Comparing perceived and objectively measured changes in physical activity and sedentary time after a 12-week Internet-based intervention

![CDATA[Purpose: Evidence suggests that physical inactivity and sedentary time (ST) are independent risk factors for poor health. While there have been many successful interventions targeting these lifestyle behaviours, few studies have examined the relationship between perceived and objective changes in physical activity (PA) and ST. The purpose of this study was to compare participants' perceived changes in PA and sedentary time ST with objectively measured changes in PA and ST after a 12-week Internet-based intervention to increase PA and reduce ST in middle-aged male university employees. Methods: ManUp UWS was a 12-week, two-arm randomised controlled trial comparing an Internet-based intervention (receiving access to the ManUp UWS website and printed materials on PA, ST, and health) to a printed resources comparison group (receiving only printed materials). PA and ST were objectively measured using the ActiGraph GT3X+ accelerometer. Due to poor compliance with accelerometer wear, data were included for all participants who had a minimum of10 hours wear time per day, for at least 3 days (n=35). Self-reported changes in PA and ST were assessed using a process evaluation questionnaire. Agreement between the two measures was assessed using Cohen's kappa coefficient. Results: The mean (±SD) age of these 35 participants was 49.0y (±7.4), the mean BMI was 30.3kg/m2 (±5.0) and 71.4% of participants were non-academic staff. At end of intervention, 25/35 participants (71.4%) reported that their PA had "increased" since baseline, however, objectively measured accelerometer data did not reflect these changes, with only 8/35 participants (22.9%) increasing their PA (kappa coefficient= .119). At end of intervention, 15/35 (43.0%) reported that their ST had "decreased", however, only 7/35 participants (20.0%) reduced ST (kappa coefficient= .250). Conclusions: Participants' perceived changes in PA and ST were overestimated and not reflective of the objectively measured changes in these behaviours. Objectively measured data is less prone to measurement error and may provide a more accurate representation of an individual's PA and ST, however, it is important to examine perceived changes in lifestyle behaviours to provide a more complete picture of intervention effectiveness.]]