Sherry Owens

Date of Graduation

2016

Document Type

Dissertation

Degree Type

PhD

College

School of Public Health

Department

Epidemiology

Committee Chair

Haslyn Hunte

Committee Co-Chair

Jamison Conley

Committee Member

Robert Duval

Committee Member

Michael Mann

Committee Member

Douglas Myers

Abstract

Neighborhood Disadvantage is the sum of a series of socioeconomic indicators that triangulate disadvantaged living conditions in a neighborhood (i.e., poverty, unemployment, home vacancies, female headed households, educational attainment, and segregation). Associations between Neighborhood Disadvantage (ND) and health outcomes have been widely explored. However, findings on mental and behavioral health outcomes remain inconclusive, and individual-level covariates often attenuate relationships between ND and health outcomes. This inconclusive evidence has denied researchers the opportunity to focus on neighborhoods as points of intervention. A potential reason for inconclusive findings is that neighborhoods and individuals are both subject to macro-level socioeconomic events, such as segregation and deindustrialization (loss of manufacturing jobs). This may generate between-level multicollinearity issues in traditional multilevel models because both neighborhoods and individuals are subject to these macro-level events. Instead, this dissertation focused on classifying neighborhoods based on their socioeconomic trajectories. A "trajectory" was defined as the changes in ND scores that occurred in a neighborhood over time. The objective of this dissertation was to determine whether ND trajectories from 1970 to 2000 were associated with residents' health outcomes in a 2001-2003 study of Chicago residents. This method intended to capture the types of socioeconomic influences, such as segregation and deindustrialization, which may have contributed to variation in neighborhoods' health resources and social norms in later years. Residents' health outcomes were compared across trajectories. This approach was compared to the traditional multilevel model used to analyze associations between ND and health outcomes.;In the first method, The Long-Term Census Tract Database (LTDB) was used to create the Neighborhood Disadvantage. I employed a latent profile analysis of ND scores across 343 Neighborhood Clusters in Chicago from 1970-2000. A multiple-regression was performed to investigate the association between Neighborhood trajectory classifications and three outcomes: depressive symptoms, smoking, and drug dependence symptoms among Chicago Community Adult Health Study (2001-2003) participants (n=3,105). The adjusted models indicated that: residents in Long-Term (LT) Very Disadvantaged and LT Inequality trajectories had significantly greater depressive symptom scores than the LT Advantaged trajectory. Residents of Declining trajectories were 1.66 (95% CI: 1.08-2.55) times more likely to smoke compared to the LT Advantaged trajectory. Residents of LT Very Disadvantaged trajectories were 3.25 times more likely to suffer from drug dependence symptoms than the LT Advantaged trajectory (95% CI: 1.32-8.05). The second method was a mixed-effects multilevel analysis of year 2000 ND and each aforementioned health outcome. ND was significantly and positively associated with depressive symptoms, not associated with drug dependence symptoms, and was negatively associated with smoking in the unadjusted and adjusted models. Overall, the Neighborhood trajectory method for identifying neighborhoods as points of intervention shows promise. Future work may investigate whether the neighborhood trajectory technique is useful in identifying neighborhoods for specific health interventions.

Download

Share

COinS