Saturday, December 30, 2006
Medical Bankruptcy: Myth vs Fact
Recently, medical bloggers like NHS Doctor have referenced Himmelstein's classic study which claims 54.5% of all bankruptcies filed in 2001 were medically related. However, as happens all to often, people fail to critically analyze published articles that support a particular preconceived viewpoint. Researchers David Dranove and Michael Millenson have published an excellent analysis of Himmelstein's data to outline the disappointingly poor quality of this study and its conclusions.
Abstract:
Abstract:
David Himmelstein and colleagues recently contended that medical problems contribute to 54.5 percent of personal bankruptcies and threaten the solvency of solidly middle-class Americans. They propose comprehensive national health insurance as a solution. A reexamination of their data suggests that medical bills are a contributing factor in just 17 percent of personal bankruptcies and that those affected tend to have incomes closer to poverty level than to middle class. Moreover, for national health insurance to have an impact, it would have to define “medical” expenses in a much broader way than is now typical of either private or government-funded plans.
Posted by JoshMD |
link
| Comments (4)
| 
Comments:
And while we are at it we can have universal home mortgage coverage, universal gambling coverage, universal disability, life, and accident insurance. Go ahead, chain smoke or stuff those fries in your face while riding that hog without a helmet. And let's fund it with taxpayer dollars.
Himmelstein's is a ludicrous argument.
Himmelstein's is a ludicrous argument.
# posted by 
: 7:33 PM
: 7:33 PM
When will we expose the lie that lawsuits are bankrupting doctors all over the country regularly?
# posted by : 9:05 AM
: 9:05 AM
Himmelstein does appear to go too far, but so do Dranove and Millenson.
In particular, the inclusion of families in the "medical bankruptcy" total who simply reported significant medical expense but who did not cite medical expense as a reason for bankruptcy is clearly a stretch. But is also a stretch to discount such families entirely.
Indeed, it is also possible that there are cases where the cost of medical care is a contributing factor to bankruptcy even when not one penny was spent on medical care! For example, suppose you have someone who is clinicly depressed but who, because of cost issues, fails to receive proper diagnosis and treatment. They lose their job as as result of their depression and are forced into bankruptcy. They may not even be aware of their situation and hence aren't going to cite a medical cause.
Dranove and Millenson also make much of the distinction between causation and contribution. However, this ignores the fact that bankruptcy is often the end result of what engineers call a cascade failure: Many factors had to align to produce that result. In such situations it is pointless to try and determine the definitive causative factor.
But there's an even bigger problem in this study, which is that financial disasters don't always result in bankruptcy, for a variety of reasons: It may not make financial sense, people may be unaware of this option, or they may be afraid of the social stigmata that's still associated with it in some circles. So what we're looking at may be no indication of the degree of hardship medical costs actually impose.
The bottom line is that this study fails to answer the real questions people want answered: If my family
suffers some sort of expensive medical crisis, statistically what will it do to our standard of living? And if there is significant risk here, what changes to the health care system would reduce that risk?
It would very interesting to see the results of a study constructed to answer these questions. However, the only approach I see would be to look at the outcomes for a random sample of families who have had such a crisis. Given medical privacy issues that may be hard for an independent research group to assess.
Post a Comment
In particular, the inclusion of families in the "medical bankruptcy" total who simply reported significant medical expense but who did not cite medical expense as a reason for bankruptcy is clearly a stretch. But is also a stretch to discount such families entirely.
Indeed, it is also possible that there are cases where the cost of medical care is a contributing factor to bankruptcy even when not one penny was spent on medical care! For example, suppose you have someone who is clinicly depressed but who, because of cost issues, fails to receive proper diagnosis and treatment. They lose their job as as result of their depression and are forced into bankruptcy. They may not even be aware of their situation and hence aren't going to cite a medical cause.
Dranove and Millenson also make much of the distinction between causation and contribution. However, this ignores the fact that bankruptcy is often the end result of what engineers call a cascade failure: Many factors had to align to produce that result. In such situations it is pointless to try and determine the definitive causative factor.
But there's an even bigger problem in this study, which is that financial disasters don't always result in bankruptcy, for a variety of reasons: It may not make financial sense, people may be unaware of this option, or they may be afraid of the social stigmata that's still associated with it in some circles. So what we're looking at may be no indication of the degree of hardship medical costs actually impose.
The bottom line is that this study fails to answer the real questions people want answered: If my family
suffers some sort of expensive medical crisis, statistically what will it do to our standard of living? And if there is significant risk here, what changes to the health care system would reduce that risk?
It would very interesting to see the results of a study constructed to answer these questions. However, the only approach I see would be to look at the outcomes for a random sample of families who have had such a crisis. Given medical privacy issues that may be hard for an independent research group to assess.
# posted by : 5:01 PM
: 5:01 PM
