assistant to distribute, a self-administered questionnaire to debtors appearing at the bankruptcy meeting. Questionnaires (which were available in English and Spanish) included a cover letter explaining the research project and human subjects protections and encouraging debtors to consult their attorneys (who were almost always present) before participating.

The questionnaire asked about demographics, employment, housing, and specific reasons for filing for bankruptcy; it also asked whether the debtor had medical debts exceeding $1,000, had lost two or more weeks of work-related income because of illness, or had health insurance coverage for themselves and all dependents at the time of filing, and whether there had been a gap of one month or more in that coverage during the past two years. In joint filings, we collected demographic information for each spouse.

During the spring and summer of 2001 we collected questionnaires from consecutive debtors in each district until the target number was reached.[14]

Follow-up telephone interviews. The written questionnaire distributed at the time of bankruptcy filing invited debtors to participate in future telephone interviews, for which they would receive $50; 70 percent agreed to such interviews. We ultimately completed follow-up telephone interviews with 931 of the 1,771 debtor families, a response rate of 53 percent.[15] The telephone interviews, conducted between June 2001 and February 2002 using a structured, computer-assisted protocol, explored financial, housing, and medical issues. Many debtors also provided a narrative description of their bankruptcy experience.

Detailed medical questions. Each of the 931 interviewees was asked if any of the following had been a significant cause of their bankruptcy: an illness or injury; the death of a family member; or the addition of a family member through birth, adoption, custody, or fostering. Those who answered yes to this screening question were queried about diagnoses, health insurance during the illness, and medical care use and spending. Interviewers collected information about each household member with medical problems. In total, we collected in-depth medical information on 391 people with health problems in 332 debtor households.

Data analysis. We used data from the self-administered questionnaires (and court records) obtained from all 1,771 filers to analyze demographics, health coverage at the time of filing, and gaps in coverage in the two years before filing.

We also used the questionnaire to estimate how frequently illness and medical bills contributed to bankruptcy. We developed two summary measures of medical bankruptcy. Under the rubric ¿Major Medical Bankruptcy¿ we included debtors who either (1) cited illness or injury as a specific reason for bankruptcy, or (2) reported uncovered medical bills exceeding $1,000 in the past years, or (3) lost at least two weeks of work-related income because of illness/injury, or (4) mortgaged a home to pay medical bills. Our more inclusive category, ¿Any Medical Bankruptcy,¿ included debtors who cited any of the above, or addiction, or uncontrolled gambling, or birth, or the death of a family member.[16]

Data from the 931 follow-up telephone interviews were used to analyze hardships experienced by debtors in the period surrounding their bankruptcy, including problems gaining access to medical care. The in-depth medical interviews regarding 391 people with medical problems are the basis for our analyses of which household members were ill, diagnoses, health insurance at onset of illness, and out-of-pocket spending. Two physicians (Himmelstein and Woolhandler) coded the diagnoses given by debtors into categories for analysis.

SAS and SUDAAN were used for statistical analyses, adjusting for complex sample design. To extrapolate our findings nationally, we assumed that our core sample was representative of the 1,457,572 households filing for bankruptcy during 2001. Human subject committees at Harvard Law School and the Cambridge Hospital approved the project.

Study Findings
Who files for bankruptcy? Exhibit 1 displays the demographic characteristics of our weighted sample of 1,771 bankruptcy filers. The average debtor was a forty-one-year-old woman with children and at least some college education. Most debtors owned homes; their occupational prestige scores place them predominantly in the middle or working classes.

On average, each bankruptcy involved 1.32 debtors (reflecting some joint filings by married couples) and 1.33 dependents. Extrapolating from our data, the 1.5 million personal bankruptcy filings nationally in 2001 involved 3.9 million people: 1.9 million debtors, 1.3 million children under age eighteen, and 0.7 million other dependents.

