A Side Trip into Diabetes: Selling the Wrong Disease Model

This blog is about the ethics of the relationship between medicine and the pharmaceutical industry. Sometimes it seems that it is also about cholesterol and statin drugs. I've posted a lot on the latter subject because I think that the controversies over statins highlight a number of issues pertinent to the subtle effects of the drug marketing system on medical practice. Ditto for my occasional comments on why the serotonin model of depression is so widely believed even though research has thoroughly debunked it. Based on this and previous posts, I will now have to add Type 2 (adult) diabetes to my list of "regular side trips."

I have also had occasion in many previous posts to tip a hat to my friends Rick Bukata and Jerry Hoffman of Primary Care Medical Abstracts. (A footnote about them at the end.) For their June edition Rick and Jerry decided to go on vacation and turn their audiotaping over to two other good friends of mine, Gary Ferenchick (internal medicine) and Henry Barry (family medicine) at Michigan State. Most of the rest of this post is prompted by some of Gary and Henry's discussion.

One of the articles Gary and Henry reviewed was by a team out of Cardiff, Wales headed by Craig J. Currie, and interestingly enough including some people from Eli Lilly, in Lancet (subscription required). Basically the article addresses the fact that most diabetes treatment guidelines act as if lowering blood sugar--or lowering the hemoglobin A1C, which measures blood sugar trends over the long haul--is the be-all and end-all of Type 2 diabetes treatment, and the lower the A1c the better. Currie's team did a retrospective cohort study, which as such is unable to prove cause and effect. But its results jibe with many other controlled trials, all of which showed no outcomes benefit in Type 2 with tighter control of blood sugar, but an added risk when blood sugars are dropped too low by medication.

Currie's gang plotted all-cause mortality against A1C levels in nearly 50,000 older adults followed over a 20-year period. The graph they should have ended up with if the guidelines are correct would be a straight line downward toward the left, with fewer deaths the lower the A1C. What they actually found was a U-shaped plot, where the best overall A1C was about 7.5%, and death rates went up on either side. One scary feature of their diagram was that if you had a hemoglobin A1C of 6.5, generally considered "excellent blood sugar control," your death rate was actually higher than if you had a level of 9.5, which today would be viewed as rotten blood sugar control.

My pals Gary and Henry went on to chat at length about the new pay-for-performance systems whereby the amount of money docs get from insurance companies or Medicare depends on how well you meet various "quality indicators" set by current guidelines. A physician who practices according to what Gary, Henry, and I would consider to be the best available evidence on diabetes Type 2 would probably mostly ignore blood sugar anyway, and stress smoking cessation, lifestyle changes (especially exercise), and aggressive management of heart-disease risks (especially blood pressure control). If this evidence-based physician did pay attention to blood sugar, she would aim for a moderate A1C of around 7.5, as indicated by the Currie research. Many pay-for-performance programs, by contrast, would ding the doc for having patients above 7.0 and would pay a bonus only for patient who were below 7.0 A1C. So basically you have physicians being paid a bonus for putting their patients at higher health risk, all in the name of so-called "evidence-based" guidelines.

Two basic take home messages here. First, as we have been ragging about in the Avandia case, how did the basic disease model of "Type 2 diabetes = problems with high blood sugar" continue to gain such a powerful hold over the minds of physicians, biomedical scientists, and policymakers, when the best evidence today says rather that "Type 2 diabetes = increased cardiovascular risk, not necessarily made any better by tight sugar control"? Is it just a coincidence that the first model sells a lot more drugs than the second model? Is it just a coincidence that the marketing juggernaut of Big Pharma sides squarely with the first model and wants us not to hear about or think about the second?

The second take-home message is the unholy alliance between industry interests and the push to pay-for-performance. The P4P movement seems on the surface to be a no-brainer. Why pay docs for providing bad care when you could pay them only for providing good care? The devil-in-the-details of course is agreeing on what good care is and measuring it accurately in the real world. There, you create a drug marketer's paradise. In olden days the drug company could sell a drug like Avandia only by marketing individually and persuading (also known less delicately as "bribing") 800,000 free-range physicians in the U.S. Today, you can kiss off 799,00 of those guys and just focus on bribing 1000 docs--so long as you can get to the ones who write the P4P guidelines. Then you have the interesting coincidence that all the guidelines fly in the face of the most persuasive studies by still harping on blood sugar control; and that a majority of the writers of the guidelines happen to be paid speakers and consultants for the drug companies.

Thus we are back to Avandia, which as all readers no doubt know by now dodged the FDA advisory committee bullet this past week and will no doubt remain on the market. The stories in the media all say how Avandia was once the best selling diabetes drug and now has fallen on hard times due to fears of excess heart risks. None of the stories address the core question, How the heck did Avandia ever get to be the top-selling drug for Type 2 diabetes, when no study ever showed that it reduced any long-term cardiovascular risks from diabetes? The only way this could happen was for the industry first to"sell" medicine and the public a false model of the disease--which they seem to have done very well. They've done it so well that a doc who actually tries to practice evidence-based diabetes care will probably go broke under P4P, and will have his patients leave in droves because he won't recommend that they all get their free glucometers that are advertised all over daytime TV.

The footnote I promised is that Gary and Henry began their audiotape with a plea to Rick and Jerry that I would strongly second, even though I am sure it is futile. The three of us are all former faculty in the Primary Care Medical Abstracts CME courses. These courses were a lot of fun to teach, attempted to present evidence-based medicine, and were completely paid for by the attendees' registration--no drug company advertising of any sort. As you could guess, when all the competing CME courses were heavily subsidized by drug dollars, Rick and Jerry eventually had to quit offering these courses to primary care docs due to low enrollment. (As embarrassing for me as a primary care person to admit it, the ER docs still manage to pay for their own CME through the emergency medicine courses that Rick and Jerry continue to teach.)

Currie CJ, Peters JR, Tynan A, et al. Survival as a function of HbA1c in people with type 2 diabetes: a retrospective cohort study. Lancet 375:481-89, Feb. 6, 2010.