Individual and contextual effects on the risks of adult mortality in the United States by race and Hispanic origin

The purpose of this dissertation is to examine how individual and contextual level factors work together to shape race/ethnic differences in adult mortality outcomes. Existing research on adult mortality is limited in three important ways: (1) the majority of research has focused primarily on black/white differences; (2) most research has focused on only one level of analysis – ignoring multilevel effects; (3) the few studies that have been conducted using a multilevel approach have a number of methodological problems, such as the inability to isolate the independent contributions of individual and community level effects. Further, past research has not addressed how individual level and community level factors influence the mortality risks of multiple Hispanic subgroups. This dissertation research fills these gaps by accomplishing the following aims. First, this dissertation analyzes three Hispanic subgroups – Puerto Ricans, Mexican Americans, and other Hispanics – in addition to non-Hispanic whites and blacks. Second, this dissertation investigates multilevel effects of socioeconomic status variables such as individual and neighborhood-level income, education, and employment status on all cause adult mortality using multilevel models. Third, this dissertation investigates the utility of small contextual units of data aggregation – very small areas (VSAs) – for mortality research. The data are from the 1986-1994 National Health Interview Survey (NHIS) matched to deaths from the National Death Index (NDI) through 1997. The data analysis is conducted using Cox proportional hazards models and multilevel logistic regression models. Results from this dissertation research indicate that higher mortality risks are associated with not only individual socioeconomic, family status, and demographic characteristics, but also with neighborhood characteristics. Blacks, Mexican Americans, Puerto Ricans, and other Hispanics are more likely than whites to live in communities with high levels of unemployment, poverty, low educational attainment, as well as racial and ethnic segregation. Such neighborhood characteristics contribute to higher mortality risks. Individuals can expect to live longer lives if they live in neighborhoods with high income, access to quality education, and strong community ties. The findings of this dissertation suggest the need for social policies that can improve not only individual risk factors but also neighborhood characteristics.