Randomized Recruitment Trial: Analyzing Differences in Survey Response Rates and Email Campaigns for Recruiting Latinos.

INTRODUCTION

In the USA, the Hispanic/Latino population is the largest minority racial/ethnic group; yet, data have shown that they are underrepresented in trials. We conducted an experiment to assess the effect on response rate and cost of three email formats to engage Hispanic/Latino patients.

METHODS

Patients were randomly allocated to either Arm 1 ("Survey Research Center (SRC) Plain"), which was a plain email script with text only; Arm 2 ("SRC Formatted"), which included photos of Latino individuals; or Arm 3 ("Vendor Formatted"), which had a visual layout like Arm 2, but the email distribution was routed through a third-party's software. The overall response rate (RR1) was defined as the number of respondents who engaged with the survey (partial or complete response). The completed response rate (RR2) is the number of complete responses compared to patients invited to complete the survey. RR1 and RR2 were compared between arms using the Chi-Sq test. Analyses were performed in SAS v9.4 (SAS Institute; Cary, North Carolina).

RESULTS

In total, 89,778 participants were emailed. The mean age of respondents was 54.2 years (SD 14.5 years), and most respondents, 65.7%, were female, with no significant differences across all three arms. The overall response rate (RR1) for all 89,778 individuals was 2.3% (2004 responses). The overall completion rate was not significantly different across the three arms (69.6%, 66.9%, 68.4%, respectively, p = 0.589). Actual costs for implementing the three arms are proprietary. Using Arm 1 (Plain Text) as the reference, Arm 2 (SRC Formatted) costs were 1.5 times higher than Arm 1, and Arm 3 (Vendor formatted) costs were 30 times higher than Arm 1.

CONCLUSIONS

Although the costs differed substantially between Arms 1, 2, and 3 in this study, we found no statistically significant differences in response rate between arms amongst English and Spanish speakers. These findings suggest that investigators should avoid spending additional resources on email campaigns incorporating images or outsourcing to third-party vendors when contacting Hispanics/Latinos in English or Spanish for research-related studies and focus on other ways to help foster engagement amongst this population, including incorporating community partners and incorporating Hispanic/Latino cultural values.