Descriptive Study Of Accelerometer-derived Measures Of Sedentary Behavior In Middle-aged Women

Abstract

PURPOSE: To characterize accelerometer-derived indicators of sedentary behavior collected over 5 consecutive weeks in middle-aged women representing pre- to post-menopause. METHODS: Data were obtained from the Evaluation of PA Measures in Middle-Aged Women (PAW) Study and included 62 women (93.9% of total sample) with 5 consecutive weeks of valid (i.e., ≥ 4 of 7 days with ≥ 10 hours of wear time/day) accelerometer (ActiGraph GT1M) data. Data were collected using 60 s epochs and indicators of sedentary behavior included: a) sedentary time (minutes/day), b) number of breaks in sedentary time/day, and c) number of breaks in sedentary time/day adjusted for sedentary time. Sedentary time (min/d) was quantified as the sum of activity counts between 0-100 and sedentary breaks were defined as an interruption in which a period of sedentary time was immediately followed by a minute or more above 100 counts. Repeated measures ANOVA using PROC MIXED (SAS/STAT software, v. 9.2) were used to describe day-to-day variation in sedentary time measures over 5 weeks of data collection. RESULTS: Participants were 52.7±5.4 years, 85.5% non-Hispanic white, 62.9% post-menopausal, with a BMI of 26.5±4.7 kg/m2. When compared to weekdays, mean sedentary minutes/day and number of breaks in sedentary time/ day were significantly lower on Saturday when compared to weekdays (all p<.05). Similar differences were noted when Sunday was compared to weekdays (all p<.0001). When weekend days were compared, the average number of breaks in sedentary time/ day were significantly higher on Saturday (p=.002). After adjusting the number of breaks in sedentary time per day by sedentary time (hr/d), differences between week- and weekend- day were no longer statistically significant. CONCLUSIONS: These data suggest that average sedentary time is higher on week vs. weekend days, which provided more opportunities to accumulate transitions from sedentary to active behavior. Day-to-day variation attenuated when breaks were adjusted for sedentary time. These data suggest that accelerometer-derived estimates of patterns of sedentary behavior are more stable after adjustment for sedentary time and have important implications for researchers interested in a) quantifying sedentary behavior and b) establishing associations with health outcomes.