• Provider-induced demand

    We all know that patients get loads of advice on what treatments might be best for them from physicians. It follows that the majority of care is provider-driven. Patients aren’t really demanding care so much as it is being forced down their throats by greedy doctors. That’s what’s revealed by the literature that shows tremendous variation across geographic regions in health care utilization and spending. It’s all the providers’ fault.

    Ha! Did you really think I’d write that? Well, yeah, I did, but it’s meant as an example of what I’d never write. It’s not the TIE way. We need evidence of provider-induced demand, not just speculation. Variations in utilization are suggestive, but not definitive. Maybe a lot of it is due to patient culture. Maybe providers are just doing what patients want. Isn’t this the era of “patient-centeredness” anyway? (And, for the record, if provider-induced demand is a large phenomenon, greed need not be and probably isn’t the dominant motivation.)

    What papers would help tease apart provider-inducement from patient culture? I emailed Chapin White this question. Here’s what he suggested. All links are ungated.

    Barnato, A. E., M. B. Herndon, D. L. Anthony, P. M. Gallagher, J. S. Skinner, J. P. W. Bynum, and E. S. Fisher. 2007. “Are Regional Variations in End-of-Life Care Intensity Explained by Patient Preferences?: A Study of the US Medicare Population.” Medical Care 45(5), 386-93.

    Dartmouth Center for the Evaluative Clinical Sciences, Supply Sensitive Care, Dartmouth Atlas Project Topic Brief, Lebanon, N.H. (Jan. 15, 2007).

    Finkelstein, A. 2007. “The Aggregate Effects of Health Insurance: Evidence from the Introduction of Medicare.” Quarterly Journal of Economics 122(1), 1-37. [Chapin says this is suggestive but not a smoking gun.]

    Newhouse, J. 1992. Medical Care Costs: How Much Welfare Loss? Journal of Economic Perspectives 6(3):3-21. [Chapin says this is a classic.]

    All are now on my reading list. If you’re interested in finding more literature on provider-induced demand, here are a few things you could do: (1) Search the term in Google Scholar; (2) Consult relevant citations in the above articles; (3) Find each of the above articles in Google Scholar and see what cites it. If you come across anything useful, please suggest it in the comments.


    • Hey Austin, the references above are in Tracking Medicine, except for Newhouse. I’m looking forward to your interpretation.

    • You might try to look at some attempts to replicate the supply-cost-outcomes analyses. It turns out, the evidence is rather mixed on this.

      Ricketts TC, Holmes GM. (2007) Mortality and physician supply, does region hold the key to the paradox. Health Services Research, Volume 42 No. 5. 2233-2250.

      Ricketts, T. C. and D. W. Belsky (2011). “Medicare Costs and Surgeon Supply in Hospital Service Areas.” Annals of Surgery. Online ahead of print Oct 2011.

      The Dartmouth group looks pretty carefully as the ecological associations in this article: Chang, C. H., T. A. Stukel, et al. (2011). “Primary care physician workforce and Medicare beneficiaries’ health outcomes.” JAMA : the journal of the American Medical Association 305(20): 2096-2104.

    • “And, for the record, if provider-induced demand is a large phenomenon, greed need not be and probably isn’t the dominant motivation.”

      Good common sense, as only an economist could provide.

    • You’re absolutely correct, Bob. Also for the record, “Supply-sensitive care” is the technical term. Provider induced demand is more reactionary – which Austin used most likely because it works. Just look at the reactions.

    • Gawande has suggested that practice patterns learned as residents largely determine utilization. Will look back and see if he cited literature for this, or if it is just personal belief.


    • A note on terminology: Gawande’s statement reflects the stability of regional practice patterns. A region’s “Surgical Signature” is remarkably stable over time. (Read more about regional surgical signatures in “Tracking Medicine”.) The key word is regional.

      Dartmouth’s findings should not be interpreted as provider induced demand. That would be a case of extending the data beyond the findings.

