Sometimes a person might wonder how much difference our myriad regulations and reporting requirements make. Does it really affect anything to answer (or analyze) the same question repeatedly? Financial conflicts of interest could seem an example. What kind of impact does the reporting of financial interests have? Recently, the possible influence of financial ties got attention at one of the highest levels of authority in clinical research: an FDA advisory committee meeting.
In September an FDA advisory committee considered the results of a study involving Pfizer’s smoking cessation treatment, Chantix. The stakes for Pfizer were high: it hoped to remove a black box warning that Chantix has carried since 2008 about possible psychiatric side effects, but a possible conflict of interest threatened this outcome. The study before the committee involved over 8,000 subjects at 139 sites around the world, and nestled among those sites were some investigators who had reported significant payments from Pfizer outside of the study.
In cases where possible financial conflicts of interest exist, the investigator, his/her institution, and the ethics review committee (aka the IRB) must review the circumstances and determine what measures might be appropriate.
IRBs have a range of tools at hand: they can require that the consent form identify the financial conflict of interest; they can exclude the investigator from participating in the informed consent process; or they can remove the investigator from the study entirely. More broadly, IRBs and the designers of studies also can consider how well the protocol itself shields participants and any study results from conscious or unconscious bias.
A report earlier this year suggested that even inexpensive gifts can predispose physicians to prescribing a product. The potential of financial influence is also a concern in clinical research. Financial conflicts of interest may have factored into the high-profile research tragedies of Jesse Gelsinger at the University of Pennsylvania in 1999 and Dan Markingson at the University of Minnesota in 2004. FDA analysts may have had these cases in mind when they presented the results of the Pfizer study.
In the Chantix study, the potential role of additional payments to investigators appeared on pages 63-66 of a 564-page memo. The memo observes that serious psychiatric episodes—that is, possible negative effects of the medication the participants were taking—were reported at lower-than-average rates at sites where investigators had received more than $25,000 from Pfizer. In one arm of the study, sites without the additional financial involvement reported events in 2.2 percent of participants, while those with the larger financial ties reported events in 1.5 percent of participants. In a second arm of the study, the difference was larger: 6.4 percent versus 1.4 percent.
After identifying these possible correlations, the report also suggested the apparent trend could be a coincidence. Of the 132 investigators in the study, 32 reported financial ties, a possibly insignificant number. In addition, circumstances at the 32 sites seemed too varied to demonstrate a clear pattern. As a third point, the protocol required blinding the researchers to which treatment anyone received. (It could have been Chantix, a competing GlaxoSmithKline product, or placebo.)
The memo left it to the committee to decide. In its summary of the possibilities, the memo states : “(T)he observed difference in NPS (neuropsychiatric) event rates between sites with and without financial involvement may be attributable to chance due to the small sample size among sites with financial involvement and the high heterogeneity in event rates across sites.”
The committee may have reached the same conclusion, and decided that the financial ties did not affect the study results. In the end, the advisory committee recommended that the FDA remove the black box warning, and, in early December the FDA did just that.
Conflicts of interest receive a lot of attention in clinical research, as they should. Financial (and other) ties can erode confidence in the objectivity of results. But transparency has value; it is reassuring to come across an instance where investigators communicated their financial ties, and those ties could be assessed alongside concrete study results at a moment when it mattered most: deciding whether a medication is safe.