ecological momentary assessment; adherence; self-initiated surveys; physical activity

Background: Assessing the context of physical activity (PA) is challenging with objective methodsalone. One way to assess PA context is through self-initiated surveys. However, factorsinfluencing adherence to such surveys over time remain unclear. Understanding the factors thatunderly adherence to ecological momentary assessment (EMA) protocols is critical for evaluatingdata quality. This study investigates adherence patterns in a year-long EMA study. Methods: 12-month EMA study with four 14-day measurement bursts was conducted among1314 Czech adults aged 18-65. Participants were instructed to complete a self-initiatedsmartphone survey (via study app) after each purposeful PA throughout the 12 months. Tosupport adherence, participants received one reminder before each burst period. Kruskal-Wallistest was used to examine differences in the average number of completed surveys per participantby age groups (18–29, 30–49, 50–65), calendar month, and weekday. Wilcoxon rank-sum testevaluated differences by sex and burst period. Pairwise Wilcoxon tests were used for post-hoccomparisons between age groups. Linear mixed-effects model was applied to assess the effectof time on adherence, accounting for repeated measures and individual variability. Results: The study analyzed 77401 surveys from 1084 participants (mean=71.40, median=31.50,IQR=103.25). Fewer average surveys per participant were completed by young adults vs. bothearly (p<0.01) and middle- aged adults (p<0.01) and on non-burst days (p<0.01). No differenceswere found by sex (p=0.68), calendar month (p=0.44) or weekday (p=0.42). Over time, participantscompleted 0.24 fewer surveys per month (p<0.01) and 0.01 fewer per week (p<0.01). Conclusions: Adherence to self-initiated PA surveys in a year-long EMA study was similarbetween sexes and across all calendar months and weekdays but varied by age group, withadherence declining over time. Reminders seemed effective in increasing survey completionrates during burst periods. These findings are relevant for long-term EMA studies evaluating PAcontext.

Background: Assessing the context of physical activity (PA) is challenging with objective methodsalone. One way to assess PA context is through self-initiated surveys. However, factorsinfluencing adherence to such surveys over time remain unclear. Understanding the factors thatunderly adherence to ecological momentary assessment (EMA) protocols is critical for evaluatingdata quality. This study investigates adherence patterns in a year-long EMA study.Methods: 12-month EMA study with four 14-day measurement bursts was conducted among1314 Czech adults aged 18-65. Participants were instructed to complete a self-initiatedsmartphone survey (via study app) after each purposeful PA throughout the 12 months. Tosupport adherence, participants received one reminder before each burst period. Kruskal-Wallistest was used to examine differences in the average number of completed surveys per participantby age groups (18–29, 30–49, 50–65), calendar month, and weekday. Wilcoxon rank-sum testevaluated differences by sex and burst period. Pairwise Wilcoxon tests were used for post-hoccomparisons between age groups. Linear mixed-effects model was applied to assess the effectof time on adherence, accounting for repeated measures and individual variability.Results: The study analyzed 77401 surveys from 1084 participants (mean=71.40, median=31.50,IQR=103.25). Fewer average surveys per participant were completed by young adults vs. bothearly (p<0.01) and middle- aged adults (p<0.01) and on non-burst days (p<0.01). No differenceswere found by sex (p=0.68), calendar month (p=0.44) or weekday (p=0.42). Over time, participantscompleted 0.24 fewer surveys per month (p<0.01) and 0.01 fewer per week (p<0.01).Conclusions: Adherence to self-initiated PA surveys in a year-long EMA study was similarbetween sexes and across all calendar months and weekdays but varied by age group, withadherence declining over time. Reminders seemed effective in increasing survey completionrates during burst periods. These findings are relevant for long-term EMA studies evaluating PAcontext.