The FDA went against the recommendations of advisory committees 22% of the time between 2008 and 2015, according to a study.
Writing in the Milbank Quarterly, researchers present an analysis of 376 voting meetings requested by the FDA. Most of the meetings discussed whether to recommend the initial approval of biologics and small molecules, although the data set also features a small number of votes on medical devices, supplemental indications and safety questions.
The FDA disagreed with the recommendations of the committees more than one-fifth of the time. The agency was less likely to go against a recommendation when the panelists were united in their opinion. None of the other factors looked at in the study, such as public speaker favorability, were statistically correlated with the likelihood of the FDA rejecting the committee’s consensus view.
While the study shows the FDA is willing to go against committees, particularly when there is some disagreement among the experts, the findings suggest it typically takes a more restrictive position than panelists.
Of the 83 cases of discordance between the FDA and committees, 62 relate to times when the agency issued an unfavorable decision after receiving a positive recommendation from the panelists. That suggests companies that have submissions knocked back by advisory committees should hold out little hope of the FDA approving their products. But the picture may be more nuanced than that.
Almost half of the times the FDA went against a favorable committee recommendation took place in 2008 and 2009, the first two years of the eight-year period analyzed in the study. Less than 10% of the times the FDA took a more restrictive view than a committee happened in the final two years of the data set. The FDA took more restrictive positions less and less over the years.
There is less of a clear trend in the data on less-restrictive decisions. Of the 21 times the FDA reached a less-restrictive position than committees, 52% of cases happened in the second half of the analyzed period.