Friday, February 04, 2005

Medical Bankruptcy

Fox IslandConsiderable media attention has been paid to a recent study demonstrating a strong link between bankruptcy and medical illness. When I first read of the study in the Wall Street Journal, I must confess I was skeptical. Studies which receive large media attention are sometimes driven by an underlying agenda, and in some instances, have used questionable study populations and statistical analysis to produce a result less scientific than political in nature. One such controversial study was the Institute of Medicine (IOM) Medical Error Report of estimated morbidity and mortality from medical mistakes. The study has been widely criticized for its overly broad definition of a medical error, retrospective analysis of data over 10 years old, study populations which may not be representative, and the statistical sleight-of-hand of multiplying the findings in a small population nationwide produce a staggering estimate of deaths and injuries due to medical mistakes. Nevertheless, the study has fully entered the public consciousness, and is often quoted as a solid fact in health policy discussions.

The bankruptcy article, published in Health Affairs, initially raised some red flags. Its chief author, David Himmelstein, is a health care activist and strong proponent of nationalized healthcare -- a fact not mentioned in the media reports. While this is an entirely legitimate philosophy to promote, it is also true that such a bias might produce a study strongly inclined to reach conclusions supporting the authors preconceived philosophy.

However, after studying the article, I was pleased to find a thorough and detailed methodology which suggests a strong effort to obtain useful data as little influenced by bias as possible. Nevertheless, the big media splash suggests the authors conclusions were in line with their philosophical bent. Yet reading the study, such conclusions seem strained.

The authors reviewed nearly 1800 personal bankruptcy files from five federal courts (geographic locations not specified), and followed up approximately half of these with detailed, in-person interviews. The demographics showed bankruptcy filers to be predominantly middle or working-class, with women slightly outnumbering men. Approximately one quarter of bankruptcy filers cited illness or injury as the specific reason for bankruptcy, and a similar number cited medical expenses as a contributing factor, with uncovered medical bills in excess of $1,000. However, somewhat surprisingly, medical debtors were no less likely than other debtors to have coverage at the time of bankruptcy filing.

The media headlines trumpeted one conclusion of the study: that 55% of bankruptcies were related to medical illness or expenses. The inference in several of the newspaper articles was that lack of health insurance or high medical bills was the cause of over half of all bankruptcies. Reviewing the study, however, such a startling conclusion does not appear warranted. First of all, the 55% figure is a fruit salad of risk factors for bankruptcy. The study groups together medical bills in excess of $1,000 with major medical illness or injury resulting in loss of work time and work-related income. The more accurate conclusion, to my reading, is that illness or injury resulting in loss of job or loss of income is significantly more important, although often added to, unpaid medical bills. This conclusion is reinforced by the fact that many of those filing for bankruptcy were not without health insurance coverage. In many cases, the illness resulted in loss of employment, and subsequent loss of health insurance.

Somewhat surprisingly, few medical debtors studied had elected to go without insurance coverage. Furthermore, debtors out-of-pocket medical expenses were often well below catastrophic levels. In the year prior to bankruptcy, the average out-of-pocket medical expenses was $3,700, and the average out-of-pocket costs since onset of illness was less than $12,000.

A far more significant risk is the loss of work and income due to illness or injury. The intimate connection between employment and health insurance doubles the bet, as prolonged illness means lost income, lost jobs, and lost health insurance.

The medical dollar amounts -- annual medical expenses over $1,000, multi-year expenses over $10,000 -- certainly represent a significant burden to low or middle-income families. Yet to a far greater degree they reflect the extreme financial brittleness of those who end up in bankruptcy. The numbers cited are far less than many families spend on an automobile, or on car repairs, or to fix a leaking roof. How many of these families were living beyond their means, overextended by credit card debt or large mortgage payments, is not clear. How many women were abandoned by husbands with children to raise, and little or no income to support them? Such issues are social in nature, reflecting the breakdown of marriage commitment, the indirect effects on families from an epidemic of drug and alcohol abuse, public policies discouraging savings, individual lack of financial discipline, and a potential host of other social pathologies unrelated to health care costs. Indeed, the authors themselves admit that even universal health care coverage would not have prevented many of the bankruptcies. Medical expenses are certainly one factor, and medical catastrophy can and does result in financial devastation to some, but health care costs alone do not shoulder the responsibility of over half of all bankruptcies.