Children's Accelerometer-Determined Sedentary Time Reliability: A 12-Country Study

• Published in: Research quarterly for exercise and sport 2016-06, Vol.87 (S2), p.A35
• Main Authors: Barreira, Tiago Veloso, Schuna, John M, Tudor-Locke, Catrine, Chaput, Jean-Philippe, Church, Timothy S, Fogelholm, Mikael, Hu, Gang, Kuriyan, Rebecca, Kurpad, Anura, Lambert, Estelle V, Maher, Carol, Maia, José A, Matsudo, K R, Olds, Timothy, Onywera, Vincent, Sarmiento, Olga L, Standage, Martyn, Tremblay, Mark S, Zhao, Pei, Katzmarzyk, Peter T
• Format: Article
• Language: English
• Subjects: Academic Achievement; Accelerometers; Child Health; Children & youth; Females; Generalizability Theory; Girls; Males; Obesity; Physical Fitness
• ISSN: 0270-1367; 2168-3824

Previous accelerometer-derived sedentary time (SED) reliability research has been limited to single small samples. The International Study of Childhood Obesity, Lifestyle and Environment was conducted at sites from 12 countries across the globe and gives a unique opportunity to study reliability in a number of samples. Thus, the purposes of this multicountry analysis were: (a) to determine the obtained reliability of SED estimates, and (b) to estimate theminimumnumber of days necessary to achieve desirable reliability ($0.8) for SED. Children were asked to wear the accelerometer at the waist for 7 consecutive days (24 hr/day) and remove it only when in contact with water. Time in SED was objectively assessed using the ActiGraph GT3X þ accelerometer (# 100 activity counts/min). SAS PROC GLM was used to perform the intraclass correlation (ICC; 2, 1) computational method. Generalizability theory using R software was used to estimate the number of days required to achieve a desirable reliability estimate (G $.8) for each site separated for boys and girls. A total of 6,548 children (54% girls) aged 9 to 11 years old provided valid accelerometer data ($4 days including 1 weekend day with $10 hr of wake wear). The average number of valid days across sites varied from 6.0 ± 0.9 to 6.9 ± 0.4 for boys and girls, respectively. SED ICC ranged from 0.67 to 0.79 for boys and 0.64 to 0.80 for girls. The minimum number of days required to achieve a G$.80 ranged from 7 to 11 for boys and 6 to 7 for girls. An ICC $.80 was achieved only for girls at the China site; however, in only 3 instances, the ICC was less than .70. In only 6 instances, the number of required days was #7. Measuring SED for only 4 days, which is common in pediatric accelerometry research, will not provide reliable children's SED estimates. These results indicate that it is unlikely for researchers to realize high levels of reliability for SED measurement with a period of only 7 days of data collection because it is rare to obtain complete valid data from all participants.