Abstract: This paper explores the financial impact of medical problems, using data from Phase III of the Consumer Bankruptcy Project, a survey of 1,974 individual bankruptcy petitioners conducted during the first quarter of 1999 in eight federal judicial districts. Although the questionnaire covered a variety of topics, this paper focuses on debtors' identification of illness or injury as a reason for filing, medical debts, and health insurance coverage. One out of four debtors, or an estimated 326,441 families in 1999, identified an illness or injury as a reason for filing for bankruptcy. One third of the debtors said that they had substantial medical bills, i.e., that they had incurred $1,000 or more in medical bills not covered by insurance. Combining those identifying medical reasons with those indicating substantial medical debts (an overlapping but not perfectly coextensive group), the financial consequences of medical problems were a factor in the bankruptcy cases of an estimated 596,198 families in 1999. Health insurance coverage was sparse for the group, with one in five debtors reporting that they had no health insurance for any family member. The absence of insurance, however, did not correlate with a debtor's identifying a medical problem. Those who had insurance and those who did not were about equally distributed among those who identified a medical problem and those who did not. The data were re-analyzed by separating the responses of single filing men, single filing women, and joint filing married couples. Households without a male present were nearly twice as likely to file for bankruptcy giving a medical reason or identifying a substantial medical debt as households with a male present. The proportion of debtors providing medical reasons for filing also varied with the debtors' age. Of debtors 65 or older, 47.6 percent listed a medical reason, as compared with 7.5 percent of debtors under 25. Previous studies considering medical problems and bankruptcy in the United States are summarized, although the methods used and results obtained are not directly comparable with the current findings.