Mixed Delivery Mode Reminder Strategy to Tackle Low Participation in an e-Cohort: Randomized Controlled Trial Nested Within the Le French Gut e-Cohort

Abstract Background Typically, e-cohorts are susceptible to low participation rates, undermining representativeness. Frequent reminders can be a cost-effective strategy to increase response rates. However, their effectiveness may vary depending on the delivery mode and message content. Objective This study aimed to evaluate different reminder strategies (ie, emails and/or text messages with standard and/or institutional wording) and compare their effectiveness in increasing the response rate in a population-based e-cohort of healthy adults. Methods We conducted a 4-arm parallel randomized controlled trial nested within the Le French Gut e-cohort (registration number 2021-A01439-32). In November 2024, we included participants who were enrolled online (SKEZIA platform) but were not yet active participants (ie, eligibility, consent form, and/or personal information questionnaire had not been completed). Computer-generated randomization was stratified by time since enrollment. We sent 3 reminders, 72 hours apart, following 4 experimental designs: group 1, standard emails only; group 2, text messages only; group 3, institutional emails only; and group 4, standard email, followed by a text message and an institutional email. The primary outcome was the 30-day completion rate of the personal information questionnaire. We also measured completion time (ie, time between reminder received and questionnaire fully completed), online login rate, login time, email opening, and click-through rate. Participants were unaware of the trial, and investigators were blinded to group allocation. Results At the end of the trial, of the 20,487 eligible participants, 19,525 received at least 1 reminder (n=5225 in group 1; n=3858 in group 2, because n=1365 of the initial 5223 had no phone number and could not receive text messages; n=5222 in group 3; n=5220 in group 4). The per-protocol completion rate was 8.4%, with a higher rate in group 4 (9.4% vs 7.4%, 8.4%, and 8.3% for groups 1, 2, and 3, respectively; P <.001). Completion time was faster in group 2 (mean 6.3, SD 4.3 d) compared to groups 3 and 4 (mean 7.0, SD 3.5 and mean 7.2, SD 3.8 d, respectively; P =.003). The online login rate was higher, and the login time was faster in group 4 (login rate: 15.6% vs 12.1%, 14.3%, and 15.3% for groups 1, 2, and 3, respectively; P <.001; and login time: mean 7.1, SD 4.3 d vs mean 6.7, SD 4.3 d, mean 6.2, SD 4.6 d, mean 6.8, SD 3.9 d for groups 1, 2, and 3, respectively; P =.002). For emails, opening rates were similar ( P =.87), but the click-through rate was higher for institutional emails (23.1% vs 19.6% for standard formatting; P <.001). Conclusions For the first time in a specifically designed randomized controlled trial, we directly compare email-only, text message–only, and a combined reminder strategy in a large e-cohort. While previous studies focused on specific populations, our work examines a general adult population, demonstrating that a mixed delivery mode strategy effectively increases response rate by 27% compared to a single-channel approach. Text messages yielded faster responses, and institutional emails, with formal wording and coordinator’s signature, seemed more appealing than less formal wording. These results provide concrete guidance for researchers seeking to improve recruitment in real-world e-cohorts.