Recruitment strategies and lessons learned from a large genetic study of African Americans

Abstract

Genetic studies must enroll large numbers of participants to obtain adequate statistical power. Data are needed on how researchers can best use limited financial and practical resources to achieve these targets, especially in under-represented populations. This paper provides a retrospective analysis of the recruitment strategies for a large glaucoma genetics study in African Americans. The Primary Open-Angle African American Glaucoma Genetics study enrolled 10,192 African American subjects from the Philadelphia region. Major recruitment approaches included clinic enrollment from University of Pennsylvania (UPenn) sites, clinic enrollment from external sites, sampling of Penn Medicine Biobank (PMBB), and community outreach. We calculated the enrollment yield, cost per subject, and seasonal trends of these approaches. The majority (65%) of subject were enrolled from UPenn sites with an average cost of $133/subject. Over time, monthly case enrollment declined as the pool of eligible subjects was depleted. Expanding to external sites boosted case numbers ($129/subject) and the biobank provided additional controls at low cost ($5/subject), in large part due to the generosity of PMBB providing samples free of cost. Community outreach was costly with low return on enrollment ($978/subject for 220 subjects). Summer months (Jun-Aug) produced the highest recruitment yields (p<0.001). Genetic studies will benefit from a multi-pronged and culturally sensitive recruitment approach. In our experience, the biobank was most cost-effective for control enrollment, while recruitment from clinics (including expansion to new sites) was necessary to recruit fully phenotyped cases.

Fig 1. Monthly enrollment of glaucoma cases for the Primary Open-Angle African American Glaucoma Genetics study.

Monthly case enrollment declined over time, despite increased number of Clinical Research Coordinators (CRCs) to enroll subjects. The arrows denote five significant shifts in the enrollment process: (1) November 2014: Switch from blood samples to saliva samples for DNA collection, (2) May 2015: Addition of first external site (Windell Murphy, MD), (3) July 2015: Re-started enrollment at second external site (Temple University), (4) September 2015: Addition of samples from the Penn Medicine Biobank, and (5) March 2018: Addition of a fourth UPenn site (VA Hospital). Monthly enrollment is shown from 07/2010-07/2018, as after this time point, enrollment was intentionally slowed as CRC efforts were needed for other efforts such as phenotype collection or database management.

Fig 2. Seasonal trends in recruitment in the POAAGG study, displayed as average number of enrolled subjects/month per Clinical Research Coordinator (CRC).

The months of July and August saw the highest enrollment yield, while spring months showed the lowest yield. Recruitment from the biobank is excluded from this figure. Data from 07/2010-07/2018 is included, as after this time point, enrollment was intentionally slowed as personnel efforts were needed for other efforts such as phenotype collection or database management.