Monday, April 27, 2009

Two of my favorite medical bloggers seem to disagree regarding CER

When I read a recent blog posting by DB,who postings inspired me to start to blog and with whom I rarely disagree, I was sharpening up my blogging tools to respectfully offer a view counter to his, but as is often the case DrRW beat me to the draw (See here).

Those who oppose or at least have serious doubts about government- run CER are likely not against comparing the effectiveness of various medical treatments ( i.e.. not against CER) but are very skeptical of the ability of a government agency to regularly turn out results that will be free of bias and may also reject the thesis that all drug company sponsored research is biased.

Research can be spun one way or the other-there are so many ways to stack the deck and cook the books that lurk beneath behind the pages of a research project.To believe that a government agency is a) without an agenda and b) incapable of making such an agenda-generated bias operational requires much more faith in the integrity of the government that I summon and seems to ignore what we have learned about how governmental entity really operate versus the high school civics views that describes things in terms of purported aims rather than in real results. A leisurely half hour or so with a discussion of Public Choice theory might disabuse many of the vision of government as a hybrid between Santa Claus and superman.

Dr. RW's final paragraph deserves much consideration:

No one that I know of is objecting to more unbiased data. But CER is not inherently unbiased. Moreover, it is inherently susceptible to design flaws for reasons I pointed out here, with several examples. Bias has more to do with who’s sponsoring the research than the type of research. There’s no reason to think that the government would introduce less bias. In fact, the government policy makers who are pushing CER are explicitly very biased. If you don’t believe me just read the Congressional Budget Office paper which was pushing for CER, which I cited here.

No comments: