Trends in clinical trial investigator workforce and turnover: An analysis of the U.S. FDA 1572 BMIS database

Abstract

Background: High turnover rates among clinical trial investigators contribute to inefficiency, instability, and increased costs for the clinical research enterprise; however, factors contributing to investigator turnover have not been well characterized.

Methods: Using information from the U.S. Food and Drug Administration's Bioresearch Monitoring Information System (BMIS), we examined trends in the overall clinical investigator workforce and within specific "phenotypes" as well as differences by investigator location (U.S.-based vs. non-U.S.-based). We identified unique investigators within the database, stratifying them into one of three "phenotypes": those with one Form FDA1572 submission across the study interval ("one-and-done"); those with two or more submissions but with substantial intervals between trials ("stop-and-go"); and those with two or more submissions and continuous involvement in multiple trials ("stayers").

Results: Of the 172,453 unique investigators who submitted a Form FDA 1572 during the study interval (1999-2015), 85,455 were classified as "one-and-done" investigators; 21,768 as "stop-and-go" investigators; and 65,231 as "stayer" investigators. The total number of investigators declined across the study interval. Among all subgroups, only "one-and-done" investigators showed growth across the study period, largely driven by increases in non-U.S.-based investigators. "Stop-and-go" investigators showed declines for both U.S.-based and non-U.S.-based investigators, as did "stayers," who showed the largest absolute and proportional declines of all subgroups.

Conclusions: From 1999 to 2015, investigators submitting a Form FDA 1572 to the BMIS database declined by approximately one-third and the proportion of investigators involved in only one trial increased, signaling potential adverse trends in the clinical investigator workforce. Strategies for sustaining investigator engagement warrant further exploration.

Keywords: Drug trials; FDA; Investigator turnover.

Fig. 1

Temporal trends in investigator turnover, all investigators, 1999–2015.

Fig. 2a

Temporal trends in investigator turnover, 1999–2015, according to subgroup “one-and-done.” Investigators were classed as one-and-done if they had only 1 Form FDA 1572 submission across the entire study interval.

Fig. 2b

Temporal trends in investigator turnover, 1999–2015, according to subgroup “stop-and-go.” Investigators were classed as stop-and-go if they had at least 2 Form FDA 1572 submissions, but with the time to second submission occurring beyond the 75th percentile of the interval between the first and second submissions. Note: There were very few “stop-and-go” investigators after 2012 due to the 1013-day minimum period between first and second submissions required to be classified in this category.

Fig. 2c

Temporal trends in investigator turnover, 1999–2015, according to subgroup “stayer.” Investigators were classed as stayers if they had at least 2 Form FDA 1572 submissions, but with the time to second submission occurring within the 75th percentile of the interval between the first and second submissions.