Medical causes of bankruptcy. Exhibit 2 shows the proportion of debtors (N = 1,771) citing various medical contributors to their bankruptcy and the estimated number of debtors and dependents nationally affected by each cause. More than one-quarter cited illness or injury as a specific reason for bankruptcy; a similar number reported uncovered medical bills exceeding $1,000. Some debtors cited more than one medical contributor. Nearly half (46.2 percent) (95 percent confidence interval = 43.5, 48.9) of debtors met at least one of our criteria for ¿major medical bankruptcy.¿ Slightly more than half (54.5 percent) (95 percent CI = 51.8, 57.2) met criteria for ¿any medical bankruptcy.¿

A lapse in health insurance coverage during the two years before filing was a strong predictor of a medical cause of bankruptcy (Exhibit 3). Nearly four-tenths (38.4 percent) of debtors who had a ¿major medical bankruptcy¿ had experienced a lapse, compared with 27.1 percent of debtors with no medical cause (p < .0001). Surprisingly, medical debtors were no less likely than other debtors to have coverage at the time of filing. (More detailed coverage and cost data for the subsample we interviewed appears below.)

Medical debtors resembled other debtors in most other respects (Exhibit 1). However, the ¿major medical bankruptcy¿ group was 16 percent (p < .03) less likely than other debtors to cite trouble managing money as a cause of their bankruptcy (data not shown).

Privations in the period surrounding bankruptcy. In our follow-up telephone interviews with 931 debtors, they reported substantial privations. During the two years before filing, 40.3 percent had lost telephone service; 19.4 percent had gone without food; 53.6 percent had gone without needed doctor or dentist visits because of the cost; and 43.0 percent had failed to fill a prescription, also because of the cost. Medical debtors experienced more problems in access to care than other debtors did; three-fifths went without a needed doctor or dentist visit, and nearly half failed to fill a prescription (Exhibit 4).

Medical debt was also associated with mortgage problems. Among the total sample of 1,771 debtors, those with more than $1,000 in medical bills were more likely than others to have taken out a mortgage to pay medical bills (5.0 percent versus 0.8 percent). Fifteen percent of all homeowners who had taken out a second or third mortgage cited medical expenses as a reason. Follow-up phone interviews revealed that among homeowners with high-cost mortgages (interest rates greater than 12 percent, or points plus fees of at least 8 percent), 13.8 percent cited a medical reason for taking out the loan.

Following their bankruptcy filings, about one-third of debtors continued to have problems paying their bills. Medical debtors reported particular problems making mortgage/ rent payments and paying for utilities (Exhibit 4). Although our interviews occurred soon after the bankruptcy filings (seven months, on average), many debtors had already been turned down for jobs (3.1 percent), mortgages (5.8 percent), apartment rentals (4.9 percent), or car loans (9.3 percent) because of the bankruptcy on their credit reports.

Medical diagnoses, spending, and type of coverage. Our interviews yielded detailed data on diagnoses, health insurance coverage, and medical bills for 391 debtors or family members whose medical problems contributed to bankruptcy. In three-quarters of cases, the person experiencing the illness/ injury was the debtor or spouse of the debtor; in 13.3 percent, a child; and in 8.2 percent, an elderly relative.

Illness begot financial problems both directly (because of medical costs) and through lost income. Three-fifths (59.9 percent) of families bankrupted by medical problems indicated that medical bills (from medical care providers) contributed to bankruptcy; 47.6 percent cited drug costs; 35.3 percent had curtailed employment because of illness, often (52.8 percent) to care for someone else. Many families had problems with both medical bills and income loss.

Families bankrupted by medical problems cited varied, and sometimes multiple, diagnoses. Cardiovascular disorders were reported by 26.6 percent; trauma/orthopedic/back problems by nearly one-third; and cancer, diabetes, pulmonary, or mental disorders and childbirth-related and congenital disorders by about 10 percent each. Half (51.7 percent) of the medical problems involved ongoing chronic illnesses.

