Wednesday, June 9, 2010

Why The Dartmouth Flap Matters For Obamacare

Cross-posted from National Review Online.

The liberal end of the blogosphere is aflame with criticism of press coverage of a health-care database at Dartmouth College. While the debate may seem arcane, it is worth a closer look. For it relates to one of the key justifications for Obamacare: that the rising cost of health insurance is largely due to the greed and incompetence of ordinary physicians, and that government alone can fix this problem.

First, some background. Over twenty years ago, Dr. Jack Wennberg of Dartmouth Medical School started tracking regional variations in Medicare expenditures; his diligent compilations came to be known as the Dartmouth Atlas of Health Care. The group that tends this database has gained broad academic respect over the years. In 1997, Wennberg’s son David helped to found Health Dialog, a consulting firm based on the Dartmouth work; in 2007, it was sold to British insurer Bupa for $775 million.

It wasn’t until 2009, however, that the Dartmouth Atlas became truly famous. In June of that year, Atul Gawande of The New Yorker published an article, “The Cost Conundrum,” describing medical practices in McAllen, Texas, an area that the Dartmouth Atlas identifies as one of the most expensive health-care markets in America. “In 2006,” observed Gawande, “Medicare spent fifteen thousand dollars per enrollee here, almost twice the national average.”

But, wrote Gawande, there was no discernible justification for McAllen’s high costs. In El Paso, a town that is in many ways similar to McAllen, Medicare spent only half of what it was spending in McAllen: $7,504 per enrollee. Importantly, the Dartmouth data suggested that McAllen’s increased medical utilization wasn’t leading to better results: Indeed, based on Medicare’s rankings, McAllen’s hospitals were performing worse than El Paso’s. The reasons, it appeared, were ignorance and avarice. Many doctors were simply unaware of how their practices deviated from medical convention. In addition, Medicare mostly pays doctors and hospitals on a fee-for-service basis; therefore, health-care providers have incentives to order more tests and conduct more operations, so they can collect more fees.

Gawande’s piece gained enormous attention. David Brooks called it “the most influential essay of 2009.” It became required reading in the White House. Recounted Sen. Ron Wyden (D., Ore.), “[The President], in effect, took that article and put it in front of a big group of senators and said, ‘This is what we’ve got to fix.’” Peter Orszag, the White House Budget Director, used the Dartmouth data to argue that health-care expenditures could be reduced by $700 billion—over 30 percent—without sacrificing quality. The Dartmouth data allowed Democrats to imagine a glorious possibility: that the increased expense of universal health insurance could be paid for by intimately regulating medical practice in a way that wouldn’t harm, and might even improve, the quality of care.

Fast-forward to last week, when Reed Abelson and Gardiner Harris published a front-page article in the New York Times that was highly critical of the Dartmouth work. “Dartmouth’s claims about which hospitals and regions are cheapest may be suspect,” they wrote. “The real difference in costs between, say, Houston and Bismarck, N.D., may result less from how doctors work and live. Houstonians may simply be sicker and poorer than their Bismarck counterparts.” Abelson and Harris wrote that, among other things, the Dartmouth Atlas didn’t take local cost of living into account when making its measurements, which meant that expensive cities like New York might suffer if the government tried to correct for variations in health-care spending.

The article provoked a fierce rebuttal from Jonathan Skinner and Elliott Fisher, the Dartmouth researchers who now maintain the Atlas. They correctly pointed out that the Atlas does remove cost-of-living considerations from the equation by measuring utilization (i.e., the number of tests and procedures) as well as dollar expenditures. Maggie Mahar of the Century Foundation, an excellent liberal health-care blogger, tracked down the other academics quoted in the Times article, who told her that the Times had misrepresented their views. “It sounds as if it were written by someone’s ex-spouse,” one of her sources told her. “Harris and Abelson were determined to write a story that would ‘take down Dartmouth.’”

Technically, Skinner and Fisher are right: The Times article contains factual and analytical errors. But there are numerous other critiques of the Dartmouth methodology that have not been fully resolved. Rates of chronic disease and poverty are not well accounted for in the Dartmouth statistics, though the Dartmouth team is doing more in this realm than before. Another problem is that the Dartmouth work tries to separate out health spending by medical diagnosis, but does not take into account the severity of illness or the incidence of multiple diseases in the same patient. Another fault is that the study fails to take into account the impact of racial and economic diversity within a given area. Yet another is that the Atlas measures Medicare fee-for-service spending alone, and does not study health-care spending for individuals with private insurance, or with Medicare HMOs. In addition, the data do not take into account regulatory or legal variations among the states, such as differences in malpractice law.

The list of disputations goes on, but you get the point. The Dartmouth group, to its credit, does its best to produce a thoughtful, empirical analysis of variations in health-care spending. And it is of course true that some doctors are unaware of best practices, and that others try to game the system. But it is one thing to catalogue unusual spending, and another to decide which portions of that spending are unnecessary, and quite another altogether to fashion federal regulations that will eliminate the spending that has been deemed unnecessary. There are several fallacies in these great leaps of logic that led the Obama administration to the conclusion that the Dartmouth Atlas can teach the federal government how to micromanage the practice of medicine.

