Motivation-informed digital reminders do not improve task completion compared to standard wording in melanoma trialStandard Digital Reminders Work Well for Melanoma Monitoring Tasks

OBJECTIVES: Digital reminders are widely used in clinical trials, but their design is rarely evidence-based. We evaluated the effectiveness of two types of digital reminders for trial task completion. STUDY DESIGN AND SETTING: This randomized Study Within A Trial (SWAT) was embedded in the MEL-SELF trial which evaluates patient-led vs. clinician-led melanoma surveillance. SWAT participants were randomized 1:1 to motivation-informed or standard digital reminders for two tasks: monthly health-resource diaries (all participants) and 3-monthly image submissions of self-detected skin lesions (MEL-SELF intervention arm only). Primary outcomes were the per-participant proportion of diaries completed on time (within 14 days) and images completed on time (42 days). Secondary outcomes were on-time completion of the first diary and first scheduled image, any-time completion, time to completion, number of reminders required, and costs. RESULTS: Between 16 August 2023 and 3 June 2024, 260 participants were randomized to motivation-informed (n = 130) or standard (n = 130) reminders. On-time completion of the first diary was 82% (106/130) for both reminder types (risk difference (RD): 0.0%; 95% CI, -9.5% to 9.5%) and any-time per-participant diary completion was also high (median 100% both arms). Across 2821 diary submission tasks, participants received 5710 automated reminders (mean 2.0 per task), with 21.6% requiring staff follow-up; mean cost per completed diary was AUD 4.15 (Australian dollars). For images, on-time first submission was 51% (33/65) in the motivation-informed arm and 57% (36/63) in the standard arm (RD: -6.3%; 95% CI, -22.9% to 10.6%); per-participant any-time completion of scheduled image submissions was also modest in both arms (median 50% motivation-informed vs 67% standard). For 489 image episodes, there were 1929 automated reminders (mean 3.9 per episode); 56.0% required trial staff follow-up, and the mean cost per submitted image was AUD 17.43. In prespecified subgroups, differences between reminder arms were small, with diary completion higher in both SWAT arms for participants in the MEL-SELF control arm compared with the intervention arm. CONCLUSION: Motivation-informed reminders did not improve completion of trial tasks compared with standard wording. For low-complexity tasks, standard reminders may be sufficient; for high-complexity tasks, changing message content alone is unlikely to offset burden or logistical barriers. PLAIN LANGUAGE SUMMARY: Clinical trials often rely on participants to complete tasks such as filling in online diaries or submitting information at scheduled time points. Missing tasks can reduce the quality of trial data. Many trials use digital reminders to help people complete these tasks, but we do not know whether certain types of reminder messages work better than others. In this study, we tested whether reminders written to reflect people's motivations for joining a trial would improve task completion compared with standard reminders. We embedded this research into a melanoma surveillance trial. Participants received monthly reminders to complete online health diaries, and some also received reminders every 3 months to submit skin images. We compared motivation-informed reminders with standard reminders, and we measured how often participants completed these tasks on time, as well as the amount of staff time needed to support completion. We found that motivation-informed reminders did not increase task completion. Almost all participants completed their online diaries, regardless of the reminder wording. Image submissions were more challenging: only about half of participants completed them on time, and many required additional help from trial staff. These findings show that changing reminder message content did not meaningfully improve task completion. For simple tasks, standard reminders appear sufficient. For more complex tasks, improving the process itself (e.g., making image capture easier or providing more hands-on support) may be more important than changing the message wording.