What This Means
This research examined how physical activity, recreational screen time (such as watching TV or playing video games), and dietary fat intake cluster together in Chinese children and adolescents, and whether these behavioral clusters were associated with weight status. Using data from nearly 4,100 observations of school-attending children aged 6-17 in the China Health and Nutrition Survey collected between 2004 and 2011, the researchers identified five distinct groups: a very low activity group, a low-to-moderate activity/lower fat group, a group reporting no entertainment screen time, a higher fat intake group, and a highly active group. Children in the very low activity group had the highest rate of overweight or obesity (about 24%), compared to about 16% in the reference group.
However, when the researchers adjusted for other factors, the differences in weight status between behavioral groups were no longer statistically meaningful. The adjusted odds ratio for overweight/obesity in the very-low-activity group compared to the reference group was 1.18 (meaning roughly 18% higher odds), but the confidence interval spanned from 0.81 to 1.69, indicating no statistically significant difference. A test comparing all five profiles together also showed no significant overall association (p = 0.806). Multiple different statistical approaches all reached the same conclusion. Additionally, nearly half of the children could not be clearly assigned to a single profile with high confidence, suggesting these behavioral combinations were not sharply distinct.
This research suggests that grouping children by combinations of physical activity, screen time, and dietary fat did not clearly identify subgroups with meaningfully different weight outcomes once other factors were taken into account, at least during this historical period in China. The authors emphasize that these findings reflect children surveyed between 2004 and 2011 and should not be used to draw conclusions about current patterns across China or to justify specific profile-based health interventions. The findings highlight the complexity of studying how multiple health behaviors relate to childhood weight status and point to the challenges in using behavioral clustering approaches for public health targeting.