Our in-depth interviews with medical debtors confirmed that gaps in coverage were a common problem. Three-fourths (75.7 percent) of these debtors were insured at the onset of the bankrupting illness. Three-fifths (60.1 percent) initially had private coverage, but one-third of them lost coverage during the course of their illness. Of debtors, 5.7 percent had Medicare, 8.4 percent Medicaid, and 1.6 percent veterans/military coverage. Those covered under government programs were less likely than others to have experienced coverage interruptions.

Few medical debtors had elected to go without coverage. Only 2.9 percent of those who were uninsured or suffered a gap in coverage said that they had not thought they needed insurance; 55.9 percent said that premiums were unaffordable; 7.1 percent were unable to obtain coverage because of preexisting medical conditions; and most others cited employment issues, such as job loss or ineligibility for employer-sponsored coverage.

Debtors¿ out-of-pocket medical costs were often below levels that are commonly labeled catastrophic. In the year prior to bankruptcy, out-of-pocket costs (excluding insurance premiums) averaged $3,686 (95 percent CI = $2,693, $4,679) (Exhibit 5). Presumably, such costs were often ruinous because of concomitant income loss or because the need for costly care persisted over several years. Out-of-pocket costs since the onset of illness/injury averaged $11,854 (95 percent CI = $8,532, $15,175). Those with continuous insurance coverage paid $734 annually in premiums on average, over and above the expenditures detailed above. Debtors with private insurance at the onset of their illnesses had even higher out-of-pocket costs than those with no insurance (Exhibit 5). This paradox is explained by the very high costs¿$18,005¿incurred by patients who initially had private insurance but lost it. Among families with medical expenses, hospital bills were the biggest medical expense for 42.5 percent, prescription medications for 21.0 percent, and doctors¿ bills for 20.0 percent. Virtually all of those with Medicare coverage, and most patients with psychiatric disorders, said that prescription drugs were their biggest expense.

The human face of bankruptcy. Debtors¿ narratives painted a picture of families arriving at the bankruptcy courthouse emotionally and financially exhausted, hoping to stop the collection calls, save their homes, and stabilize their economic circumstances. Many of the debtors detailed ongoing problems with access to care. Some expressed fear that their medical care providers would refuse to continue their care, and a few recounted actual experiences of this kind. Several had used credit cards to charge medical bills they had no hope of paying.

The co-occurrence of medical and job problems was a common theme. For instance, one debtor underwent lung surgery and suffered a heart attack. Both hospitalizations were covered by his employer-based insurance, but he was unable to return to his physically demanding job. He found new employment but was denied coverage because of his preexisting conditions, which required costly ongoing care. Similarly, a teacher who suffered a heart attack was unable to return to work for many months, and hence her coverage lapsed. A hospital wrote off her $20,000 debt, but she was nonetheless bankrupted by doctors¿ bills and the cost of medications.

A second common theme was sounded by parents of premature infants or chronically ill children; many took time off from work or incurred large bills for home care while they were at their jobs.

Finally, many of the insured debtors blamed high copayments and deductibles for their financial ruin. For example, a man insured through his employer (a large national firm) suffered a broken leg and torn knee ligaments. He incurred $13,000 in out-of-pocket costs for copayments, deductibles, and uncovered services¿much of it for physical therapy.


Bankruptcy is common in the United States, involving nearly four million debtors and dependents in 2001; medical problems contribute to about half of all bankruptcies. Medical debtors, like other bankruptcy filers, were primarily middle class (by education and occupation). The chronically poor are less likely to build up debt, have fewer assets (such as a home) to protect, and have less access to the legal resources needed to navigate a complex financial rehabilitation. The medical debtors we surveyed were demographically typical Americans who got sick. They differed from others filing for bankruptcy in one important respect: They were more likely to have experienced a lapse in health coverage. Many had coverage at the onset of their illness but lost it. In other cases, even continuous coverage left families with ruinous medical bills.

