Validation of methods of measuring physical activity using The Observer and actiwatch in young children

C.P. Khokhar³, R. Juyall³, U. Dhingra², A. Dutta2 M. Osmany¹, R.E. Black², A. Sarkar¹, P. Dhingra ¹, P. Verma¹, M. Black⁴, R.M. Angulo-Kinzler⁵, V.P. Menon¹, J. Kumar¹, G.S. Hiremath² and S. Sazawal^1,2

¹Center for Micronutrient Research, Annamalai University, India
²Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore USA
³Department of Psychology, Gurukul Kangri University, Haridwar, India
⁴University of Maryland, School of Medicine, Baltimore, USA
⁵Division of Kinesiology, Center for Human Motor Research, University of Michigan, Michigan, USA

Physical activity and active exploration play an important intermediary role in a child’s developmental processes. Assessing the levels of physical activity of young children in free living conditions with non-invasive techniques remains a challenge and studies have used direct observation of the activity. Direct observation is costly, tedious and accuracy may decrease as the observation period lengthens. With the advent of smaller watches using motion sensors, measurement of physical activity has gained popularity. Validation studies in young children of active watch are lacking. We validated the Actiwatch activity monitor (Mini- Mitter Company Inc.) against direct observation using The Observer^® (Noldus Information Technology bv, The Netherlands). Among 100 children aged 13-76 months from peri urban population in Delhi, one hour of direct observation of activity at home while wearing an omni directional sensory monitor on wrist (W) as well as ankle (A) was undertaken. From direct observation, an overall activity score was estimated using proportion of time spent in each of the 5 activity levels (gross and .ne motor), and scoring using Children’s Activity Rating Scale (CARS). The activity counts from watch were correlated well with the CARS scoring W: ß 0.18, R2 0.03 (p=0.07), A: ß 0.30, R2 0.09 (p=0.003), C: ß 0.31, R2 0.099 (p=0.001)] and were not affected by age [W: ß -0.04, R2: 0.03 (p=0.68), A: ß -0.04, R2 0.09 (p=0.72), C: ß -0.05, R2 0.10 (p=0.63)]. When children were categorized into 4 quartiles based on CARS score, mean scores with watch across four categories showed a significant rising trend. Our results indicate that 1 hour observation CARS score for activity correlates with 1 hour activity counts of W and A, however, combined scores (W+A) reflected better association, validating the use of the activity monitors in assessing physical activity in young children.

Paper presented at Measuring Behavior 2005 , 5^th International Conference on Methods and Techniques in Behavioral Research, 30 August - 2 September 2005, Wageningen, The Netherlands.

© 2005 Noldus Information Technology bv