Conflicts of Interest as a Health Policy Problem: Industry Ties and Bias in Drug Approval

On September 30th 2014, Genevieve Pham-Kanter, assistant professor in the School of Public Health at Drexel University, presented her research on conflicts of interest in the biomedical industry in the Edmond J. Safra Center’s third lab seminar of the semester. Pham-Kanter first presented a case for why conflicts of interest in medicine can be productively viewed as a health policy problem rather than a moral dilemma faced by individual physicians, and then moved on to explain her research on how different types of ties to industry may reflect bias when it comes to drug approval by the advisory committees of the Food and Drug Administration (FDA).

In order to provide a framework for her research, Pham-Kanter gave a short introduction to the problem of conflicts of interests in medicine. Pharmaceutical and medical device companies give money to physicians and medical researchers in exchange for their expertise in the clinical and medical field. The concern is that these relationships may cause physicians and medical researchers to act, consciously or unconsciously, in favor of their own financial interests and not in the interest of their patients or the general public. Pham-Kanter argued that not all of these relationships are inherently bad as they provide benefits to both the firms and the doctors. She argued that instead of the historical view of conflicts of interest in medicine as an individual dilemma or problem of moral conduct we should switch from focusing on the individual to focusing on the system.

Pham-Kanter then moved on to the second part of her presentation, in which she explained her research on how financial conflicts of interest on FDA advisory committees may lead to biased drug approval. The FDA makes approval decisions for drugs sold in the United States and their decision relies heavily on the FDA advisory committees that are made up of external experts such as researchers and physicians. The question is whether the impartiality of these external experts is affected by financial relationships with the manufacturer or the manufacturer’s competitors. Her research also focused on distinguishing between financial ties that are innocuous or beneficial and those that are harmful. This research is important because it has far-reaching consequences in terms of public trust and public health, and it is a good topic to study as there is detailed information on the financial interests of the people on the advisory committees. In addition, drug approval is a clear measurable outcome that would benefit the pharmaceutical companies. 

The research study used 15 years of data, from 1997 to 2011, and focused on 15 Center for Drug Evaluation and Research committees. A lot of the committees varied significantly in the amount of conflicts of interest but overall the most prevalent type of financial tie was consulting, at 34%. The second most prevalent tie was ownership interest or royalties, at 25%; the third was paid speaker for the company, at 19%; and the fourth type of financial tie was being a member of the company’s advisory board, at 14%. Surprisingly low on the list was research grant or contract recipient, at 9%.

Before discussing the results, Pham-Kanter went over some possible results that might be expected and then explained a range of possible mechanisms that might generate different kinds of findings. Thinking about the mechanisms that might explain observed behavior is important because it is crucial to know why people act the way that they do in order to develop effective policies. The two mechanisms she discussed were influence and selection. With the influence mechanism, money is what influences people’s behavior and makes them act a certain way, whereas with the selection mechanism, people might be—for pre-existing reasons—predisposed to favor pharmaceutical companies and so self-select to work with industry. Selection could also occur when physicians and researchers with a lot of expertise in that area are being selected by the pharmaceutical companies to work for them. This becomes important when we begin to think about policies because, if the external experts were biased because of influence, then the optimal policy would be to ban financial ties. Policy choices become more complicated if the potential driving mechanism is selection, because then the industry could be selecting these physicians and clinicians due to pre-existing bias or because they just have a lot of expertise in that area. If a ban is enacted then you could possibly be getting rid of the good with the bad.

The results of the research study found that members of the advisory committee with an exclusive tie to the sponsor company often favored the sponsor. It is not possible with this data to tell if this is due to influence or selection. Another finding is that members of advisory committees with multiple financial ties did not favor the sponsor. This tells us that there was definitely no selection on the basis of pro-industry bias. There may have been selection on the basis of expertise, or the committee members may have simply been less influenced by any one company because of the multiple financial ties. Another interesting finding is that, when board members were on the advisory board for the sponsor and that was their only financial tie, then they were significantly more likely to vote in favor of the sponsor. Paid speakers, on the other hand, were weakly associated with voting in favor of the sponsor, whereas paid speakers for multiple companies were not more likely to vote for the sponsor than were those with no financial ties.

In conclusion, Pham-Kanter found that FDA conflicts of interest policy should focus on exclusive ties, especially ties to advisory boards, as these seem to be the largest source of bias producing behavior. From this study we also learned that multiple ties are not necessarily bad, as they do not seem to always reflect pro-industry bias. They more likely signify expertise in that field or reflect that fact that the biasing influences of multiple firms simply cancel each other out. 

- Summary composed by Lianna Llewellyn