More on Conflicts - in Business and Medicine

The chairman of the Federal Reserve of New York, Stephen Friedman, resigned recently after he faced questions regarding how he could simultaneously sit in a “public” seat and chair the New York Fed (including while it approved Goldman Sach’s request to become a bank holding company), serve on Goldman’s Board, and purchase 37,000 shares of Goldman stock – the value of which has risen $1.7 million since he purchased them in December. http://dealbook.blogs.nytimes.com/2009/05/07/friedman-resigns-as-chairman-of-new-york-fed/

According to Dealbook, the General Counsel of the NY Fed reports that Friedman did nothing wrong. Perhaps this is true. If so, I think there must be several thousand unemployed (fill in the blank) who would be happy to take a crack at rewriting the Fed’s conflict-of-interest and self-dealing rules to make sure such conduct would be wrong in the future.

My attention, meanwhile, has been instead drawn to conflicts in medical research. An article regarding how conflicts of interest impact medical research claims that a review of 1500 studies shows significant pro-sponsor bias. The article was published May 11 in the on-line version of Cancer.  Its relevant findings summarized as follows:

Analysis of 124 oncology clinical trials showed that those with a conflict of interest – either direct industry funding or an author’s declarations of financial relationships – were more than twice as likely to find significantly improved patient survival, according to Reshma Jagsi, M.D., D.Phil., of the University of Michigan, and colleagues.

. . .

Among 52 randomized, controlled trials with no conflict of interest, 14% found significantly better survival with the intervention relative to control, 72% found equivalent survival, and 6% significantly favored the control.

In 72 similar trials with a conflict of interest, 29% found in favor of the intervention, 61% showed no difference, and none reported better survival with the control (P=0.04 for trend relative to nonconflicted studies).

See John Gever, Industry Ties Predict Outcomes in Cancer Research, Medpage Today, available at http://www.medpagetoday.com/PublicHealthPolicy/ClinicalTrials/14133.

In other words, conflicted studies were more than twice as likely to find a positive correlation between the sponsor’s product and positive outcomes than nonconflicted studies. And the likelihood of a negative correlation shrank from 6 percent for nonconflicted studies to 0 percent for conflicted studies.

Too bad legal practice does not lend itself to such a clean comparison of conflicted and nonconflicted representations.