Tuberculosis (TB) infects over 9 million people annually and is responsible for approximately 1.5 million deaths each year. Despite high treatment success rates, declines in incidence have averaged only 1.5% per year over the past decade. The newly adopted End TB Strategy sets a progressive agenda that moves beyond the medical sphere and includes a focus on upstream social determinants aimed at prevention. The strategy specifically calls for stronger poverty alleviation and social protection programs and policies. This dissertation sets out to support the need for social protection programs to combat TB, as well as attempting to answer the questions of who should be the recipients of such programs and in what form should they take.

Paper one uses country-level data to show that spending on social protection is associated with decreased TB prevalence, incidence, and mortality. This work has a global purview and predicts a drop of 18 per 100,000 persons in TB prevalence rate from 1% increase in social protection spending. This is true even after adjusting for factors associated with TB rates such as level of economic development, the strength of the health system, and HIV burden.

The second paper examines the relationship between household poverty and TB disease. Using household asset and characteristic data, this work creates household socioeconomic quintiles and attempts to show the relationship between this measure and individual active TB disease. The work utilized eight national TB prevalence surveys; large household surveys with a rigorous diagnostic TB algorithm. Although this approach found lower risk of TB disease for individuals in the poorest quintiles in four countries, a dose-response relationship was not observed. This paper also created an absolute wealth estimate, a US dollar-based measure of household wealth, which allowed for comparability across settings and also pooled country models. This measure of household socio-economic level did not have a clear association with individual TB risk. This work suggests novel ways of assessing the relationship between poverty and TB at the individual level that have the potential to be more efficient and to further the field of the social determinants of TB.

The final paper of this dissertation focuses on a TB patient cost survey in Myanmar. This nationally representative survey includes 966 TB patients across the country and gathers information about their income and costs while seeking care. This survey is the first step in measuring the percentage of TB-affected households experiencing catastrophic costs due to the disease, defined as costs exceeding 20% of annual household income. This metric is one of the high level indicators in the End TB Strategy and will be measured by the World Health Organization and its partners in the majority of high burden countries in the coming years. The Myanmar survey found 65% of households experiencing catastrophic costs due to TB, with major cost drivers being patient’s time and additional food and/or nutritional supplements required because of the disease.

Together these three papers support the need for social protection, alongside appropriate and timely medical care, in order to reach the ambitious targets set forth by the End TB Strategy.

Community-based participatory research (CBPR) is a paradigm for developing partnerships to address health disparities, however few studies have examined the role of CBPR principles in contributing to organizational capacity and leadership development, especially among community-based organizations (CBOs) in a partnership. A keener understanding of influencing factors for organizational capacity and development helps an organization to strengthen its capacity and sustain services. The Orange County Asian and Pacific Islander Community Alliance, a nonprofit organization based in Orange County, CA, has been engaged in CBPR partnerships for over 15 years. While they have engaged in the CBPR paradigm, they have not had the opportunity to evaluate its impacts on their partnerships and organization. The study examined PATH for Women, a breast and cervical cancer disparities collaborative framed in CBPR principles. This was a qualitative cross-sectional study focused on PATH for Women. The study explored CBPR principles, infrastructure, relationships and leadership and their associations to organizational capacity and leadership development. Key informant semi-structured interviews were conducted with PATH for Women partner organizations. The study revealed that the nine core CBPR principles were present and promoted in the partnership contributing to individual and organizational capacity building and leadership development. In particular the principles of building on strengths and resources, facilitating collaboration, and recognizing the community as a unit of identity were highlighted as important facilitators for organizational development. The study also revealed positive correlations between facilitating collaboration, integrating knowledge and action for social change, and the empowering process to address social inequalities with organizational visibility. In addition, the CBPR principles of an empowering process for social inequalities, integrating action for social change and facilitating collaboration were also positively correlated with leadership development. Increased skills, leadership development, increased and additional funding support, increased and leveraged relationships and networks, organizational recognition and visibility, and the ability to sustain programs and meet mission were described. The study identified that the PATH for Women partnership, framed in CBPR principles, provided growth and development for the participating organizations. Recommendations for consideration in implementing CBPR principles and promoting organizational capacity and development were also provided for consideration by community-based organizations.

Social inequalities can produce disparities in healthcare access and quality. This dissertation explores relationships between two social stratification processes- community residential segregation and social capital- on the supply of U.S. urban safety net providers.

The first paper, "Community residential segregation and the local supply of Federally Qualified Health Centers," used data from the Area Resource File and the U.S. Census to examine growth in FQHCs in urban counties from 2000 to 2007. Residential segregation by poverty and race/ethnicity were measured using the dissimilarity index. Logistic and negative binomial regression models were used for dichotomous and count outcomes, respectively. Residential segregation measures were associated with both county FQHC supply at baseline and the addition of new FQHCs over time. Residential segregation may produce geographic segregation of health services, such that FQHCs may be required to fill the gaps arising from provider maldistribution.

The second paper, "Residential segregation and the survival of U.S. urban public hospitals," used data from the American Hospital Association Annual Survey from 1987 to 2007. Cox proportional hazards models were used to estimate competing risks of hospital closure versus privatization. Poverty rates, intermediate poverty segregation, a low proportion of black residents, and low black residential segregation were associated with closure. Poverty associations suggest that areas with a high need for safety net services may be at risk to lose them, but segregated black communities may successfully advocate for maintenance of public hospitals. In contrast, Hispanic residential segregation was associated with privatization. Areas with segregated Hispanic communities may be less inclined to support public provision of services and have reduced opposition to privatization.

The third paper analyzed the same sample of urban public hospitals in relation to measures of community social capital. Voting rates were associated with closure, whereas bridging social capital among elites was associated with privatization. The findings suggest that social capital among privileged groups bears more influence on public hospital outcomes than vertical connections between the disadvantaged and those in power.

Taken together, the three papers suggest that social determinants may dictate both the need and societal response for the safety net.

The Earned Income Tax Credit (EITC) is the largest poverty alleviation program for American families, yet one in five do not take it up. Everyone who is eligible has a low income, so why are some people able to overcome structural and personal barriers to EITC access while others are not? Building on and beyond Bourdieu’s cultural capital, I argue for a dialectical, intersectional, and critical quantitative analysis that looks at structure, agency, and how they inform each other to study social (im)mobility through within-group differences. Using the Survey of Income and Program Participation (SIPP) data from the 2018 panel covering 2017–2020, cultural capital as operationalized was not a significant predictor of uptake in the main structural model; however, citizenship is. Furthermore, gender is a significant modifier in exploratory moderation analysis, highlighting the need for further work on quantifying intersectionality; incorporating other Bourdieusian concepts such as habitus, shadow capital, and fields; and more asset-based capitals including community cultural wealth.

Background: Cancer is the second leading cause of death in the United States, accounting for 22% of all deaths among women. Despite the importance of the Pap test in preventing and detecting cervical cancer, screening rates among poor women remain low. The pathways linking poverty with lower Pap test use remain unclear. The screening disparity for this low-cost test suggests that poor women may face other transaction costs, opportunity costs and barriers in accessing Pap tests that are not faced by women living above poverty.

Data/Methods: The data set merged the 2005 California Health Interview Survey, 2000 Census, and safety net clinic data. I examine elements of the patient-provider interaction (racial/ethnic discrimination) and healthcare system (distance to the nearest safety net clinic) as possible mediators between individual-level poverty and Pap tests. I then examine the role of individual-level poverty in moderating the association between discrimination and Pap test use and safety net distance and Pap test use. I also examine whether the neighborhood-level poverty moderates the association between individual-level poverty and Pap tests, discrimination and Pap tests, and distance and Pap tests. The sample included women, 18-64, with a physician visit in the previous 5 years and no history of cervical cancer. Mediation was tested using Generalized Estimating Equations (GEE) and a weighted product-of-coefficients test. Moderation was tested using GEE and including interaction terms.

Results: Poor women were significantly less likely to report a timely Pap test, had significantly higher odds of reporting discrimination, and were more likely to reside closer to a safety net clinic. Discrimination did not predict Pap tests and was not a mediator between individual-level poverty and Pap tests. Safety net distance predicted Pap tests and was a mediator between individual-level poverty and Pap tests. Neither individual-level poverty nor neighborhood-level poverty modified the proposed relationships with Pap test use.

