Monday, July 21, 2008
Solution to improve integrity in medical research was proposed surrounding statistician's participation
This may be one of the most interesting articles from the president's column of AMSTAT news I have read recently. It is a reaction to an editorial of the Journal of American Medical Association that "impugning the integrity of medical science." As part of the proposed solution for the integrity problem in the pharmaceutical industry, it is suggested that clinical trial data collected by a for-profit organization (usually a company) should be analyzed by an academic statistician. The AMSTAT news article challenged this proposal by first pointing out the implicit accusation against industrial statisticians, then suggesting the remaining conflict of interests for an academic statisitican who is paid to do the analysis. Another problem with this solution, according to the AMSTAT news article, is the segregation of the design and experiment phase of the study from the analysis phase. I tend to agree with the AMSTAT news article a lot. As someone who has done some analysis for medical investigators, my first reaction is "since when have statisticians become the leading investigators of such studies? who would decide on actions that will affect the integrity of a study."