• White vs. Skinner on the uses & misuses of the Dartmouth Atlas

    The following is a guest post by Harold Pollack, the Helen Ross Professor of Social Service Administration at the University of Chicago.

    Pardon the self-promoting press release. Incidental Economists readers may be especially interested in a new section I edit for the Journal of Health Politics, Policy, and Law.

    The feature is called “Point-Counterpoint.” Right out of the gate, political scientist Joseph White hurls a broadside against the Dartmouth Atlas. Dartmouth’s Jon Skinner, a major investigator in the Atlas efforts, respectfully disagrees. In part this is a specific debate about the Dartmouth Atlas. In part, this is a more fundamental debate about regulating prices vs. regulating utilization as paths to improving the quality and efficiency of our health care system.

    To give a sense of the emphatic tone, White starts as follows:

    • The most common claims about the extent of health care system costs that are caused by unnecessary services, both in and based on the Dartmouth scholarship, appear excessive.
    • The Dartmouth scholars have gone out of their way to claim their work shows that volume is the answer; implicitly and sometimes explicitly they argue that paying attention to prices is wrongheaded. This part of their argument is not the strongest analysis within their own work and sometimes is based on only a partial view of other relevant literature.
    • The emphasis on volume in the variations literature is contradicted by other evidence, which is at least as relevant to cost-control choices. The variations literature is widely interpreted as showing that the problem is paying fee-for-service. There is no ideal way to pay for medical care, but in the United States the biggest problem with fee-for-service is the specific fees per service. 
    • The emphasis on reducing unnecessary volume rather than lowering prices has a further flaw: we know very little about how to do the former and a lot about how to do the latter. Many of the theories generated by the Dartmouth scholars or by people interpreting their work are either not supported by evidence or face severe obstacles to practical application.
    • For all these reasons, the Dartmouth scholars’ overemphasis on volume has probably done more harm than good to the cause of better controlling health care costs in the United States. This is, simply, a shame. The work makes an important contribution to our understanding of medical economics. It both suggests some of the concerns that should be targeted in price regulation and provides ideas about what to do in addition to taking the right steps on prices. The work’s misuse for policy analysis is unnecessary.

    Jon Skinner sees things differently:

    Since the Dartmouth Atlas research was often used by then Office of Management and Budget Director Peter Orszag and others in support of the Obama administration’s health care reform efforts, it is not surprising that opponents of health care reform should have criticized it and questioned its validity.

    The combative commentary by Joseph White in this volume takes aim at Dartmouth researchers from another angle: their role in one of the greatest wrong turns ever in health policy. In his view, the Dartmouth research turned policy makers away from the more popular and sensible approach to saving money — cutting prices — and toward the chimerical and unpopular quest to cut back on quantities in order to save money.

    This narrative certainly has its high points — for example, the idea that academic researchers could lead seasoned policy makers, sheep-like, down the primrose path. Even so, there are several points with which I take exception. First, the author argues that the Dartmouth Atlas overstates the extent of variation and the degree of inefficiency in the U.S. health care system, relying in part on a now-outmoded study by the Medicare Payment Advisory Commission (MedPAC). Yet even the most conservative estimates imply waste in U.S. health care comprising 2.5 percent of gross domestic product (GDP)….. Several recent independent studies have each reached the conclusion that the degree of inefficiency in the U.S. health care system is 30 percent or more.

    Second, White’s commentary highlights some of the confusion around the role of prices and quantities in health care…..

    Readers can make their own judgments about who came out ahead. I appreciate the extensive efforts by both authors. I plan to hold a White House beer summit to help them work things out.

    Here are the links: Joe White, Jon Skinner, and my own introduction.

    • If you send me $15 I might consider reading these. Stupid academic journals…

    • Price x Volume = Total Cost
      Why can’t we simply say both price and volume are important, and both White and Skinner are right?

    • Without detracting from the fascinating work the Dartmouth folks publish–and of which I am a great devotee, and as much as While may take an aggressive tone (I am assuming based on limited cuts from above), he does have his points.

      From many of the publications I am reviewing nowadays, yes, prices a majority piece of the cost puzzle; and state, region, hospital, and physician variation is so nuanced as to render broad stroke generalizations difficult at times. So much so, that implementing policy based on results would probably create more problems than then solved–given 2011 science.

      Dartmouth publishes great stuff, but I am a cautious in readily accepting every conclusion. This is healthy I believe, and is far from the denialists who wrongfully, as mentioned, politicize their findings for advancing political agendas.


    • Harold is kind to get the ball rolling, but I’d like to summarize my argument here:

      (a) quantity is REALLY important in Medicare as shown by price-adjusted expenditure and utilization measures varying as much as 8-to-1 across regions (www.dartmouthatlas.org)

      (b) prices are also VERY important in the under-age-65 private markets where market power matters a great deal!

      So prices and quantities matter. There’s more difference between us on the policy prescription; I’d rather cut quantities than slash prices.

    • jJon
      If you would be so kind to comment, is this the MedPac study you are referencing?


      Given variation vs prices, not at the fringes, but in the middle (FIg 2, pg 5), can the mantra of “its the volume” be the conclusion?


    • The Medpac report that Brad and Professor White refers to shows ‘service use’ and ‘spending’ in figure 2. I believe White refers to the ‘service use’ aspect to indicate that in fact, after reasonable adjustment, regional variations are not very large. He concedes that illness adjustment likely over adjusts but he reasonably concludes probably by not that much. In a paper under review we (Dartmouth TDI) found that at the tails the over adjustment (for mortality) is likely on the order of 30%- really large. While there are certainly differences in illness between areas, measuring it via claims data (and HCCs) produces unacceptable biases- one appears to be closer to the truth by simply age, sex and race adjustment.