NIH Research Festival
FARE Award Winner
Background: Youth use different screen types (e.g., streaming, gaming) that may be uniquely related to body mass index (BMI). Time spent on screens is interconnected with other activities, such as physical activity and sleep, that affect BMI. Compositional data analysis (CoDA) models the association between interconnected behaviors and health outcomes. We used CoDA to examine how different screen types and other interconnected behaviors relate to BMI one year later.
Methods: Baseline and one-year follow-up data were from the Adolescent Brain Cognitive Development Study (N=5,165, mean [SD] baseline age=10.0 [0.6] years, 49% female, 48% non-White). Participants reported baseline screen time (streaming, gaming, and socializing), physical activity, and sleep. Sex-stratified linear regressions estimated the association between baseline relative proportion of time spent in each behavior and follow-up BMI z-score, adjusted for baseline BMI z-score, participant characteristics, and depressive/anxiety symptoms.
Results: Mean time (min/day) spent in different activities was 152.0 streaming, 72.7 gaming, 46.8 socializing, 29.7 physical activity, 534.9 sleep and 603.9 other. In males, a greater proportion of time spent in baseline socializing via screens, relative to time spent in the remaining behaviors, was related to a higher follow-up BMI z-score (Œ≤ [95% CI] = 0.05 [0.02 to 0.08]). No other significant associations were observed.
Conclusions: CoDA can advance our understanding of the distinct BMI implications of different screen types, independent of other relevant behaviors. Further investigation of possible biological and behavioral mechanisms underlying the observed sex differences in the association between screen time and BMI is warranted.
Scientific Focus Area: Social and Behavioral Sciences
This page was last updated on Monday, September 25, 2023