|Skip left side navigation and go to content||
Obesity and the Environment: The National Longitudinal Study of Adolescent Health (Add Health)
Penny Gordon-Larsen PhD for the Add Health Obesity and Environment Investigators
There are substantial racial and ethnic differences in patterns of obesity, chronic disease risk factors, morbidity and mortality that are related to disparities in socio-demographic and environmental stressors. The Obesity and Environment project focuses on disparity in environmental and socio-demographic resources, lifecycle changes in these contexts and resources, and how this disparity impacts obesity and obesity-related behaviors. This work provides a national perspective on the complex environmental, biological, behavioral, and social relationships underlying the obesity epidemic.
The Obesity and Environment project has produced several papers on obesity and its biological, sociodemographic, behavioral, and environmental correlates in the National Longitudinal Study of Adolescent Health (Add Health). Add Health is a nationally representative, school-based, prospective longitudinal study of adolescent health. The Add Health cohort includes 20,747 ethnically diverse US youths (grades 7-12), followed with multiple interview waves (1995, 1996, and 2001) into young adulthood.
Longitudinal changes in overweight. Longitudinal multivariate models were used to assess the association of overweight with initial (and one-year change in) physical activity, controlling for age, ethnicity, SES, urbanicity, smoking, and region (Gordon-Larsen, et al., 2002). Findings suggest differential sex and ethnicity effects, with greater impact of physical activity on overweight among white versus minority adolescents. A further analysis explored obesity incidence in the transition to adulthood. During a five-year period, obesity incidence was 12.7%, with 9.4% remaining obese, and 1.6% shifting from obese to non-obese. Incidence was highest among non-Hispanic black and Hispanic females (Gordon-Larsen, et al., 2004a). The Add Health analysis sample represents about 15.6 million 13-20 year old students at public and private schools in the US and indicates that more than 1.9 million adolescents became obese and an additional 1.5 million adolescents remained obese during the five-year study period.
Add Health physical activity/inactivity data. One descriptive paper (Gordon-Larsen et al. 1999) on physical activity and inactivity patterns in waves I and II of Add Health shows higher inactivity among minority adolescents, with the exception of Asian females. Another paper (Gordon-Larsen et al., 2004b) explored the longitudinal shifts in activity and inactivity between Waves I, II, and III. Findings indicate that a majority of individuals who were physically active as adolescents became relatively inactive as young adults. Declines in physical activity were greatest among females, particularly non-Hispanic black females (Gordon-Larsen, et al., 2004b).
SES and overweight. Gordon-Larsen, et al. (2003a) examined the extent to which ethnic differences in income and education account for disparities in overweight prevalence. Analyses indicated that changing only family income and parental education, without manipulating environmental factors, had a limited effect on overweight disparities. Ethnicity-SES-overweight differences were greater among females. Given that overweight prevalence decreased with increasing SES among white females and remained elevated and even increased among high SES black females, black-white disparity in overweight increased at highest SES. Conversely, disparity was smallest at highest SES for white, Hispanic, and Asian females. Among males, disparity was lowest at average SES. The benefits of increased SES, in terms of reduced overweight, shown among white adults was not present in other gender-age-ethnic groups.
Acculturation. Popkin and Udry (1998) investigated patterns of overweight by immigrant status and found large and statistically significant changes in overweight prevalence for foreign-born (relative to US-born) Hispanic and Asian children. A more detailed paper (Gordon-Larsen, et al., 2003b) on determinants of overweight, including neighborhood context, explored the differences in overweight in foreign-born (relative to US-born) Hispanic adolescents. Findings include important changes in dietary patterns as adolescents immigrate to the US, with healthier diets in more recent immigrants. In sum, these findings suggest a rapid assimilation to an increase in overweight-related behaviors from first to subsequent generations of US residence (Gordon-Larsen et al., 2003b).
Maturation and Breastfeeding. An analysis of the timing of sexual maturation in relation to overweight found early (relative to average) maturation nearly doubled the odds of overweight (Adair and Gordon-Larsen, 2001). Nelson et al. (in press) investigated the relationship between breastfeeding and overweight in the Add Health cohort and in the Add Health sibling pair sample (controlling for unmeasured genetic and/or environmental factors). Findings from the cohort data suggest that the odds of overweight decrease significantly (females only) as breastfeeding duration increases. However, sibling analyses indicate that this relationship may be attributable to unmeasured confounding related to mothersí choice to breastfeed or other childhood risk factors for overweight.
Environmental correlates of physical activity. Gordon-Larsen et al. (2000) used Wave I contextual data and self-reported use of community facilities to illustrate large and significant influence of contextual factors on physical activity and inactivity. These findings show a major impact of environmental factors, such as PE, community recreation center use, and neighborhood crime on being physically active. Importantly, these findings indicate that physical activity is more influenced by environmental factors, while sociodemographic factors had greater impact on inactivity. Another analysis investigated the inequitable distribution of physical activity-related resources and facilities in Add Health neighborhoods varying in SES and ethnicity (Gordon-Larsen et al., submitted). Physical activity resources and facilities in each of the >45,000 block groups within 5 miles of the participantsí residences (representing approximately 20% of the US census block groups) were surveyed. Findings show that high SES, low minority communities were significantly more likely to have various types of activity-related resources compared to low SES, high minority communities.
Future Directions. A major push for the Obesity and Environment project is the linkage of physical environment measures from existing databases to Add Health respondentsí geographic locations. A central task is the development of a diverse and detailed database comprised of time-varying modifiable environmental factors for waves I, II, and III using data sources, such as the US Geologic Survey, Census, and Department of Labor Statistics. Measures of access to recreation facilities, transportation options, and other data (e.g., crime, cost of living, land use, walkability, climate) will be added to the Add Health dataset. The ultimate aim of this work is to assess the impact of density and proximity of these resources on physical activity and obesity patterns.