After the intervention ends: Fitbit Use and Activity levels 2 years post intervention (Preprint)

BACKGROUND There has been a rapid increase in the use of commercially available activity trackers, such as Fitbit, in physical activity intervention research. However, little is known about the long-term sustained use of trackers and behavior change after short-term interventions. OBJECTIVE The aims were to use minute level data collected from a Fitbit tracker for up to 2 years after the end of a randomized controlled trial to examine patterns of Fitbit use and activity overtime. METHODS Participants in this secondary data analysis were 75 female breast cancer survivors who had been enrolled in a 12-week physical activity randomized controlled trial (RCT). Participants randomized to the exercise intervention (Full Intervention arm) received a Fitbit One, which was worn daily throughout the 12-week intervention, and then were followed for 2 years post-intervention. Participants randomized to the waitlist arm, after completing the RCT, received a Fitbit One and a minimal version of the exercise intervention (Light Intervention arm), and then were followed for 2 years post-intervention. Average and daily adherence and MVPA were compared between the two groups in the interventional and post-interventional periods using both linear and generalized additive mixed effects model. RESULTS Adherence to wearing the Fitbit during the 12-week intervention period was significantly higher in the Full Intervention arm than the Light Intervention arm (85% vs 60%, p<0.001). Average adherence was significantly lower for both study arms during the follow-up period than the intervention period; however, there were statistically different patterns of adherence during the follow-up period with the Light Intervention arm having steeper declines that the Full Intervention arm over time (p<0.001). Similar to the adherence results, mean minutes of Fitbit-measured MVPA was higher for the Full Intervention arm than the Light Intervention arm during the 12-week intervention period (mean MVPA: 27.89 min/day, SD=16.38 vs 18.35 min/day, SD=12.64, p<0.001). During the follow-up period, average MVPA was significantly lower than the 12-week intervention period for both the Full Intervention arm (21.74 min/day, SD=24.65; p=0.002) and the Light Intervention arm (15.03 min/day, SD=13.27; p=0.004). Although the mean MVPA in each arm was similar across the follow-up period (p=0.329), the pattern of daily MVPA was significantly different between the two groups (p<.001). CONCLUSIONS While adherence to wearing activity trackers and maintaining physical activities declined after completion of a 12-week exercise intervention, a more active interventional strategy resulted in greater wear-time and activity levels during the intervention and more stable patterns of adherence and activity in the long-term. An improved understanding of long-term maintenance patterns may inform improved exercise interventions that result in sustained increases in physical activity. CLINICALTRIAL ClinicalTrials.gov: NCT 02332876