What Atul Gawande and the Dartmouth Atlas Completely Missed
Conventional research based on Medicare spending alone has led to the claim that some areas are high resource users because of the way doctors practice medicine. It is why so much emphasis is being placed on evidenced-based medicine and changing the way doctors practice. See Gawande and Dartmouth and our response to both here and here.
Yet a new study, consistent with NCPA research, finds that where Medicare spending is high, non-Medicare spending is often low.
The study examined private insurance claims … [in areas] throughout the country that spent the most and least on health care across demographics. The researchers concluded that their findings challenged other established research that traditionally has used Medicare data for the elderly to determine geographic spending variations for people of all ages.
[For example], McAllen, Texas — a city known for having the highest Medicare spending in the U.S. — ranked among the 10 lowest-cost cities for health care among individuals with employer-sponsored health insurance, the report found.
Excellent post. And much needed in the light of so much misleading rhetoric on the other side.
I agree with Paul. Thanks for posting this.
I find the research on regional variation in medical spending to be interesting — but not always convincing. For instance, Greg Scandlen and Linda Gorman have both pointed out instances where the conclusions were wrong. In Linda’s case, Grand Junction, Colorado is often used as an example of a highly efficient health care region. I don’t doubt that it is. But Colorado has a healthier population than other states and many seniors needing costly care will jump in their car and take I-70 east to Denver.
To add to what Devon’s saying: I’m not sure that it’s entirely possible to risk-adjust populations. There is a lot that we don’t understand about medical risk. Saying that one population is utilizing care more than would be expected for the same risk-adjusted population in another area is not necessarily a fair statement.
I also liked this. Very good counter to what we ordinarily read.
There didn’t seem to be any pattern at all to the variation in spending by age group. Maybe if we knew the proportion of the population that was in each age group by MSA we could make some sense of it.
For example, do MSAs that have an extraordinarily large proportion of seniors in the population also have higher than average Medicare costs per capita, and vice versa?
This could mean (at least) two things: Seniors flock to MSAs where health care is more accessible/available. Or, providers in MSAs thick with seniors exploit that situation by oversupplying care.
Reed Abelson and Gardner Harris wrote an article in the New York Times on June 2, 2010, which cited a few of the Dartmouth Atlas’ critics (http://tinyurl.com/2er96m8).
@Devon Herrick “Grand Junction, Colorado is often used as an example of a highly efficient health care region. I don’t doubt that it is.”
Devon, having personally worked as a physician in Grand Junction I can say that perhaps you should doubt it. There are several factors about G.J. other than the supposed efficiency of the healthcare delivery system that are relevant to health spending, none of which were addressed in the Dartmouth Atlas. Grand Junction is geographically isolated. The medical community is dominated by family practice physicians, a single not-for-profit insurance company, an active community-based hospice program and an overgrown community hospital that advertises as a “regional medical center”, despite several clinical specialty areas being notably underrepresented. Patients with more complex disease tend to leave Grand Junction. Complex patients are actively steered toward hospice care. Financial incentives for primary providers promote underutilization. I can cite multiple instances of patients who were not even offered the option of seeing a specialist for a complex condition, or where aggressive and potentially life-saving treatment was not even discussed with patients. The attitude sometimes seems like “Die sooner, save us all some money!”
Grand Junction residents tend to be stoics of northern European ancestry (e.g. farmers) who tend to accept their lot in life without whining. They also tend to be fiscally conservative, and will often choose to sacrifice their own needs for those of their children and grandchildren, not wanting to prolong their own lives if it will burden their family. The health of the population is generally better than in urban areas, possibly due to self-selection, i.e. complex patients don’t voluntarily migrate to Grand Junction.
The Dartmouth Atlas raises some interesting questions regarding Medicare spending in the last two years of life, as you point out, but it does not address the ethical basis of medical decision making within Medicare. And as John’s post points out, it does not address non-Medicare spending and its potential effect on total health spending.