• Legal obstacles to shared decision making

    I’ve just finished Jack Wennberg’s book Tracking Medicine. It may be the most important book about the US health system I’ve read. It’s not perfect (an unfairly high standard), but it gets far more things right than wrong.* Do yourself a favor and read it.

    Anyway, here’s a passage that made me want to hurl my iPad through a window in frustration:

    Under current law, physicians who engage in shared decision making [SDM] may expose themselves to malpractice suits. In a now famous case, at least among family practice physicians, a young resident named Daniel Merenstein helped a middle-aged patient decide whether he wanted to undergo a PSA test. The patient decided against the test, but when he subsequently went to another physician, he was given one without his knowledge. The test showed his PSA was high, and a subsequent biopsy found advanced prostate cancer. The patient successfully sued the large family practice where Merenstein was training, despite extensive documentation by Merenstein that the patient was fully informed of the tradeoffs when he made his initial choice not to get tested.

    Wennberg cites a paper by King, Stables, and Moulton (ungated pdf) that documents how the law is at odds with SDM. The paper is on my reading list.

    * The one big thing wrong, in my view, is use of “flat of the curve” language even though elsewhere in the book the evidence that contradicts just that is provided. (See my simple argument.) My only other gripe, which applies to almost every book, is that it is repetitive, making it unnecessarily long. Almost no authors are succinct enough to satisfy me, so this is not a unique criticism. If you’re pressed for time, read Part IV (Chapters 13-16) first and consume earlier chapters of the book for the supporting evidence.


    • I’m acquainted with Dr. Wennberg and look forward to reading it. But will it topple Topol? Plus he seems to have succumbed to nan increasingly common error: Using the adjective “famous” when “infamous” is called for.

    • I love the Dartmouth Atlas and honor Dr. Wennberg for the work he has done to create and inspire it. I wish two things for them both: 1) That they would take price a lot more seriously as a variable influencing U.S. costs — not price differences as explanations of Medicare spending variation (which are not consequential), but price differences as a factor in U.S. costs. Yes, the U.S. is a veritable soup of variation, but I am not aware of any evidence that we are different in this respect from other OECD countries. It always feels to me that the DA’s call for utilization reduction as the primary strategy of cost savings is a covert appeal to leave physician prices alone. Eliminate waste and pay less for it, a more complete prescription for reform.

      Second I wish that they would return seriously to the problem of finding explanations for variation, especially for what makes for low use, low cost geographies. There are many such areas that are not ‘integrated’ and many ‘integrated’ systems that are not particularly low use, and the DA seems to have given up on a rigorous description of that complexity.

      Still, what a marvelous thing is the Dartmouth Atlas and its pioneering discovery of the dialectology of American medicine. Geography is destiny. What dialect does your doctor speak?

      • I thought price was taken seriously in the book, as well as variation within integrated systems.

        • I could have missed something new on pricing, but I think the point on integration holds.

          Variation as a fact within integrated systems, they do pretty well, a whole Atlas, I believe a few years ago, and in that work they find reason to be skeptical about the degree of standardization that is found in those systems. But despite that, they still cite integration as an important contributor to lower cost, lower utilization, effective care and accountable care. What I don’t see is that they have ever made the case for that with data and hypothesis testing, and there are some possible counter arguments. Integration is not a predictor of similar patterns of care; there are more than a few low utilizing geographies with no integration; there are integrated systems with high use rates; there are highly integrated markets with low base rates but very high recent growth.

          I once overheard a Dartmouth Atlas researcher speculate that where integration resulted in low use rates it was because integrated groups acted to discourage competitors from entering the market. This he suggested resulted in the market then having a lower supply of providers, which was the causal link to lower utilization. That’s a whole different argument than integration as a vehicle for standardization.

          • Integration and continuity of care are important, but that alone doesn’t account for the variation in utilization. What integrated practices have done, whether intentionally or not (and it would seem in some cases it is not intentional) is limit investment in capital resources such as ICU beds and in labor, particularly specialist labor. They have also been successful at limiting regional supply by discouraging competitors (think Mayo in Rochester).

            John Grima’s comment about price is important, because we also know that health care markets that are more concentrated are better able to extract higher prices out of private payers. So even though their utilization rates may be lower, overall per capita spending by private payers may not be because of higher prices.

            • I wanted to return to my comments. I will confess to having spoken without reading much of this latest book by Dr. Wennberg, but I have now remedied that. I’ve just searched through Tracking Medicine for a discussion of price and reviewed the comments on integration and use rates.

              I see the usual — and so far as I know, sound — arguments about price NOT being a factor in the variation in Medicare spending in the U.S. But I don’t see a discussion of price in the sense of how much more U.S. consumers pay for health services because of the weaker position of purchasers in the U.S. vis a vis providers. We know how to change prices; it requires shifting power to purchasers and likely capping rates of increase. Primitive, but proven, politically unlikely, and ignored by Dr. Wennberg.

              We know something about changing use rates, but less than Dr. Wennberg wants to claim. We know — thanks to him and his colleagues — that shared decision making, in the context of really well worked out decision tools, works. We know that constraining supply works — but we don’t know how to constrain supply effectively over time without constraining overall budgets. I like that Dr. Wennberg plays with the idea of a transfer tax as a sort of two steps removed penalty on capacity. But his core suggestion for changing use rates is to encourage the formation of integrated systems of providers, and the DA data don’t really lead to that conclusion. The data at best seem to suggest that some integrated systems with a particular kind of culture CAN limit wasteful utilization. So can natural communities like Grand Junction — and several others that are never mentioned.

              It’s not like I know this stuff so well as not to have missed something. But the concern might be then that despite assertions to the contrary, DA research does not lead to the conclusion that ACO like organizations will achieve lower use rates, not without some vital other ingredient. And if it is the prices where we should focus, well then ACOs are provider concentrations more or less by definition, and that would an unfortunate thing to have deduced from some very beautiful ethnography.