Predictors of Change in Objectively-Measured Physical Activity: One Year Results From the SPEEDY Study

Purpose To inform intervention development by investigating predictors of one-year change in objectively-measured intensity of physical activity (PA) in a cohort of 9-10 year old British school children.

Methods In a cohort of 753 children in Norfolk (UK) (41.7% male, baseline age (mean±SD): 9.7±0.4 years) PA was measured over 3 or more days by Actigraph accelerometer. Outcome measures were change in percentage time spent in moderate PA (MPA, 2000-3999 counts/min) and vigorous PA (VPA, ≥4000 counts/min), separately for weekdays and weekends. A total of 26 biological, psychological, socio-cultural and objectively-assessed environmental factors were assessed at baseline. Multiple multilevel linear regression was conducted, adjusted for baseline PA, and interactions with sex explored.

Results Weekend PA declined (MPA: -4.1±22.0 minutes; p<0.001 and VPA: -2.0±20.4; p=0.012) but not weekday PA. Baseline PA (all p<0.001) and sex (girls vs. boys, all p<0.02) were negatively associated with all outcome measures. Other predictors differed by activity intensity and time of week. For example, availability of PA equipment at home was positively associated with change in weekday VPA (ß: 0.23; 95%CI: 0.01; 0.44), but negatively with weekend MPA (ß: -0.54; 95%CI: -0.96; -0.12). Family logistic support was associated with change in weekend activity (MPA: ß: 0.15; 95%CI: 0.05; 0.25; VPA: ß: 0.18; 95%CI: 0.06; 0.30).

Conclusions Predictors of change in PA differ by PA intensity and time of the week and this should be taken into account when designing future interventions. Strategies aiming to prevent a decline in PA may be targeted at weekends and focus on socio-cultural factors.