This is the last of a series of imaginary lectures for President Obama. I am hoping to educate him by criticizing three people who influenced him – Peter Orszag, Atul Gawande, and Elliott Fisher and his colleagues at Dartmouth. In this last installment I focus on Gawande.
Obama was deeply impressed by Gawande’s “The Cost Conundrum,” an article published in the New Yorker in June 2009. By June 2009 Obama had already adopted the managed care diagnosis (overuse) and the latest iteration of the managed care solution (ACOs, “medical homes,” and pay-for-performance, all of which will allegedly be facilitated by electronic medical records). “The Cost Conundrum” did not convert Obama to managed care ideology, but it did strengthen his belief in it.
“The Cost Conundrum” illustrates the good and the bad effects the Dartmouth Atlas has had on American health policy and on intelligent people like Gawande and Obama. The article is about Gawande’s trip to McAllen, Texas to see why per capita Medicare spending in that small town was the highest in the country.  Gawande knew it was high because the Dartmouth Atlas said so. Asking why Medicare spending in McAllen was so high was a legitimate question to ask.
But Gawande went way beyond exposing problems with Medicare spending in McAllen. He told his readers that the problems he uncovered in McAllen – overuse of some Medicare services induced by fee-for-service payment – afflicted vast swaths of the medical profession and that “accountable care organizations” were the answer. He specifically singled out the Mayo Clinic in Rochester, Minnesota and an informal cartel in Grand Junction, Colorado as examples of ACOs that had allegedly already proven they could provide high-quality care at very low cost. 
But within a few years, research would turn Gawande’s characterization of Mayo and Grand Junction upside down. It would turn out that both the Mayo Clinic and Grand Junction are costly places to be treated when all medical spending, not just Medicare spending, is taken into account. Oops.
But before I elaborate on that mistake, I want to give Gawande credit for the good “Conundrum” did do.
What was good about “The Cost Conundrum” was that it illustrated the capacity of the Dartmouth Atlas data, crude as it is, to generate hypotheses worth investigating. According to Gawande, Dartmouth Atlas data showed that per capita Medicare spending in McAllen was “almost exactly” equal to the national average in 1992, but had gone up so much by 2006 McAllen held first place in Medicare spending. “Conundrum” presented convincing evidence that something had gone wrong in McAllen between 1992 and 2006, and substantial overuse of medical care was the result.
Gawande’s article appears to have had a good effect. Between 2008 and 2012, Medicare spending in McAllen dropped from 40 percent above the Texas average to 16 percent, primarily due to a large reduction in the use of post-hospitalization care. (I am unaware of research that demonstrates the eliminated services were all unnecessary services, but it seems likely most were.) In a follow-up article about “an avalanche of unnecessary care” entitled “Overkill,” Gawande reported that many McAllen doctors attributed the drop in Medicare costs to the combination of simply learning they were ordering more services than other doctors plus the glare of the media. (Will Gawande ever write an article about the “avalanche of underuse”? Will it bear the title, “Underkill”?)
In Overtreated, Shannon Brownlee tells a similar story about doctors who didn’t know they were ordering services at rates much higher than colleagues elsewhere and who reacted appropriately when they saw the data. She reports that John Wennberg, the father of regional-variation research, visited five doctors in Morrisville, Vermont in the mid-1970s to show them data indicating they were doing tonsillectomies at a rate substantially above the Vermont average. Brownlee reports those doctors responded by reducing the Morrisville tonsillectomy rate by two-thirds within five years (p. 27). (Will Brownlee ever write a book entitled Undertreated?)
The McAllen and Morrisville anecdotes illustrate how the appropriate use of crude regional-variation data (sharing it with doctors, as opposed to inflicting suffocating P4P schemes like MACRA on all doctors) can lead to good outcomes.
But Gawande did not see the Dartmouth Atlas as a mere generator of hypotheses. He saw it, rather, as the great arbiter of high- and low-cost regions and hospitals. “Medicare expenditures [are] our best approximation of overall spending patterns,” he wrote in “Conundrum.” If Elliott Fisher concluded, on the basis of Atlas data, that the Mayo Clinic was a shining example of an ACO that provides superior care at a very low price, that was good enough for Gawande. He would stake his credibility on it. “[T]he Mayo Clinic … is among the highest-quality, lowest-cost health-care systems in the country,” he wrote in “Conundrum.”
