The Family Life, Activity, Sun, Health, and Eating Study: Updated data resources for analyzing eating and physical activity behaviors among parent-adolescent dyads

Abstract

Objective: The Family Life, Activity, Sun, Health, and Eating (FLASHE) Survey was conducted in 2014 to collect information on eating, physical activity, sedentary behaviors, and behavioral correlates from a national sample of parent-adolescent dyads in the United States. FLASHE is a publicly available data resource offering opportunities to investigate research questions about psychosocial, generational, household, and neighborhood correlates of health behaviors (https://cancercontrol.cancer.gov/brp/hbrb/flashe.html). This presentation describes additional FLASHE data resources that can enhance analyses of the earlier survey datasets and are relevant to ISBNPA's research interests. Methods: Three supplemental FLASHE datasets include: a 'geoFLASHE' dataset, a multiple imputation dataset, and an adolescent accelerometer dataset. The geoFLASHE project used parent-provided address information for their home and adolescents' school to geocode these locations and compute a set of variables applied to several different neighborhood definitions, including both circular and streetnetwork buffers with distances ranging from 400, 1200 meters. The multiple imputation dataset was produced using weighted sequential hotdeck imputation to address a high missing data rate for eight items in the parent physical activity survey. A dataset of accelerometer variables was computed for a subset of adolescents who wore an Actigraph GT3X+ for seven days and includes estimates from raw and activity counts data. Results: The resulting geoFLASHE dataset includes variables for neighborhood socioeconomic status index, factor scores for three built environment characteristics (high density, older neighborhoods, and short commutes), and other contextual variables, for each of the buffer configurations. The multiple imputation dataset allows researchers to analyze data on parent-reported life goals for their child and/or physically limiting health conditions with sufficient sample size. The adolescent accelerometer dataset provides several key summary variables of accelerometer data and minute-level estimates of light, moderate, and vigorous activity using Crouter, Chandler, and GGIR processing methods. Conclusions: These tools offer opportunities for researchers to apply objective and imputed data in their studies of diet and physical activity behaviors among families. These resources complement the public use survey data and can be accessed and used in a variety of individual and dyad level analyses to understand activity and nutrition within families.