Study strengths and limitations. Our study¿s strengths are the use of multiple overlapping data sources; a large sample size; geographic diversity; and in-depth data collection. Although our sample may not be fully representative of all personal bankruptcies, the Chapter 7 filers we studied resemble Chapter 7 filers nationally (the only group for whom demographic data has been compiled nationally from court records).[17] Several indicators suggest that response bias did not greatly distort our findings.[18]

As in all surveys, we relied on respondents¿ truthfulness. Might some debtors blame their predicament on socially acceptable medical problems rather than admitting to irresponsible spending? Several factors suggest that our respondents were candid. First, just prior to answering our questionnaire, debtors had filed extensive financial information with the court under penalty of perjury¿information that was available to us in the court records and that virtually never contradicted the questionnaire data. They were about to be sworn in by a trustee (who often administered our questionnaire) and examined under oath. At few other points in life are full disclosure and honesty so aggressively emphasized.

Second, the details called for in our telephone interview¿questions about out-of-pocket medical expenses, who was ill, diagnoses, and so forth¿would make a generic claim that ¿we had medical problems¿ difficult to sustain. Third, one of us (Thorne) interviewed (for other studies) many debtors in their homes. Almost all specifically denied spendthrift habits, and observation of their homes supported these claims. Most reflected the lifestyle of people under economic constraint, with modest furnishings and few luxuries. Finally, our findings receive indirect corroboration from recent surveys of the general public that have found high levels of medical debt, which often result in calls from collection agencies.[19]

Even when data are reliable, making causal inferences from a cross-sectional study such as ours is perilous. Many debtors described a complex web of problems involving illness, work, and family. Dissecting medical from other causes of bankruptcy is difficult. We cannot presume that eliminating the medical antecedents of bankruptcy would have prevented all of the filings we classified as ¿medical bankruptcies.¿ Conversely, many people financially ruined by illness are undoubtedly too ill, too destitute, or too demoralized to pursue formal bankruptcy. In sum, bankruptcy is an imperfect proxy for financial ruin.

Trends in medical bankruptcy. Although methodological inconsistencies between studies preclude precise quantification of time trends, medical bankruptcies are clearly increasing. In 1981 the best evidence available suggests that about 25,000 families filed for bankruptcy in the aftermath of a serious medical problem (8 percent of the 312,000 bankruptcy filings that year).[20] Our findings suggest that the number of medical bankruptcies had increased twenty-threefold by 2001. Since the number of bankruptcy filings rose 11 percent in the eighteen months after the completion of our data collection, the absolute number of medical bankruptcies almost surely continues to increase.[21]

Policy implications. Our data highlight four deficiencies in the financial safety net for American families confronting illness. First, even brief lapses in insurance coverage may be ruinous and should not be viewed as benign. While forty-five million Americans are uninsured at any point in time, many more experience spells without coverage. We found little evidence that such gaps were voluntary. Only a handful of medical debtors with a gap in coverage had chosen to forgo insurance because they had not perceived a need for it; the overwhelming majority had found coverage unaffordable or effectively unavailable. The privations suffered by many debtors¿going without food, telephone service, electricity, and health care¿lend credence to claims that coverage was unaffordable and belie the common perception that bankruptcy is an ¿easy way out.¿

Second, many health insurance policies prove to be too skimpy in the face of serious illness. We doubt that such underinsurance reflects families¿ preference for risk; few Americans have more than one or two health insurance options. Many insured families are bankrupted by medical expenses well below the ¿catastrophic¿ thresholds of high-deductible plans that are increasingly popular with employers. Indeed, even the most comprehensive plan available to us through Harvard University leaves faculty at risk for out-of-pocket expenses as large as those reported by our medical debtors.

Third, even good employment-based coverage sometimes fails to protect families, because illness may lead to job loss and the consequent loss of coverage. Lost jobs, of course, also leave families without health coverage when they are at their financially most vulnerable.