That was a mistake. Within a year, researchers began to inquire whether McAllen really was as expensive, and Mayo and Grand Junction really were as inexpensive, as Gawande said they were. That research would show that McAllen was nowhere near as expensive as Gawande said it was, and that the Mayo Clinic and Grand Junction were very expensive. Because Gawande exalted Mayo more than any other clinic or system, I’ll focus on the research that destroyed the myth that the Mayo Clinic is unusually efficient.
The ascendance of the Mayo myth
By the early 1990s managed care advocates were constantly repeating the myth that Mayo Clinic provides medical care at very low costs. (I am calling the claim that Mayo is unusually inexpensive a myth. I am not questioning Mayo’s reputation for excellent medical care.) Hillary Clinton mentioned the Mayo Clinic and Minnesota so often in her 1993 testimony before the Senate Finance Committee that Senator Dave Durenberger (R-MN), a member of the committee, thanked her for “mentioning Minnesota with some frequency.” (p. 32) As we saw in my second post in this series, Clinton could not explain to the committee what Mayo was doing to stop all the alleged overuse, but she was sure Mayo was doing something right because Medicare data assembled by John Wennberg at Dartmouth told her so.
The Mayo myth was solidified in 2008 by a study by Wennberg et al. which claimed that the Mayo Clinic’s St. Mary’s hospital in Rochester was half as expensive as Cedars-Sinai in LA and NYU Medical Center.  This study suffered from the three defects I discussed in post number three in this series, namely, it was limited to Medicare expenditures, it used the “they all died” method of adjusting for differences in patient health, and it assumed all spending on a particular patient can be attributed to one hospital. Nevertheless, the 2008 Dartmouth study set off another wave of hype about Mayo’s costs. 
The destruction of the Mayo myth
But just as 2009-2010 were the peak years of the credibility of the Dartmouth Atlas, so 2009-2010 were the peak years of the Mayo myth. By 2009 the criticism that the Dartmouth Atlas failed to take account of factors outside hospital control (primarily patient health status) was becoming more common, and that took a toll on the Mayo myth because what little evidence there was for the myth came from the Atlas.
The Washington Post was among the first to poke a stick at the myth. In a 2009 article , the Post noted that Mayo Clinic patients “are wealthier, healthier and less racially diverse than those elsewhere in the country,” in part because the local population is healthier, and in part because Mayo cherry-picks its Medicare and Medicaid patients.  But it was the release of data on prices charged by Minnesota providers in 2011 and the opening of MNSure, Minnesota’s Obamacare exchange, in October 2013 that finished off the Mayo myth.
In 2011 an organization set up by Minnesota insurance companies called Minnesota Community Measurement released data on prices paid by the four largest Minnesota insurance companies to Minnesota hospitals. Using price data for 69 of the procedures listed in the report, the American Experiment, a conservative think tank, concluded : “Mayo costs far more than other Minnesota providers. Of the 69 procedures, Mayo’s price is the highest for 11 and among the top five highest for 48. On average, Mayo’s price was 220 percent higher than HealthEast [a relatively low-cost system] and 180 percent higher than Park Nicollet [a relatively high-cost system].”
In October 2013, Obamacare exchanges, including MNSure, opened for business across the country. In an October 30, 2013 article , the Minneapolis Star Tribune reported that MNSure shoppers in the Rochester area were facing very high premiums and had no choice of insurer, thanks to “the Mayo effect.” “The lack of choice is particularly stark for residents with a Rochester ZIP code,” the article reported. “There is just one health plan offered on MNsure – a single silver-level plan from Blue Cross and Blue Shield of Minnesota. It costs a 40-year old nonsmoker $326 a month, with a $3,000 deductible. A comparable option for a Twin Cities resident would cost an average $154.” The article went on to quote an exasperated Rochester broker saying, “No other carrier wants to work in this area partly because the Mayo Clinic is becoming too expensive.”
Keep in mind that Minnesota has one of the most expensive health care systems in the country.
The Star Tribune confirmed Mayo’s high price a year later with data from the Minnesota Hospital Association. The national-level studies I discussed in installment three in this series that debunked the Atlas data were nails in the coffin of the Mayo myth.