Conclusion: Racial/ethnic discrimination did not explain disparities in Pap tests. Proximity to a safety net clinic was shown to be protective against disparities and may play a role in addressing disparities in screening. Individual-level poverty and neighborhood-level poverty did not influence the role of proposed mediators in predicting Pap tests.

Gentrification is a dynamic process that changes the physical, economic, social, and cultural characteristics of historically underserved neighborhoods. This neighborhood transition process can improve the material and environmental circumstances of some residents and bring forth harmful consequences such as heightened financial stress and residential displacement for other community members. The subsequent impact of gentrification on population health is understudied, and little is known about how gentrification influences the mental wellness of residents.

This dissertation advances the small but growing literature on the relationship between gentrification and adult mental health. Using multiple data sources, we identified Southern California neighborhoods that gentrified between 2010 and 2015 and investigated the impact of living in a gentrified neighborhood on mental health distress. Econometric techniques such as instrumental variables estimation and propensity score analyses were applied to reduce bias arising from residential selection and reverse causality.

The first study compared three quantitative approaches for identifying gentrified neighborhoods and demonstrated that each approach generated a different set of results. Findings highlighted the importance of the strategy used for identifying gentrified neighborhoods, especially when assessing gentrification’s effects on health outcomes. The second study used five years of pooled data from the California Health Interview Survey to examine the causal relationship between gentrification and adult mental health. Relative to living in a low-income and not gentrified neighborhood, living in a gentrified neighborhood was associated with increased likelihood of serious psychological distress among longtime residents, renters, and people with low incomes. In the third study, we evaluated reasons for moving between residents who moved within gentrified and not gentrified neighborhoods and found evidence that people in gentrified neighborhoods were more likely to experience within-neighborhood displacement. Residents who experienced within-neighborhood displacement had greater likelihoods of having serious psychological distress.

Taken together, findings suggest that gentrification imposes a mental health cost on longtime residents and the most financially vulnerable residents, which has important implications for population health. By elevating levels of mental health distress of population groups who are already disproportionately exposed to stressors, gentrification can exacerbate mental health inequities.

Legal status is an important social determinant of health for immigrants and children of immigrant parents, which is typically not measured in public health surveys. The sensitivity of legal status and presumed response behavior to relevant questions are primary reasons why this topic goes unmeasured. Changes in immigration enforcement likely impact the sensitivity of the topic and may compromise data quality, however, this is also likely when legal status matters most for health outcomes. This dissertation evaluates the response behavior to questions of citizenship and immigration status in the California Health Interview Survey and applies these data to identify mental health risks for Latino adolescents with an unauthorized parent.

The first study, When we ask, do they answer? Item-nonresponse to questions of citizenship and immigration status in the California Health Interview Survey, examined foreign born survey participants who did not answer questions of citizenship and immigration status

between 2001 and 2015. Nonresponse was low overall, however, increased over time and was largely attributable to respondents who were born in Mexico. The second study, When they answer, should we listen? Examining the quality of self-reported citizenship and immigration status, evaluated potential misreporting of legal status among Mexican-born participants between 2003 and 2015. This study utilized indirect estimation strategies which have been developed to produce profiles of the unauthorized population from surveys which do not ask legal status. Nearly a quarter of all Mexican-born participants reported that they were a non-citizen without a

green card, and these participants were demographically similar to external profiles of the unauthorized population. Predicted probabilities of unauthorized status produced by the indirect estimation procedure indicated that the threat of extensive misreporting was low and consistent over time. These results, paired with the findings of low nonresponse, indicate that participants were willing to answer questions of citizenship and immigration status and that these data are fit for use. The third paper, Severe Psychological Distress Among Latino Adolescents with an Unauthorized Parent examined adolescent mental health using data from 2007 to 2016 disaggregated by parental nativity and legal status. Multivariate logistic models indicated that Latino adolescents with an immigrant mother were less likely to report severe psychological distress and that children with an unauthorized father were more likely to report severe psychological distress. These findings reveal important heterogeneity among children in immigrant households and demonstrates the value of measuring legal status in a population survey. It is critical that data used to monitor public health trends more fully incorporate immigrants and their children by measuring domains which are relevant to their health and wellbeing. In addition to measuring what needs to be measured, researchers should continue to critically evaluate quality and put data which are fit to use to meaningful and timely use.