Finally, illness often leads to financial catastrophe through loss of income, as well as high medical bills. Hence, disability insurance and paid sick leave are also critical to financial survival of a serious illness.

Only broad reforms can address these problems. Even universal coverage could leave many Americans vulnerable to bankruptcy unless such coverage was much more comprehensive than many current policies. As in Canada and most of western Europe, health insurance should be divorced from employment to avoid coverage disruptions at the time of illness. Insurance policies should incorporate comprehensive stop-loss provisions, closing coverage loopholes that expose insured families to unaffordable out-of-pocket costs. Additionally, improved programs are needed to replace breadwinners¿ incomes when they are disabled or must care for a loved one. The low rate of medical bankruptcy in Canada suggests that better medical and social insurance could greatly ameliorate this problem in the United States.[22]

In 1591 Pope Gregory XIV fell gravely ill. His doctors prescribed pulverized gold and gems. According to legend, the resulting depletion of the papal treasury is reflected in his unadorned plaster sarcophagus in St. Peter¿s Basilica.[23] Four centuries later, solidly middle-class Americans still face impoverishment following a serious illness.

This paper was supported by Grant no. 042425 from the Robert Wood Johnson Foundation. The Ford Foundation, Harvard Law School, and New York University Law School provided funding for other portions of this project. Teresa Sullivan, Jay Lawrence Westbrook, Robert Lawless, Bruce Markell, Michael Schill, Susan Wachter, Katherine Porter, and John Pottow played key roles in the overall project. Joel Cantor reviewed the medical questionnaire.