First off, as we can learn from the very data compiled in the Atlas, a number of institutions—such as the Mayo Clinic in Minnesota and the Intermountain system in Utah and Idaho—have succeeded in combining high quality and low cost, without federal intervention. As the Intermountain story shows, these improvements require painstaking focus and, above all, personal relationships with participating physicians. Much more can be done by the private sector to improve the quality of health care, in particular by adopting electronic health records. But there is little evidence that the federal government has the competence to achieve such improvements by bureaucratic fiat.

Second, the government is largely responsible for the original problem. The biggest distorter of physician behavior is what the Dartmouth group delicately describes as “the current reimbursement system”—that is, Medicare. Medicare is what gives physicians incentives to spend as much as they can on patients, and it’s not simple greed: It’s about erring on the side of marshaling all available resources for your patient. In the bad old days, when more people paid for their expenses out of their own pockets, without insurance, doctors were much more conscious of keeping their patients’ financial health in mind. Now that retirees pay little to nothing for Lamborghini health care, their doctors are liberated from having to think about cost.

Third, it isn’t clear that a single-payer system would eliminate variations in health-care spending. As Cato’s Michael Cannon has pointed out, the Congressional Budget Office has looked at Veterans Administration data and found that the VA health-care system contains almost as much variation as the Medicare system, despite the government’s complete control.

Fourth, Orszag is wildly exaggerating when he claims that 30 percent of health-care spending could be saved by using the Dartmouth Atlas to eliminate all unnecessary expenditures. It would be much better for health outcomes, but just as absurd, to seek to cut costs by eliminating all obesity from the country. Eliminating the entirety of wasteful health-care spending would be like eliminating, via federal mandate, every messy desk in America. That is to say, doubtful. And it is not metaphysically possible to eliminate all waste. Quite a bit of what we call “waste” is actually uncertainty: performing heart surgery, say, on someone who may or may not require it. As in war, sports, and courtship, the best medical decision is often more obvious in hindsight than it is in the heat of the moment.

Fifth, while it is a good thing for physicians to adopt the best evidence-based guidelines for clinical practice, every human being is different. There are exceptions to every protocol, medical cases that defy conventional logic. America in particular is a genetically diverse nation; treatments that work in some populations may not work in others. Treatments that work in some individuals don’t work in others within the same population. A one-size-fits-all approach to the practice of medicine is especially appealing to those who have never done it, but it would in reality cause innumerable headaches and heartaches.

And so, the debate about the ins and outs of the Dartmouth Atlas is not merely a statistical one. It is about something more fundamental: Can government do a better job of managing medicine than doctors and hospitals can, or should the doctor-patient relationship remain sovereign? There is plenty of waste in medical care today, but the Dartmouth Atlas demonstrates that government is the problem, not the solution.


  1. Medicare is also a primary payer at the Mayo, yet they do not overutilize.

    " Quite a bit of what we call “waste” is actually uncertainty: performing heart surgery, say, on someone who may or may not require it. As in war, sports, and courtship, the best medical decision is often more obvious in hindsight than it is in the heat of the moment."

    I am a physician. Much of what you write here makes no sense, except at the extreme margins. If you do not set guidelines ahead of time, you can justify virtually any procedure at the time it is proposed. Do you know any doctors?


  2. Hi Steve, thanks for your comment.

    I'm not sure where you are disagreeing with me -- I write in the piece that "it is a good thing for physicians to adopt the best evidence-based guidelines for clinical practice." I cite Mayo as an example of a place where, in spite of Medicare's disincentives, low-cost, high-quality care has been achieved (without government mandates).

    I'm not sure guidelines are the cure-all you think they are, though. Take the mammogram controversy: clearly, it's in the interests of women to get screened at age 40. However, some cost-oriented policy types suggest that women should only begin screening at age 50. Which should it be?

    Imatinib, a drug for chronic myelogenous leukemia, achieves complete cytogenetic response in 65-70% of patients. Second-generation drugs nilotinib and dasatanib achieve complete response in 80-85% of patients. Imatinib is going generic in 2015. When that happens, should patients be given the generic drug, or the branded drugs, which are clearly better, but will cost more than $100,000? Remember that this isn't hypertension -- you only get one shot. Again, patient care would suggest going with one choice; cost-control would suggest another.

    There may be situations where there is a concordance of cost-effectiveness and best medical practice. There are plenty of others where there is not.

    As to your biographical question: I come from a family of physicians; I myself went to medical school (at Yale); I deal with physicians and treatment protocols every day in my line of work. I attend around 20 scientific conferences a year; this month alone I am attending the annual scientific meetings of the American Society of Clinical Oncology, the European League Against Rheumatism, and the American Diabetes Association.

    I certainly hope to learn from you when I make mistakes -- but I also hope you would agree that I have a legitimate vantage point from which to comment on these issues!

    Thanks again,