Patterns and Determinants of Physical Activity in U.S. Adolescents
Present address of Dr. G. A. Colditz: Washington University School of Medicine, St. Louis, MO.
Received 1 July 2007; accepted 22 November 2007. published online 20 February 2008.
Refers to article:
Physical Activity in Adolescents: From Associations to Interventions
John R. Sirard, Daheia J. Barr-Anderson
Journal of Adolescent Health
April 2008 (Vol. 42, Issue 4, Pages 327-328) Full Text |
Full-Text PDF (67 KB)
Abstract
Purpose
The objectives of this study were to describe longitudinal trends in adolescent physical activity in a sample of U.S. adolescents and to assess the effect of multiple individual, parental, and environmental factors on initial level and rates of change in adolescent physical activity.
Methods
Study subjects were 12,812 boys and girls 10 to 18 years of age who were participating in the Growing Up Today Study and their mothers. We used accelerated longitudinal analysis to describe trajectories of physical activity from 1997–1999, and random effects linear mixed models to determine which factors were independently associated with baseline physical activity and changes in physical activity over time.
Results
Mean hours of physical activity ranged from 7.3–11.6 hours per week in boys and from 8.0–11.2 hours per week in girls. Physical activity was best modeled as a quadratic function of age, increasing until early adolescence and declining after age 13 in boys and girls. Multivariable modeling demonstrated that variables associated with physical activity level at baseline in boys and girls were age, body mass index, psychosocial variables, personal attitudes about body shape, perceived peer attitudes about body shape/fitness, parental attitudes about physical activity, parental physical activity, and environmental barriers to physical activity. Age was the only factor that predicted change in physical activity over time.
Conclusions
Interventions to increase physical activity in adolescents should begin before adolescence. Interventions may be more effective if they are multimodal and focus on modifiable individual, parental, and environmental factors.
ABSTRACT