Imputing accelerometer nonwear time in children influences estimates of sedentary time and its associations with cardiometabolic risk

Table 4 Sedentary time in the nonimputed and imputed datasets

Dataset	Mean (95% CI) min/day of sedentary time	Mean (95% CI) min/day of wear time	Mean (95% CI) percentage of wear time spent sedentary
Nonimputed dataset (n = 442)	599 (594, 604)	826 (821, 831)	72 (72, 73)
Imputed dataset I (n = 452)^a	632 (627, 637)	868 (865, 873)	73 (72, 73)
Imputed dataset II (n = 452)^b	632 (627, 637)	868 (865, 873)	73 (72, 73)

^aImputation based on sociodemographic (age, sex, maturity, race, family structure, parental education and household income), health (cardio-metabolic risk Z-score, internalizing symptoms Z-score, and presence of a chronic health condition), behavioural (frequency of fast food consumption and snacking in front of a screen), and time (time of day, type of day, and season)
^bImputation based on total screen time, homework time, presence of household media rules, and the number of screen-based devices in the household, in addition to the variables used in imputed dataset I

ISSN: 1479-5868