Fellow University of Rochester Medical Center Rochester, New York, United States
Background: Engagement of diverse patient samples in clinical research is critical to produce generalizable data. To facilitate research recruitment, an institutional developmental disabilities research registry was established in 2010. The registry maintains demographic and contact information provided by families after an informed consent process and enables investigators to share approved research opportunities. However, demographic distribution of the registry is not fully representative of the diversity of our clinic population. Objective: Here, we aimed to increase registry enrollment of non-white and/or Hispanic participants by at least 20% over the span of 12 months. Design/Methods: We first identified key drivers of underrepresentation, including 1) possible bias in who is offered participation and 2) patient barriers in completion of the enrollment survey. Across iterative plan-do-study-act(PDSA) cycles, we tracked the percentage of participant records with complete demographic data as a process measure, and racial and ethnic distributions as outcome measures. At baseline, racial breakdown of registry participants with completed records was(n=661): White-83.4%, Black/African American 6.8%, Asian 2.6%, Native American 0%, Native Hawaiian/Pacific Islander 0.2%, >1 race 7.1%. The ethnic breakdown was: Hispanic 8%, Non-Hispanic 82%, Unknown/Unreported 9%. Results: In PDSA 1, we aimed to reduce ascertainment bias by standardization of enrollment offered to all new patients seeking care through our clinic. This increased the proportion of non-white participants enrolling (a trend which is ongoing). At time of submission, the racial breakdown among those who enrolled with completed records following this standardization was(n=144): White 68.8%, Black/African American 9%, Asian 0%, Native American 0%, Native Hawaiian/Pacific Islander 0%, >1 race 16.7%. There is currently insufficient data to assess the impact on ethnicity. Despite this improvement, there remained a high percentage of participants (44%) who did not submit a full demographic survey to complete enrollment. We next conducted multiple cycles exploring different methods (email, phone, newsletters) and times of the day/week for outreach. No single method of participant outreach facilitated enrollment survey completion for all participants, but including different methods was needed to reach a wider percentage.
Conclusion(s): When recruiting from a diverse population, a variety of outreach methods are needed to foster full engagement, though standardization of enrollment offer at intake is a beneficial first step in promoting a more inclusive registry.
