Validity, Practical Utility, and Reliability of the activPAL™ in Preschool Children

DAVIES, G., J. J. REILLY, A. J. MCGOWAN, P. M. DALL, M. H. GRANAT, and J. Y. PATON. Validity, Practical Utility, and Reliability of the activPAL (TM) in Preschool Children. Med. Sci. Sports Exerc., Vol. 44, No. 4, pp. 761-768, 2012. Purpose: With the increasing global prevalence of childhood obesity, it is important to have appropriate measurement tools for investigating factors (e.g., sedentary time) contributing to positive energy balance in early childhood. For preschool children, single-unit monitors such as the activPAL (TM) are promising. However, validation is required because activity patterns differ from adults. Methods: Thirty preschool children participated in a validation study. Children undertaking usual nursery activity while wearing an activPAL (TM) 1 h were recorded using a video camera. Video (criterion method) was analyzed on a second-by-second basis to categorize posture and activity. This was compared with the corresponding activPAL (TM) output. In a subsequent substudy investigating practical utility and reliability, 20 children wore an activPAL (TM) for seven consecutive 24-h periods. Results: A total of 97,750 s of direct observation from 30 children were categorized as sit/lie (46%), stand (35%), and walk (16%), with 3% of time in non-sit/lie/upright postures (e.g., crawl/crouch/kneel-up). Sensitivity for the overall total time-matched seconds detected as activPAL (TM) "sit/lie" was 86.7%, specificity was 97.1%, and positive predictive value was 96.3%. For individual children, the median (interquartile range) sensitivity for activPAL (TM) sit/lie was 92.8% (76.1%-97.4%), specificity was 97.3% (94.9%-99.2%), and positive predictive value was 97.0% (91.5%-99.1%). The activPAL (TM) underestimated total time spent sitting (mean difference = -4.4%, P < 0.01) and overestimated time standing (mean difference = 7.1%, P < 0.01). There was no difference in overall percent time categorized as "walk" (P = 0.2). The monitors were well tolerated by children during a 7-d period of free-living activity. In the reliability study, at least 5 d of monitoring was required to obtain an intraclass correlation coefficient of >= 0.8 for time spent "sit/lie" according to activPAL (TM) output. Conclusions: The activPAL (TM) had acceptable validity, practical utility, and reliability for the measurement of posture and activity during free-living activities in preschool children.