To paraphrase Gawande’s comment about McAllen (see footnote 1), the Mayo Clinic appears to be the most expensive provider of medical care in one of the most expensive states in the US, the nation with highest per capita health care costs in the world.
“Accountability” buffs should be held accountable
Once the Mayo myth was refuted, Gawande had a choice: He could either notify his New Yorker readers he was wrong to call Mayo an example of an ACO, or he could notify them he was wrong to claim that ACOs cut costs. He has done neither. He has simply gone quiet about Mayo’s costs. Nor has he stopped promoting ACOs and gigantic hospital-clinic chains. He even has the chutzpah to tell us the Cheesecake Factory should be the model for the hospital-clinic cartels that have formed over the last three decades.
In this six-part series I have severely criticized President Obama and three of his advisors for serious defects in their thinking about how to reduce US health care inflation. All four men – Obama, Fisher, Orszag and Gawande – illustrate what is wrong with the culture of the managed care movement. That culture celebrates a cavalier attitude toward evidence.
How does Obama get away with telling us in August 2016 the ACA deserves credit for the 2008-2013 inflation lull and not telling us the lull is over? How does Elliot Fisher get away with telling us that all human beings become equally sick once we enter the last six or 24 months of life and therefore hospital expenditure data does not need to be adjusted to reflect differences in patient health? How could Peter Orszag oversee research at the CBO showing that ACOs and other managed care fads won’t save money, and then tell Obama and Congress they would? How does Gawande get away with promoting the Mayo myth, and when the myth is destroyed, failing to issue a retraction?
There is a double standard at work within the US health policy establishment. That “standard” insists, on the one hand, that doctors and hospital staff should be “accountable” to society via ACOs, “medical homes,” and P4P schemes, but on the other hand those who promote ACOs etc. have no accountability to anyone. Because of this double standard, those who preach accountability for providers are free to promote the overuse myth in the face of evidence that underuse is far more common; they are free to promote solutions to the alleged overuse that have little or no evidence to support them; and, when evidence arises contradicting their diagnoses and prescriptions, they are allowed to pretend they didn’t commit a serious error.
This has to change. President Obama is both a victim and a promoter of this double standard. He placed his trust in advisors who subscribed to the double standard and who told him the managed care fads contained in the ACA would cut costs. I urge Obama to acknowledge those facts, and to use his influence after he leaves office to demolish the double standard as well as the myth that the ACA is “bending the cost curve.” At that point, we as a country might at long last have an honest debate about how to reduce the enormous waste in our health care system and thereby make universal health insurance affordable.
 Gawande said in “Conundrum” that McAllen was both the most expensive and the second-most expensive city in America. In the second paragraph, he wrote: “McAllen … is one of the most expensive health-care markets in the country. Only Miami – which has much higher labor and living costs – spends more per person on health care.” But two paragraphs later he wrote, “McAllen, Texas, [is] the most expensive town in the most expensive country for health care in the world….” A few paragraphs after that, in describing a dinner he had with six McAllen doctors, he said, “Some were dubious when I told them that McAllen was the country’s most expensive place for health care.”
 Gawande repeated his praise of the Mayo Clinic and Grand Junction in a follow-up article http://www.newyorker.com/news/news-desk/atul-gawande-the-cost-conundrum-redux in the New Yorker three weeks later.
 You can see data from the 2008 Dartmouth study for Mayo’s St. Mary’s hospital and four other hospitals midway through this comment by Maggie Mahar, an ardent fan of Gawande and the Dartmouth Atlas.
 For example, two days after Wennberg et al. released their data the New York Times editorial board, a leading managed care proponent, declared, “Medicare could save tens of billions of dollars annually – without reducing the quality of care – if all hospitals mirrored the practice patterns of the Mayo Clinic.”
 Here is how the Washington Post described Mayo’s cherry-picking strategy: “Its Rochester flagship accepts Medicare patients from outside Minnesota only if they are willing to pay a personal premium beyond normal Medicare coverage, a practice that screens out some who cannot afford to pay more. Even in Rochester, a city of 85,000, Mayo serves a higher-income echelon than the town’s other hospital, Olmsted Medical Center. Just 5 percent of Mayo’s hospital patients receive Medicaid, an exceptionally low figure, compared with 29 percent at Olmsted….”