The African-American family has experienced changes in composition since the involuntary migration from Africa to the Americas. This pivotal change in family composition has affected every aspect of life for African Americans.

Using body mass index (BMI) as a proxy for eating patterns and physical activity, this study examines the changes in family composition and its impact on BMI for African-American women. To examine the relationship between changing family composition and BMI over a 10-year period (2001-2011), this study uses publicly available data from the Panel Study of Income Dynamics. A series of bivariate correlational analyses and ordinary least-squares regressions via a mixed effects panel regression model were performed to examine women living in Traditional Households (married or cohabiting greater than or equal to 1 year) and Non-Traditional Households (female headed households no husband or cohabitor) and changes in BMI over time.

Obesity increased by 45.9% for women in Traditional Households versus 42.9% for women in Non-Traditional Households from 2001 - 2011.

The development in the last decade of methodology for measuring and scaling household food insecurity and hunger in U.S. populations makes possible systematic examination of the ways in which hunger and food insecurity affect individuals and families. The impact on children has always been of primary concern for policy, advocacy, and science because of the vulnerability of children to long-term developmental sequelae. There is an emerging and rapidly growing literature demonstrating deletrious links between inadequate food and a variety of developmental outcomes for children, including poorer health status, school absenteeism, and emotional and behavioral dysfunction. The research presented here explores the relationship of household food insecurity to children’s well-being in terms of cognitive and social development at kindergarten entry, utilizing a large and representative sample children in the United States. The timing of this evaluation, in the fall of the child’s first school experience, allows a snapshot of a child’s development throughout his/her preschool years relatively independent of the major influence that the school experience will have subsequently.

The data are from the Early Childhood Longitudinal Study of Kindergartners (ECLS-K), collected in 1998–99 by the National Center for Education Statistics, and comprise 20,929 children attending 1,000 private and public schools. Our results indicate that measures of reading, math, and general knowledge competence were not impacted by household food insecurity independent of other influences, but child emotional and functioning were negatively associated with household food insecurity even when many other relevant variables were controlled for. We also investigated the relationship of household food insecurity to children’s attained growth and found no independent relationship of household food insecurity to height for age or weight for height.

Obesity is a worldwide epidemic. While numerous obesity prevention efforts have been implemented around the world for different age groups and at different settings (i.e. school, home, community), most of these have shown short-term, moderate results if at all. Identifying successful obesity prevention strategies, therefore, remains critical. Food assistance programs (FAPs), originally designed to prevent hunger among the poor, have the potential to work as venues for obesity prevention interventions since they reach low-income populations with high rates of chronic disease. The goal of this dissertation was to examine the association between FAP participation, both in the U.S. and Peru.

For the first paper in this dissertation (Chapter 2), I assessed the individual and neighborhood-level predictors of participation in the Supplemental Nutrition Assistance Program (SNAP) among adults in LA County. Using data from wave 1 of the Los Angeles Family and Neighborhood Survey (L.A. FANS) and multilevel analyses in which people were nested in census tracts, I found that the best single predictor of SNAP participation was participation in the Temporary Assistance for Needy Families (TANF) program. Age was negatively associated with SNAP participation and Hispanics and people who cohabitated with their partners but were not married were more likely to participate in SNAP when compared to non-Hispanic Whites and people who were married, respectively. In addition, people who lived in neighborhoods with higher rates of social cohesion were less likely to participate in SNAP.

For the second paper (Chapter 3), I investigated if SNAP participation was associated with obesity among adults in LA County and if mental health was a mediator and the neighborhood food environment a moderator in the association between SNAP participation and BMI.

Again using L.A. FANS data, I found that SNAP participation was associated with BMI: an additional year of SNAP participation was associated with an increase in BMI of 0.14-0.19 Kg/m2. I also found that mental health partially and significantly mediated the association between SNAP participation and BMI. On the other hand, the neighborhood food environment, operationalized as the density (count per square mile) of unhealthy food outlets, was not a significant moderator in the association between SNAP participation and BMI.

For the third and final paper (Chapter 4) I analyzed data from Peruvian nationally representative cross-sectional surveys (2003-2010) to investigate if FAP participation was associated with overweight (including obesity) among women and preschool-aged children. For the country as a whole and all the years of study, FAP participation was not significantly associated with overweight among women and was negatively and significantly associated with overweight among children. When stratified by area of residence, FAP participation was associated with increased overweight for women who lived in Lima, the capital of Peru. For children, FAP participation was associated with lower overweight for all sub-groups except for children living in Lima.

