Buying Science

Following the money in the lab


Earlier this week the Center for Science in the Public Interest issued a report, Ensuring Independence and Objectivity at the National Academies. The National Academies of Science, like many government agencies, routinely recruit outside experts from universities, industry and other organizations to advise them on scientific matters. Just how objective are such experts? The CSPI report found that over a five year period that one out of five scientists on 21 different NAS scientific panels had "direct conflicts of interest." Famously, CSPI describes itself as the "food police" decrying all manner of fat and cholesterol laced delicacies. The CSPI released its new study at a panel discussion, Government Science Panels: Fair and Balanced?, at the National Press Club this past Monday. The CSPI is not alone in worrying about conflicts of interest on government scientific panels.

In April the activist group Public Citizen, published the results of its study of financial conflicts on the voting patterns of U.S. Food and Drug Administration advisory panels in the prestigious Journal of the American Medical Association. A total of 221 meetings held between January 2001 and December 2004 by 16 advisory committees were analyzed. Public Citizen's study found that at 73% of meetings, at least one advisor had financial ties to industry. The Washington Post ran brief story about the findings headlined, "Conflicts of Interest Influence Voting on FDA Advisory Panels, Study Finds."

"Follow the money," is the hoariest maxim of journalism and it's a pretty good one too. So it is no wonder that activist watchdogs want extra scrutiny of the people who sit on government scientific advisory panels. Government regulations based on scientific advice of these panels are pervasive and can make or break a company or product or protect or fail to protect the public's health. The FDA alone regulates companies that produce prescription drugs, medical devices, food, and cosmetics accounting for 25 percent of our national economy. With so much money at stake, it is more than reasonable to suspect corporations of trying to influence agencies to make decisions favorable to themselves and against their competitors.

So what did the CSPI and Public Citizen watchdogs turn up? Let's consider the Public Citizen study first. Surprisingly, that report found "there was no relationship between the conflict rate and voting outcome" on FDA advisory panels. The Public Citizen researchers did find that for each advisor with a conflict, there was a 10% greater likelihood that the meeting would favor approving a particular drug. That being said, the researchers also admit that they found that excluding all panel members with conflicts would not have changed whether the vote result was favorable or unfavorable toward any drug. "An FDA adviser's financial connections to the drug companies had no statistically significant effect on the approval of new drugs," is the way that a Harvard Medical School professor and Massachusetts General Hospital endocrinologist tartly summarized the findings in a recent op/ed in the Washington Post. The Washington Post's headline about financial conflicts influencing voting on FDA panels is correct, but the Boston Globe's headline "Study rebuts conflict fears around FDA" seems ultimately more accurate.

In any case, the House of Representatives passed legislation that would ban any scientists and doctors with conflicts of interest from serving on FDA advisory panels. And the FDA announced this week that it was formulating new guidelines that would for the first time set explicit rules for when scientists with ties to industry may appropriately serve on scientific advisory committees.

So what about the CSPI's study of financial conflicts at the National Academies of Science? The NAS is a Congressionally chartered non-profit corporation made up of distinguished scientists who provide free scientific and technical advice to the government. The NAS operates by creating ad hoc committees of researchers assembled to answer questions posed by Congress and various federal agencies. The CSPI looked at 320 members of 21 NAS committees over the past 3 years to find at least 18 percent of them had direct conflicts of interest. The CSPI defines direct conflicts of interest as "a financial tie within the last five years to a company or industry that is relevant to the committee topic."

Unlike scientific advisory panels at agencies like the FDA or the EPA which often take formal votes on issues, the NAS committees issue reports generally based on consensus. Since no formal votes are taken at NAS committee meetings, the CSPI report simply compiled evidence that 66 of the 320 panel members analyzed did "lean to industry" while only 9 did "lean to public interest groups." The CSPI wants future NAS committees to be "balanced" and recommended that the NAS expand its definition of "balance" to include "bias and point of view," not just scientific expertise. The CSPI defined balance "as never having more than three [pro-industy] scientists on any committee and always balancing pro-industry scientists with at least an equal number of public health oriented scientists." Imposing this kind of "balance" would explicitly inject politics into what here-to-fore has been conceived of as fact-based activity. Would modeling scientific advisory panels on boards like the Federal Elections Commission which is "balanced" with equal numbers of Democrats and Republicans really improve their scientific advice?

During the panel discussion at the National Press Club, Cleveland Clinic cardiologist Steven Nissen, who is fierce critic of the FDA, warned against "intellectual conflicts of interest" on scientific panels. Nissen evidently meant that people who are known as strong advocates for a particular therapy or view should not serve on panels or should recuse themselves. But might not such a rule routinely exclude "public health oriented scientists"?

The CSPI also recommended that NAS adopt rules that would "exclude any scientists with conflicts of interest from committees unless their expertise is crucial to the successful completion of the committee's task." So if all scientists with direct conflicts of interest which, as defined by CSPI, means having a financial connection with a company in the past five years are excluded—who then are the "pro-industry" scientists that must be balanced against?

Did the CSPI study uncover egregious behavior by NAS advisory panels? Not really. Michael Jacobson, head of the CSPI admitted, "Whether [sic] complete avoidance of conflicts of interest on committees would have improved the committees' recommendation is impossible to know." Jacobson also acknowledged, "NAS reports invariably earn high marks from the scientific community, and this study, which did not evaluate the quality of any particular NAS report, makes no effort to question that consensus view." It seems that the report has not uncovered any specific problem other than the fact that scientists of the sort the CSPI likes are underrepresented on the NAS panels.

The good news for us all from these activist efforts to "follow the money" at FDA and NAS scientific advisory panels is that their scientific advice doesn't seem to have greatly distorted by the influence of filthy lucre. This does not mean that there are not instances and institutions where money has distorted scientific conclusions. And this does not mean that our scientific institutions cannot be further insulated from the temptations of cash.

One intriguing proposal about how to limit conflicts of interest on scientific advisory panels was suggested by George Washington University public health professor David Michaels during the panel discussion organized by CSPI. Michaels explained that the International Agency for Cancer Research regularly organizes scientific panels to evaluate various substances for their carcinogenicity.

According to Michaels, the IARC excludes any scientists with real or perceived financial conflicts of interest from panel membership. However, scientists representing industry, activist groups, government agencies and any other interested parties are invited to make scientific presentations to the panel. Such outside experts may not vote on or write the text of the panel's conclusions. Of course, the devil is in the details of selecting the voting panelists. It is not unreasonable to suspect that IARC, being a World Health Organization agency, might tend to select scientists more at the public health oriented end of the spectrum which could make them too cautious when it comes to weighing evidence. Nevertheless, the IARC model bears looking into farther.

Science, like all other human enterprises, is buffeted by the same influences that batter us all—ambition, money, fame, politics, religion, and ideology. In 1997, an editorial in the British medical journal The Lancet got it about right:, "The only way to minimise bias among interpretations is to allow maximum dialogue from all parties, irrespective of their interests."

Disclosure: I have been accused of all manner of conflicts of interest generally summed up by my detractors as being a "corporate shill." Whatever motivates my reporting and analysis—maybe it's mental illness– it sure is not money. I do plead guilty to having what Dr. Nissen called "intellectual conflicts of interest," but then so does everybody else.