Neighborhood environment, stress, and obesogenic behaviors among adults
Barrington, Wendy E.
MetadataShow full item record
<bold>Background: </bold>Socio-ecologic models convey a conceptualization of how levels of environment surrounding individuals impact health behaviors and subsequent health outcomes. Neighborhood context, defined as area-level social or physical systems or structures, has been recognized as an important social determinant of many health outcomes across the lifespan. The overall goal of this dissertation was to evaluate the effects of worksite and home neighborhood contexts on biobehavioral pathways associated with stress. In particular, this dissertation evaluates whether workplace neighborhood context contributed to obesity-related behaviors, as most research to date has been in relation to the home neighborhood. It also examines how neighborhood contexts of the home and workplace influence levels of stress, a hypothesized pathway for the embodiment of contextual effectson chronic disease processes. <bold>Methods: </bold>Chapters 1 and 2 used Promoting Changes in Activity and Eating (PACE) data, which included approximately 2400 individuals at follow-up within randomly selected smaller worksites in Seattle, to evaluate associations. Chapter 3 evaluated associations among 510 non-Hispanic white women in Seattle who represented a broad range of individual and neighborhood-level socioeconomic status (SES). Chapter 4 evaluated associations among 541 older professional white men and women within the British civil service. To evaluate neighborhood and individual-level associations, all analyses used models that accounted for multiple sources of variance (i.e. random effects models, generalized estimating equations (GEE)) and included relevant covariates. <bold>Results:</bold> We found in chapter 1 that higher worksite-level SES was significantly associated with more walking for at least 10 minutes in the previous week among employees. Built environment attributes were also significantly and independently associated with obesity-related behaviors. Specifically, higher residential density around the worksite was significantly associated with walking and eating 5 or more servings of fruits and vegetables, independent of worksite-level SES. More food and activity destinations as well as more intersections around the worksite were also significantly associated with more walking before accounting for worksite-level SES. We also found that higher residential density mediated the association of higher worksite-level SES with eating 5 or more servings of fruit and vegetables and with walking. Chapter 2 evaluated associations between social and built environment attributes surrounding worksites, perceived global and work-related stress, and C-Reactive Protein (CRP), a biomarker of inflammation that is associated with stress. We concluded that worksite context was not significantly associated with perceived stress or CRP in these data. Worksite contextual variables were associated with work demands and worker social support, however, and these relationships varied by gender. Specifically, work demands varied significantly by both worksite-level SES and surrounding residential density among men. White-collar worksite class was also associated with fewer work demands and greater social support. Among women, a significant linear trend between higher residential density and greater work demands was demonstrated. Worker social support also varied significantly by residential density. Chapter 3 evaluated associations between home neighborhood SES and general perceptions of stress and found relationships which depended on individual-level SES. Specifically, measures of higher neighborhood SES were associated with lower stress among lower educated women whereas the opposite was suggested for higher educated women. Similar relationships were suggested when using family income as a measure of individual-level SES, albeit not as consistent. Strong associations between perceived neighborhood characteristics and stress were exhibited for all women. Specifically, greater neighborhood problems and dissatisfaction as well as lower perceived walkability were all highly significantly associated with greater stress. Chapter 4 evaluated associations between home neighborhood SES (i.e. income and employment deprivation), perceived neighborhood characteristics (i.e. social control, neighborhood threat, fear of crime), and stress reactivity (i.e. cortisol). Income and employment deprivation were significantly associated with lower social control, higher neighborhood threat and greater fear of crime after accounting for individual demographic characteristics. Fully adjusted models also indicated that higher levels of socioeconomic deprivation were associated with cortisol non-response, an indicator of a "blunted" cortisol profile which has been associated with chronic disease risk. Perceived neighborhood characteristics were also significantly associated with cortisol non-response independent of individual-level demographic covariates. Mediation analyses indicated that the association between neighborhood SES and cortisol non-response was partially mediated by social control and neighborhood threat. <bold>Conclusion: </bold>These studies contribute to the collective understanding of how environments may impact health behaviors and subsequent risk of disease among adults through the stress process. Importantly, findings indicate that multiple neighborhood contexts may be salient to biobehavioral processes associated with stress. Perceived measures of the neighborhood are important determinants of stress which not only work independently, but also mediate relationships between neighborhood-level SES and the stress process. These more modifiable factors may provide targets for addressing neighborhood disparities in health associated with stress.
- Epidemiology