Results from these three studies indicate that FAP participation is a potential risk factor for adult obesity, both in the U.S. and in Peru. Efforts should be made to better understand the mechanisms linking FAP participation and obesity. In the U.S., having a poor mental health seems to be one link associating SNAP participation and obesity. In Peru, on the other hand, the high caloric content as well as the high leakage rates seen in most FAPs could be reasons why FAP participants have higher overweight rates than their counterparts. Regardless of the reasons, FAP reform is warranted and the high rates of obesity experienced by FAP recipients should be taken into account when re-defining FAP benefits.

Climate change affects households, villages, countries, and regions worldwide. The populations and individuals most vulnerable to the negative impacts of climate change live in the poorest and least developed regions of the world, engage in climate-dependent livelihoods and have few assets. Ethiopia, one of the poorest countries in the world, with a history of climate- and politically-induced famines, has 80 percent of its population engaged in rain-fed agricultural livelihoods.

This dissertation sought to determine empirically how individuals and households in Ethiopia cope with climate change impacts on their livelihoods and food security. Ethiopia, one of the poorest countries in the world, where 80 percent of the population engages in rain-fed agricultural livelihoods, has a history of climate- and politically-induced famines. This dissertation consists of three studies. The first two studies were qualitative primary interviews where recruitment, consent, structured interviews, and protocols were approved by the UCLA Institution Review Board, but used mixed methods including t-tests and Fisher's exact tests in the analysis. The third study was quantitative and used panel data from seven waves of the Ethiopian Rural Household Survey (1994-2009).

The first study involved 59 rural-to-urban migrants in Addis Ababa, Ethiopia. Findings indicated that migration for some individuals was induced due to having insufficient tangible and intangible assets as a result of decreased or variable crop-yields, forced sale of livestock, sale of other assets, or need for additional income. Small plots of land, low levels of education, fertilizer loans, and tax debts were additional contributors to low asset levels. Migration for others was induced by volition or the desire to increase social or economic status with better jobs or education. Individuals who migrated by volition often reported having sufficient land, crop-yields, and food. Few individuals had access to safety-net programs and those who did reported insufficient or undesired provisions.

The second study involved 35 heads or proxy-heads of household in three rural villages. Findings indicated that households cope with asset shortfalls and climate-change with a handful of strategies including selling animals, eating other less-preferred foods, decreasing the amount eaten, and selling other assets. Nearly all interviewees worried about money, land, and food, and needed more money, land, and animals. Climate change was reported to affect crop-yields. Ethiopia is a three-meal per day culture. In good-food months, adults and children eat three meals per day. In bad-food months, both adults and children eat fewer meals, and adults eat fewer meals than children to preserve child-nutritional status. Adults from households with the least amount of land and the most household members eat the fewest number of meals in bad months. Only in the worst times did individuals report anyone migrating from the household though most reported a family member migrating in the past. Some believe they will migrate in the future, primarily for land or education.

The third study used longitudinal panel data from the Ethiopian Rural Household Survey, covering 1,477 households over 20 years. Analytical methods involved hierarchical linear modeling of bivariate and multivariate relationships between outcome and predictor variables and the elaboration model to test the modification of irrigation on rainfall-adequacy and food-security status. Hierarchical linear modeling tests how much of the observed relationship can be attributed to household, village, region, and time-specific aspects.

Findings indicated that households continually vary in their food-security status over time. Where a household is located, its geography, and time were the most important predictors of household food-security status, and both are indicators of climate effects. Irrigation also demonstrated improvements in food-security status beyond rainfall-adequacy alone, indicating that while climate was an important predictor of food security, irrigation modifies the relationship, and modifies the impact of geography or household-location. Household crop storage improved food-security status, while free-food distribution and household assets had no impact. Household assets however, do influence household resilience, the type, and choice of coping-strategies households' use.

The findings of these three studies provide new and supporting knowledge on coping-strategy use and food-security status over time in Ethiopia, demonstrating the influence of climate change as well as possible causes of migration in Ethiopia. These studies also confirm current literature describing the inadequate coverage and assistance of Ethiopia's safety-net program, guiding the way for improved pro-poor and climate-sensitive policy recommendations.

This policy brief reports that, according to the 2005 California Health Interview Survey, more than 2.5 million low-income Californians (30%) cannot afford adequate food for their families on a consistent basis. For more than 775,000 adults, this problem is so severe that they frequently reduce the quality of their food intake or go hungry. This policy brief uses data from the 2001, 2003, and 2005 California Health Interview Surveys to detail the changes in food insecurity by county, income levels, and race/ethnicity. It also describes some of the recent changes in terminology used to describe food insecurity issues. Lastly, the brief discusses some policy options that could be used to address the problems described.

PROBLEM: Informal caregivers of patients with chemotherapy-induced anemia (CIA) experience unaddressed burden. Previous research typically quantified burden with a score based on an instrument scale, but did not examine drivers or relative importance of clinical, demographic, and other characteristics on burden. As caregiver outcomes may be mediated by burden (a modifiable factor), research on predictors of burden are needed to identify points to target in designing interventions and policies.

