Abstract P152: Parent-child Associations for Changes in Diet and Physical Activity Across Two Decades in Modernizing China

Abstract

Introduction: While the household context is important for lifestyle interventions to reduce cardiometabolic disease, few studies have examined parent-child associations for diet and physical activity (PA) changes over time in a rapidly urbanizing country with burgeoning cardiometabolic disease rates. Hypotheses: We tested the hypotheses that although changes in diet and PA are associated in parents and their children, the magnitude of this association will decline over time and changes in these behaviors over time differ in adults versus children. Methods: We studied dietary and PA behaviors in parent-child pairs (children aged 7-17y) using longitudinal data from the China Health and Nutrition Survey (survey years: 1991, 1993, 1997, 2000, 2004, 2006, and 2009). Sample size varied at each year and on average there were 1,692 mother-child and 1,594 father-child pairs across all visits. Three-day 24-hour recall diet data were collected to generate percentage of total energy from animal-source foods (ASF), eating away from home, and snacking. Seven day PA recall across domains of PA included screen-based activity hours per week, and metabolic equivalent hours (MET-hrs) per week from commuting and leisure-time sports. Spearman partial correlation was used to test parent-child association, adjusting for child’s age and household sociodemographics. We used the Kruskal-Wallis test to examine differences of these behaviors across years. All models were stratified by child’s sex and parent’s sex. Results: We found statistically significant positive parent-child correlations for all dietary variables (correlation coefficients ranged from 0.49 to 0.89 across all years and all parent-child combinations, p<0.01), with stronger mother-child compared to father-child correlation. Whereas, parent-child associations for PA were weaker in magnitude. Coefficients ranged from 0.1 to 0.34 for screen-based hours, 0.08 to 0.17 for commuting, and 0.07 to 0.16 for leisure-time sports (p<0.05). Associations for commuting and leisure-time sports became non-significant at later years. Children and adults experienced statistically significant increases in the percentage of energy from ASF, eating away from home, and snacking, as well as in screen-based hours over time. Parents experienced a decrease in commuting MET-hrs over two decades (p<0.05), whereas children had minimal change. Leisure-time sports MET-hrs increased in children, and decreased in adults over time (p<0.05). Conclusions: Parent-child associations were comparatively stronger for diet than for PA. We observed an increase in energy from ASF, eating away from home, and snacking in parents and children over time, increases in PA in children and declines in adults. Our findings support household-based health behavior intervention strategies, particularly around diet-related behaviors.