Prevalence of Physical Activity Maintenance Across a 12-Month Study: Comparison of Accelerometer Indicators

Maintaining physical activity (PA) is critical for reducing disease risk. Yet, lack of consensus on how to define and operationalize PA maintenance has hindered surveillance efforts. We used longitudinal accelerometer data to compare how different ways of operationalizing PA maintenance impact PA maintenance prevalence estimates. Young adults (N = 173, ages 18-29) provided up to 12 months of PA data via smartwatch accelerometers. Nonsleep movement data were processed into 7-day moving averages of Monitor-Independent Movement Summary units. PA maintenance was operationalized using combinations of 3 accelerometer-based indicators: (1) threshold (ie, level of PA required: [5.0-20.0 Monitor-Independent Movement Summary-units/min], (2) duration (ie, time required above a threshold [7-70 d]), and (3) allowance (ie, time allowed below a threshold [0-40 d]). Outcomes included the prevalence of days, episodes, and number of participants classified into PA maintenance. Increasing PA thresholds led to larger changes in PA maintenance prevalence outcomes than increasing durations or allowances. Greater changes in PA maintenance outcomes were observed when increasing thresholds up to about 12 Monitor-Independent Movement Summary-units/minute and allowances up to 7 days than when increasing above those points. Changes in PA maintenance outcomes were consistent across the entire range of durations. Threshold emerged as a more influential determinant of PA maintenance prevalence than duration or allowance, with greater changes across the lower range of thresholds. Validating these accelerometer-based indicators is a critical next step for establishing consensus regarding PA maintenance classification that can guide population-level surveillance.