Or Lubbock? Or Oklahoma City? Or New Orleans? Or any of a dozen major and minor metro areas throughout the South? According to the Medicare Payments Advisory Commission, all of them have significantly higher usage rates and costs per Medicare enrollee than McAllen, which was high-cost locale ground zero for Atul Gawande’s famous New Yorker article, “The Cost Conundrum,” which has become, to use the New York Times‘ formulation, “must reading in the White House.”
Gawande grounded his analysis on per-patient Medicare claims data compiled annually by researchers at Dartmouth Medical School. “The explosive trend in medical costs seems to have occurred here in an especially intense form,” Gawande wrote after the Dartmouth Atlas of Health showed McAllen as the highest spending region in the country outside Miami, where Medicare fraud is an especially virulent problem. Not so, MedPAC said. Adjust for prices and McAllen’s outlier status compared to the rest of Texas and large parts of the South all but disappears.
Don’t believe me? Check out the appendix to the report. Medicare service utilization rates compared to the national average: McAllen – 118%; Houston – 122%; Dallas – 117%. Other areas: Oklahoma City – 120%; New Orleans – 125%; Pascagoula, Gulfport or Biloxi, Miss. – 124%.
In other words, Texas and large parts of the South have a problem — not just McAllen. I’ll return to that in a moment.
Why am I writing about this issue today, since I wrote about it last December when the MedPAC report came out? Gawande will be the featured speaker at the National Governors Association meeting that convenes in Washington tomorrow morning. It appears that the Harvard surgeon has not only become must reading in the White House, but is also something of a rock star on the public policy circuit. If he stays on script, he’ll tell the governors that 30 percent of health care costs could be wrung out of the system if all the McAllens were magically transformed to be more like Iowa and Minnesota, where health care costs average 80 to 85 percent of the national average.
Unfortunately, Gawande still hasn’t publicly recognized the flaws in his presentation, at least according to his reaction yesterday on the New Yorker website, where he responded to a brief New York Times report this week questioning the Darmouth analysis. (Ironically, his spokeswoman passed along a “refused to comment” to Times reporter Gardiner Harris. GoozNews asks with Yiddishkeit inflection: “Reporters have spokeswomen?”)
The Times story was based on a commentary in the New England Journal of Medicine by Peter Bach of Sloan Kettering Memorial Cancer Center (a frequent source for my own writing, I should reveal), who argued adjusting the data for illness severity and outcomes — a measure of quality and efficiency — eliminates most of the differentials, at least at the hospital-provider network level.
A rejoinder by Dartmouth’s Jonathan Skinner, Douglas Staiger and Elliot Fisher rejected that conclusion. They said the cost differentials remain quite large even after quality and illness adjustments. But they do include an admission that regional variation is less significant than some interpreters (i.e. Gawande) claim:
The implication of these results is that excessive health care costs arise at the level of the hospital-provider network. Thus, incentives that are designed to reduce costs should be targeted to specific networks, rather than regions or states.
Now let’s return to what Gawande wrote yesterday in response to the Times, and presumably after reading the dueling commentaries:
The patterns of Medicare spending I found showed that McAllen’s medical community and culture simply did more-more surgery, more imaging, more specialist visits, more home-nursing visits-without clear benefit. . .There remains fierce disagreement about how much of the marked differences in spending between communities is the result of health differences between populations and how much is the result of differences in the care their clinicians provide to them. But it remains clear that there are substantial variations in the cost of care for people of similar health depending on which institutions they go to. . . Even if health reform disappears, these fundamental problems will not. The cost conundrum persists.
Note how he switches horses midstream. He starts by defending his work on regions, and concludes by claiming the problem is individual institutions.
Now back to the South. The most recent Medicare cost data revealed an interesting national phenomenon during this recession. Medicare costs soared while private health expenditures stagnated. As I wrote at the time, the most logical explanation for this trend is that local medical institutions made up for charity care given the newly unemployed and uninsured by soaking their best-paying customers: those on public programs like Medicare and Medicaid.
We’re seeing the same thing happen in the individual insurance market, where soaring rates are making headlines. In economically hard-hit California, Detroit and other troubled labor markets, the insurance companies are blaming rising rates on the declining number of workers buying policies. There’s less healthy people paying for the care of those who get sick.
But this is a phenomenon that doesn’t necessarily depend on recessionary levels of unemployment. The South, where there’s no legacy of unionization, decent wages or decent benefits, has the lowest rates of insurance coverage in the nation, with Texas the worst state in that regard. Less than three-quarters of its population has health insurance.
The Dartmouth Atlas of Health looks only at Medicare claims. Isn’t it possible that the huge outlier status for most Texas cities is, in effect, a major cost shifting from the uninsured, who still show up on institutions’ doorsteps for care, to Medicare and Medicaid, which at least pay their bills?
But wait, you say. Even utilization rates are substantially higher in Texas. Agreed. But here’s where I leave data and logic behind and resort to anecdote. In my mother’s last years (in Maryland), she was incapacitated in a nursing home where the doctor would show up about once a week to make his rounds. He would stop in every room for a few minutes, check the chart and move along. He could usually cover a dozen rooms in about an hour. Since most of the patients were in the final stages of severe dementia, there wasn’t much to discuss. Sometimes there was a family member in the room. Those visits took longer.
My point is that utilization can be an imprecise metric. Hospitals and physician offices do what they do and bill whom they can. Overutilization is endemic throughout the health care system, regardless of region. Some institutions — those that are primarily fee-for-service medicine — are more prone to overutilization compared to those that operate with salaried physicians and have incentives to hold down costs. The Dartmouth Atlas of Health’s insights into that phenomenon are spot-on and enduring. But the differences it highlights can be found anywhere.
No doubt some regions have more institutions and physician practices that engage in inefficient and poor quality care. But when it comes to gross outliers — whether Miami, McAllen, Houston or New Orleans — I suspect the differences have more to do with exogenous factors like fraud or the uninsured than a local culture of medicine that encourages overuse. That, my reporting tells me, exists everywhere.