Characterizing Physical Activity Level, Weight Perception, and BMI Among a Nationally Representative Sample of Adolescents: A Latent Class Analysis

Background and Purpose:  Although obesity rates are high among adolescents, current physical activity level recommendations are not being met.  Even so, society has placed an unrealistic thin ideal on adolescents, and negative self-image is detrimental during this critical period of development.  We used latent class analysis (LCA) to identify profiles of adolescents in the United States, with covariates, to identify predictors of class membership in an effort to better understand how students perceive their weight compared to their actual weight status and if this reflects a matching level of physical activity and/or desire to lose weight.

Methods:  We used nationally representative data (n=13, 515) to determine latent class profiles of adolescents based on physical activity level, weight perception, desire to lose weight, and BMI.  Data came from the 2011 Centers for Disease Control and Prevention’s Youth Risk Behavior Surveillance System (YRBSS).  For this study, physical activity levels were based on number of days in the past week adolescents reported being active for greater than 60 minutes and how many days they reported strength training activities.  Weight perception and desire to lose weight were self-reported.  Self-reported height and weight were used to calculate BMI (recoded as underweight, healthy weight, and overweight/obese).  LCA models were examined using Mplus 7.11.

Results:  Five latent classes of weight and activity were identified.  The largest proportion of youth were classified as high activity/healthy weight (25.8%, mean days physical activity (PA)=6.3) followed by the low activity/healthy weight group (25.3%, mean days PA=1.7).   The three remaining classes included the low activity/overweight (16.5%, mean days PA=1.6), moderate activity/healthy weight (15.9%, mean days PA=5.7), and moderate activity/overweight (16.5%, mean days PA=5.7) groups.  Using the low activity/healthy weight class as the reference category multinomial logistic regression revealed several predictors of class membership.  Males were 16 times more likely than females to be in the high activity/healthy weight class (OR=16.4, 95% CI=11.1-24.1).  African Americans and Latinos were less likely to be in the low activity/overweight class than whites (OR=0.32, CI=0.21-0.48 and OR=0.57, CI=0.40-0.83, respectively). 

Conclusions and Implications:  The results revealed disparities in activity level across students in healthy and overweight categories.  Amongst youth in the two classes that were predominantly likely to be overweight, one class was engaging in moderate average days of physical activity while the other was not, despite the fact that the majority of adolescents in the two “overweight” categories were likely to report wanting to lose weight.  Among the healthy weight categories, students reported both wanting to lose weight and desiring to stay the same weight.  More education among adolescents may be needed to promote healthy weight ideals and physical activity across BMI percentiles.