Nature Reports on Conflicts of Interest Affecting Guideline Writers

In the 20 October issue of Nature are results of a study of conflicts of interests affecting guideline writers. The study focused on 90 of the more than 200 plus guidelines in the US National Guideline Clearinghouse in 2004 which listed authors' conflicts of interests. Of these, 66 percent had at least one conflicted author (50% had at least on author on a company advisory board or who was a company consultant, 50% had at least one author whose research was funded by a relevant company, 43% had at least one author who was a member of a relevant company's speakers bureau, 11% had at least one author who owned stock in a relevant company). Of the 685 authors of the 90 guidelines, 35% had one or more conflict (21% - advisory board or consulting position, 22% - research grant, 15% - speakers bureau, 2% - stock). (See this link for detailed results.)
The article noted the example of guidelines for the treatment of anemia in HIV infection. The convening of the guideline writers was funded by Ortho Biotech, which makes the relevant drug, epoetin alpha. All six guideline writers had received speakers' fees or consulting fees from Ortho.
An accompanying editorial suggtested "time may also be ripe for a more radical overhaul of the system, particularly in the United States, where the current financial stakes are highest...." The editorial suggests that a "mechanism could be set up at the US health department... to steer the preparation of clinical guidelines." Furthermore, "if US physicians really want to separate marketing from medicine, they should support its creation."
Whether a government mechanism would be optimal, compared to, say, a private not-for-profit effort supported by practicing physicians, is unclear, at least to me.
But it is clear that the current way to write guidelines, shot through, as it is, with conflicts, cannot be trusted.
Yet these guidelines are promoted as key to improving the quality of care, and are the basis of pay for performance, the latest fashion in quality improvement, error prevention, and cost control. The only performance that may result is that of the stocks of the companies that are paying the guideline writers.