1. J. Guest, ¿High Rate Robbery,¿ Consumer Reports 67, no. 10 (2002): 7.
2. E. Warren and A.W. Tyagi, The Two-Income Trap (New York: Basic Books, 2003).
3. U.S. Bureau of the Census, Statistical Abstract of the United States, 1982¿83 (Washington: U.S. Department of Commerce, 1983); and Administrative Office of the U.S. Courts, ¿Record Breaking Bankruptcy Filings Reported in Calendar Year 2001,¿ Press Release (Washington: Administrative Office, 19 February 2002).
4. See M.B. Jacoby, T.A. Sullivan, and E. Warren, ¿Rethinking the Debates over Health Care Financing: Evidence from the Bankruptcy Courts,¿ New York University Law Review 76 (2001): 375¿418.
5. S. Fay, E. Hurst, and M.J. White, ¿The Household Bankruptcy Decision,¿ American Economic Review 92, no. 3 (2002): 706¿711.
6. E.M. Kennedy, In Critical Condition: The Crisis in America¿s Health Care (New York: Simon and Schuster, 1972).
7. P.F. Short and J.S. Banthin, ¿New Estimates of the Underinsured Younger than Sixty-five Years,¿ Journal of the American Medical Association 274, no. 16 (1995): 1302¿1306.
8. M. Merlis, Family Out-of-Pocket Spending for Health Services (New York: Commonwealth Fund, June 2002).
9. E.J. Emanuel et al., ¿Understanding Economic and Other Burdens of Terminal Illness: The Experience of Patients and Their Caregivers,¿ Annals of Internal Medicine 132, no. 6 (2000): 451¿459.
10. Access Project, The Consequences of Medical Debt: Evidence from Three Communities, February 2003, (13 December 2004).
11. NPR/Kaiser Family Foundation/Kennedy School of Government, ¿National Survey on Health Care (chartpack),¿ June 2002, January 2005); J.H. May and P.J. Cunningham, Tough Trade-Offs: Medical Bills, Family Finances, and Access to Care (Washington: Center for Studying Health System Change, June 2004); and S.R. Collins et al., The Affordability Crisis in U.S. Health Care: Findings from the Commonwealth Fund Biennial Health Insurance Survey (New York: Commonwealth Fund, March 2004).
12. The districts were California (Central District); Illinois (Northern District); Pennsylvania (Eastern District); Tennessee (Middle District); and Texas (Northern District). These were chosen to achieve geographic, social, and legal diversity. Together the five districts accounted for 13.8 percent of total U.S. bankruptcy filings in 2001.
13. The 521 extra homeowner cases make the full sample of 1,771 less representative of filers nationally than our core sample of 1,250. Therefore, we used weighting procedures to adjust for the oversampling of debtors in three districts, homeowners, and debtors filing under Chapter 13. The weighted and unweighted findings were little different.
14. Interviews with trustees indicate that response rates in the five districts varied from approximately 55 percent to nearly 100 percent.
15. It proved difficult to contact some debtors, presumably because they were experiencing major life disruptions or were afraid of calls from creditors. After ten unsuccessful attempts to telephone potential subjects, we attempted to reach them through contacts they had previously given us and via a letter. Relative to the overall sample, the 931 interviewed debtors were slightly less likely to be male, to have lost a home, or to reside outside of Illinois but did not differ in age, occupational prestige score, education, or home ownership. On occupational prestige scores, see NORC, ¿Occupational Prestige Studies/Summary,¿ .htm (13 December 2004).
16. Uncontrolled gambling is classified as a psychiatric disorder in the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) and contributed to about 1 percent of the bankruptcies.
17. Our Chapter 7 filers are similar to those nationally in income, home ownership, family size, age distribution, and marital status. See E. Flynn et al., ¿Bankruptcy by the Numbers,¿ ABI Journal 20, no. 10 (2002): 28¿29; 21, no. 3 (2002): 22, 49; and 20, no. 8 (2001): 20.
18. We achieved high response rates to our initial questionnaire, and rates of medical bankruptcy varied little between districts despite some variation in response rates. It seems plausible that more-stigmatized causes of bankruptcy (such as addiction, mental illness, or profligate spending) may be underreported.
19. Twelve percent of households reported unpaid medical debts in 1997, with 691,000 households reporting outstanding medical debts greater than $5,000. SMR Research Corporation, The New Bankruptcy Epidemic (Hackettstown, N.J.: SMR, 2001), 127. A high incidence of collection agency calls was reported in NPR/Kaiser et al., ¿National Survey on Health Care¿; May et al., Tough Trade-Offs; and Collins et al., The Affordability Crisis.
20. For the total number of bankruptcies, see U.S. Bureau of the Census, Statistical Abstract of the United States: 1986 (Washington: GPO, 1985). The estimate that 8 percent of these were medical is from T.A. Sullivan, E. Warren, and J.L. Westbrook, The Fragile Middle Class: Americans in Debt (New Haven, Conn.: Yale University Press, 2000).
21. Administrative Office of the U.S. Courts, ¿Bankruptcy Cases Continue to Break Federal Court Case Records: Total Bankruptcy Filings and Non-Business Filings Hit Highs,¿ Press Release, 18 August 2003, December 2004).
22. Between 7.1 percent and 14.3 percent of Canadian bankruptcies are attributable to ¿health/misfortune.¿ See J.S. Ziegel, ¿A Canadian Perspective,¿ Texas Law Review 79, no. 5 (2001): 241¿256.
23.Rome/Vatican City (Clermont-Ferrand, France: Michelin Travel Publications, 2001).

David Himmelstein ([email protected]) is an associate professor of medicine at Harvard medical School and a primary care physician at Cambridge Hospital in Cambridge, Massachusetts. Elizabeth Warren is the Leo Gottlieb Professor of Law at Harvard Law School in Boston. She was chief adviser to the National Bankruptcy Review Commission. Deborah Thorne is an assistant professor in the Department of Sociology and Anthropology at Ohio University in Athens. Steffie Woolhandler is an associate professor of medicine at Harvard where she codirects the General Medicine Faculty Development Fellowship Program. She practices primary care internal medicine at Cambridge Hospital.