METHODS: Data from a cross-sectional study conducted in eight oncology centers in the United States (3), Italy (3), and France (2) among 182 CIA patients and their caregivers were analyzed to assess burden and identify predictors of caregiver subjective burden, work productivity impairment (WI), activity impairment (AI), and time helping with activities of daily living (ADLs) and instrumental activities of daily living (IADLs). Descriptive, bivariate, interaction, and linear regression analyses were conducted.

RESULTS: Caregivers reported a substantial amount of time assisting with ADLs and IADLs, perceptions of burden, and considerable WI and AI. Additionally, patients reported fatigue and less than average mental and physical health. Subjective caregiver burden was significantly predicted by caregiver AI and caring for more individuals than the patient. Significant predictors of WI included the caregiver experiencing role overload in terms of feeling stressed meeting caregiving and family/work responsibilities and caregiver AI. The significant predictor of AI was the caregiver experiencing a loss of self in terms of feeling a decline in privacy. Caregiver time spent providing assistance was significantly predicted by patient physical quality of life and greater dependence on caregiver for assistance with ADLs.

CONCLUSIONS: Data supported that patients with CIA experienced disease burden and their caregivers experienced burden related to providing care in the US, Italy, and France. The data on sociodemographics, burden, and predictors of this burden added to the knowledge base, provides potential targets for future interventions, development of supportive policies, and aids health care providers, patients, and caregivers in making more informed decisions to address patient and caregiver needs. Notably, clinical and social efforts to alleviate the identified burdens would also improve the quality of life of CIA patients and their caregivers and reduce negative impacts to employers and the economy related to reduced caregiver worker productivity.

This dissertation advances understanding of how distal factors affect maternal health and health seeking behavior: by examining the links between place of residence and socioeconomic status (SES); quality of antenatal care (ANC); use of skilled birth attendants; and pregnancy outcomes. The dissertation is motivated by two main questions: Why are deliveries by skilled attendants low in most of sub-Saharan Africa, despite high antenatal attendance? And what accounts for the disparities in use of skilled birth attendants within countries? I hypothesize that differentials in quality of care partly explain the gap between antenatal attendance and use of skilled birth attendants, the rural/urban and SES differentials in the use of skilled birth attendants, as well as differentials in pregnancy outcomes.

Data for the analysis are from the Ghana Maternal Health Survey and the World Health Survey for Ghana and Burkina Faso. I employ multilevel linear and logistic regressions with mediation and moderation analyses to examine the intervening and conditional effects. The first three chapters present the introduction, background, and theory sections respectively; chapter 4 presents an overview of the data, variables, and sample distributions; and chapter 8 is the concluding chapter. The dissertation has three empirical chapters.

In the first empirical chapter (chapter 5), I examine the factors that affect the quality of antenatal care women receive--focusing on place of residence and SES (education and wealth); the interaction between these factors; and the mediating role of ANC timing, frequency, facility and provider. The results show that urban residence and higher SES are positively associated with higher ANC quality, but the urban effect is completely explained by sociodemographic factors. Specifically, about half of the urban effect is explained by education and wealth alone, with other variables accounting for the remainder. The effects of education are conditional on wealth and are strongest for the poorest women. Starting ANC visits early and attending the recommended four visits as well as receiving ANC from a higher-level facility and from a skilled provider, are associated with higher quality ANC. These factors partially explain the SES differentials in quality of ANC.

In the second empirical chapter (chapter 6), I examine the factors that influence the use of skilled birth attendants, focusing on the mediating role of quality of care. The results show that higher quality ANC, frequent ANC visits, and receiving ANC in higher level or private facilities increase the odds of using a skilled birth attendant. As expected urban residence and high SES are also associated with higher use of skilled birth attendants, but the rural/urban and SES differential in use of skilled attendants that is mediated by ANC quality is only marginally significant in most of the models. The rural/urban and SES effects are also not conditional on the ANC quality.

In the third empirical chapter (chapter 7), I examine the factors associated with pregnancy outcomes, focusing on the role of ANC quality. The results show that higher quality ANC decreases the odds of having a stillbirth by almost half, net of other factors including delivery provider and place. The other health service factor associated with lower odds of having a stillbirth in the multivariate analysis is attending at least four antenatal visits.

My findings suggest pregnant women in SSA experience significant disparities in the quality of ANC, with poor illiterate women receiving the worst care. Poor quality ANC is contributing to the low utilization of skilled birth attendants, despite high antenatal attendance, and to poor pregnancy outcomes. The differentials in quality of care are also potentially contributing to the differentials in use of skilled birth attendances by SES and place of residence, but this needs more research. Targeted efforts to increase quality of ANC could significantly improve maternal and fetal outcomes and reduce maternal health disparities in SSA.