Future jobs of FDA’s haematology-oncology reviewersBMJ 2016; 354 doi: https://doi.org/10.1136/bmj.i5055 (Published 27 September 2016) Cite this as: BMJ 2016;354:i5055
All rapid responses
I am delighted that the US FDA has performed an analysis of the subsequent careers of FDA employees. I understand why Dr. Schick and Dr. Miller, employees of the FDA, wish to downplay our findings, and conclude that “substantially fewer medical reviewers” work for the industry. Rather than the correct interpretation: the rates are largely similar and concerning: we report 15/26 (58%) and they report 10/27 (37%) medical reviewers later work for or consult for the industry they regulated.
I have 2 things to say about the analysis by Schick and Miller. First, their result has underestimated careers in the industry by 5 people (10 vs 15). Second, even if you accept their estimate, the conclusions and concerns are the same.
Bien and I & Schick and Miller reach similar conclusions that either 26 or 27 FDA hematology oncology medical reviewers from 2001-2010 left the agency. Since the FDA has access to rejected applications, but we do not, it is not surprising we differ.
Next, we document that 15 later worked for or consulted for the industry, while they document only 10. We both used similar methods to identify those who went to industry—we searched LinkedIn, Pubmed, Google, etc. Although the FDA clearly has the advantage in the denominator (27 vs. 26), there is no reason to think the FDA is better at searching LinkedIn than we are, and thus they have no advantage regarding the numerator (10 vs. 15).
Let me say: there is no doubt that 15 people later worked for or consulted for the industry. We have solid documentation of this, through screenshots. 14/15 people were found to have moved to the industry by LinkedIn. In 1 case, we ascertained subsequent employment by PubMed, but there was no uncertainty. If US FDA employees say they could only find 10 people, then either 1. Several LinkedIn profiles were changed, perhaps due to the publicity surrounding our piece or 2. They did not search hard enough. But, our systematic strategy resulted in identifying 15 reviewers who left the FDA went to work for or consult for the industry, and that is indisputable. Our percentage is also similar to other analyses of the revolving door  in other industries.
Second, now let me be kinder and say the true percent is somewhere in-between 37% and 58%. Perhaps it is. Does that matter? No, it does not. In both cases, a sizable chunk of hematology oncology medical reviewers go to work for the biopharmaceutical industry. It is also worth highlighting that this percent is incredibly high given the degrees/ training these individuals have. We found 50/55 (90%) had an MD and another 3 had provider degrees (CRNO, CRNP, and MBBS). Many reviewers have fellowship training in hematology oncology. Most people with this training don’t go to work in the pharmaceutical industry at these rates 37-58%-- they largely take jobs seeing patients (where salaries are also higher than at FDA). It is hard not to conclude former FDA reviewers go to the biopharmaceutical industry because of specific knowledge and expertise they have involving drug regulation.
Schick and Miller seem not to understand the full nature of potential bias here. They think what matters is not whether FDA employees later go to work for the industry, but whether they work for the specific companies whose applications they processed. But, either is problematic.
First, because several people go to consulting, we cannot know how many directly work with companies they once refereed. Second, there may still be a bias present even if one does not work for the specific company whose application one perused. Drug companies talk to each other. Executives move around. If a reviewer knows, in the back of their mind, that 40-60% of subsequent employment will be for some pharmaceutical company, does that make a reviewer less likely to be a difficult regulator, less likely to want to raise a red flag, or less likely to desire to be thought of as a bad gal or guy? I don’t know the answer but I don’t think the question vanishes if you work with BMS and Merck, and then go to work for Pfizer.
Other experts speculate that reviewers may use contacts they have within the FDA in their future careers to gain insights or tips . Again, I don’t know that this happens or not, but this is another way bias is potentially not limited to review company X, work for company X. Journalists have identified even more problematic cases .
Ultimately, I agree with Schick and Miller that “Future researchers are encouraged to examine the career pathways associated with every scientist contributing to the review process.” To that end, I would ask the FDA to make easily available a roster of all employees by job position on their website. This information is public, and if the FDA believes such analyses are valuable, making a roster available will speed their conduct. We found the most challenging part of our analysis was building the list of reviewers. The FDA website has a wealth of information, but navigation is terribly difficult. If the FDA does not believe it should make such lists readily available, researchers and journalists should know that Freedom of Information Act (FOIA) requests might aid in their acquisition.
Our purpose in studying subsequent careers of FDA employees is not to blame or shame individuals. In fact, quite the opposite, I believe FDA reviewers are some of the most important people in cancer medicine. Additionally, I believe FDA reviewers are underpaid, and I would support dramatic increases in salary to hire and retain the best people. However, our purpose in analyzing this question is to study structural factors that may contribute to bias. The fact that 58% of former FDA medical reviewers work for or consult for the industry is a major structural bias. We must work to identify what policies may mitigate the prevalence of this revolving door. I thank Drs. Schick and Miller for engaging with our work.
Competing interests: No competing interests
Recently published research suggests that medical reviewers at the Food and Drug Administration (FDA) regularly go on to work in companies they once regulated.1 We investigated this assertion using an internal FDA data set, and show that medical reviewers are substantially less likely to go on to work in regulated industry than was reported in past research.
Bien and Prasad (2016) examined this question for medical reviewers of new oncology and hematology drug applications between 2001 and 2010.1 Their results indicated that 27 percent of medical reviewers went on to work at or consult to pharmaceutical companies, and that medical reviewers who left the FDA were more likely to go on to work in regulated industries than non-regulated industries. Although these results are suggestive, they do not address whether medical reviewers who work on new drug applications go on to work in companies they once regulated.
We examined this question using an internal FDA data set containing every new approved and unapproved oncology or hematology drug application—small molecule and biologic—undergoing a review sometime between January 1st, 2001 and December 31st, 2009. We then linked these applications to their relevant medical reviewers, and examined these reviewers’ career paths using an internal FDA employee database and publicly available information provided on LinkedIn, Zoominfo, PubMed, and Google. Careers were examined between Jan, 2001 to Oct, 2016.
We identified 60 unique medical reviewers who evaluated 22 new biologic applications and 52 new chemical entities from 2001 to 2010. Figure 1 shows these reviewers’ subsequent jobs. Of the 60 reviewers, 33 (55%) continued to work exclusively at the FDA, 10 (17%) went to work in or consult with regulated industry directly after leaving the FDA, and 17 (28%) retired or went on to work in non-regulated institutions, such as academia, other government agencies, or medical institutions. Of the 27 reviewers who left the FDA, 10 (37%) went to work in regulated industry whereas 17 (63%) chose career outcomes not in regulated industries.
We find that substantially fewer medical reviewers in oncology and hematology went on to work in industry then was reported in Bien and Prasad (2016).1 Our results also indicate that only one reviewer directly left the FDA to work for a company whose new drug applications they reviewed. Some attrition to industry is to be expected considering the large disparities in compensation; industry oncologists earn a median salary of approximately $420,000 whereas their FDA counterparts earn approximately $210,000.2
Our study is limited by the focus on a single occupational type (medical reviewers), application type (new molecular entities) and therapeutic area (oncology and hematology). Future researchers are encouraged to examine the career pathways associated with every scientist contributing to the review process of new molecular entities, new drug applications, and investigational new drugs in each therapeutic area.
Competing interests: No competing interests