      The more technically correct description would be care that is sensitive to regional supply. For example, the average patient with high blood pressure is more likely to be seen more frequently for follow up if she is in a region with a higher number of doctors per capita. The same patient will be seen less frequently is a region with less supply.

      The supply influence is occurring at the regional level and without any conscientious effort on the part of providers or patients.

      • Great comment, as well as the one on terminology. Thanks.

      • Thom
        Your comment is a good one. I did think a bit after reading it though, and had some thoughts:

        There are many investigators who see the variation map not just at the HRR level, but at the hospital and individual one. Through that lens, there are overlooked nuances, and the answers as to why docs do “stuff” becomes less clear.

        Pockets of high utilization exist where supply is not abundant for example. Of course, operative factors in one domain may not be those exercised in another. However, if one is to consider practitioners providing services with greater intensity in these “low” supply areas, one cannot rule out a conscientious, or more overt, aggressive style of care, eg, as simple as repeat visit every 4 instead of 8 weeks.

        I raise this issue not for its theoretical potential, but on the merits. Some docs are aware and practice in this mode of care delivery. Many are inculcated into this cultural style, and its not exclusive to high supply regions. It is provider or supplier induced demand.


    • Thanks for your comment Brad. You make a bold statement when you say ” It is provider or supplier induced demand,” Bold statements need citations & data.

      If I were to write “it is possible there are healthcare providers performing unnecessary procedures in areas of high or low provider supply” is like writing “fraud exists.”

      Fair enough. Probably true. But, identifying it, quantifying it, and doing something about are another matter altogether.

      The point of my earlier comments was to clarify that Dartmouth’s findings of supply-sensitive care are regionally based. That is a strength and a limitation.

      Extending these findings to the individual level is a fallacious.argument.

      Some blatant self-promotion- I hope you visit my invited blog post that began this conversation.



      • Though I agree with Thom about the strict, proper interpretation of the Dartmouth findings, it is not unreasonable to think it is consistent with provision of both unnecessary care and some high-intensity care that is within the range of reasonableness.

        I do not think it that “it is possible there are healthcare providers performing unnecessary procedures in areas of high or low provider supply” is the same as saying “fraud exists.” Fraud, as I interpret it, implies intent, culpability, and responsibly. I do not think the physicians who perform procedures we researchers believe (yes, from studies!!!) are, in general, not worthwhile, are trying to sell snake oil and know it. I think they honestly disagree with the studies (there is some room for debate on some of them), don’t know about the studies, or are stuck in a bad situation. What would you do, at, say, age 55 when you’ve been trained to do X and now we think X is not worthwhile? Kind of hard to give up X. I don’t blame them!

        This is a systemic problem. There is overprovision of low-value and no-value care. Dartmouth results are consistent with that but do not tell us precisely where it is. Nor do they tell us what to do about it. Expecting physicians to just hang up their stethoscope and give up their livelihood because a study has found their work to be not distinguishable from placebo or no better than a cheaper alternative is asking far too much. It’s not fraud, not in general, and not at the provider level. It’s what Brad said it is, “healthcare providers performing unnecessary procedures.” That’s a problem, but the finger points in more than one direction.

    • Austin,

      We are in agreement.

      I wrote these lines earlier:

      If I were to write “it is possible there are healthcare providers performing unnecessary procedures in areas of high or low provider supply” is like writing “fraud exists.”

      I was not saying unnecessary or unproven treatment is fraudulent. I was saying that both statements as written are probably true. I was making the point to Brad that such statements are best backed by data. I know you agree there.

      Identifying unnecessary or fraudulent care at the individual level, quantifying either, and doing something about them are another matter altogether.

      Thanks for the conversation.

      • Thom
        Very rich thread. This is stimulating.

        Below link takes you to an insttructive paper that reviews some of the regional vs local/individual issues you speak about. I could be mistaken, but I dont believe the dissociation is entirely fallacious. There is some heft beyond what Dartmouth asserts.

        Also, I did not wish to convey that SID is the sole explanation for utiliztion in low supply areas. Merely, its partially causal– and coexists with other potent factors that are difficult to disentagle.