Thank you for visiting nature.com. You are using a browser version with
limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off
compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site
Data are presented as partial correlation coefficients (r) and unstandardized regression coefficients (β) with 95% CI using a linear mixed model with school as a random factor. Coefficients represent a change of 1 unit in FMI for a 100-cpm change in total physical activity, a 10-min change in time spent in MVPA, a 1% change in sedentary time, a 60-min change in screen time and sleep duration and a one-point change in CSHQ. All models were adjusted for baseline age, sex, pubertal status, sex–pubertal status interaction, month of baseline measurement and days of follow-up.
aFMI=total fat mass × height–2.
bAdditionally adjusted for the particular baseline movement behavior of interest and baseline FMI.
cAdditionally adjusted for baseline FM.
dAdditionally adjusted for the particular baseline movement behavior of interest.
Total tri-axial physical activity (100 cpm); change in behavior vs change in FMI (r=–0.03, β=–0.008, P=0.52), baseline behavior as predictor of change in FMI (r=–0.05, β=–0.15, P=0.14), baseline FMI as predictor of change in behavior (r=–0.16, β=–0.219, P<0.001).
Screen time and CSHQ were only obtained at baseline.