In the United States, lung cancer is the second most common cancer in men and women and is the leading cause of cancer death. Lung cancer survival is lower than other leading cancers, and approximately half of people with lung cancer die within one year of being diagnosed. While small advances in treatment and interventions in tobacco cessation have improved survival in recent years, persistent poor survival suggests that new approaches are needed to identify contextual risk factors to improve lung cancer survival.

It has become widely accepted that where an individual lives or works is an important determinant of health. Place-based exposures such as neighborhood contextual factors may be particularly valuable for intervention in lung cancer survival as they can be intervened on using population-level interventions such as local or state policies. Recently, studies have observed associations between two potential neighborhood contextual factors and lung cancer survival: ambient air pollution and racial residential segregation. However, neighborhood-level exposures require specific methodological considerations because they have spatial and temporal components that need to be appropriately considered in analyses.

The first chapter of this dissertation reviews the epidemiologic evidence regarding spatial heterogeneity in lung cancer survival and the relationship between both air pollution and racial residential segregation on lung cancer survival. The second chapter evaluates how lung cancer survival varies spatially in California and how these spatial patterns may change over time. The third chapter examines the association between air pollution and lung cancer survival and illustrates how a systematic error known as immortal time bias can be introduced when a time-varying exposure such as air pollution is mishandled in the context of a time-to-event outcome. The fourth chapter of this dissertation describes two considerations in the measurement of segregation, the spatial scale of geographies and the spatial relationships of populations. This chapter illustrates how these considerations impact both the absolute measurement of segregation and subsequent conclusions about the impact on lung cancer survival. The final chapter of this dissertation summarizes key findings and highlights future directions to advance the study of neighborhood contextual factors in lung cancer survival.

The last several consecutive fire seasons on the West coast, specifically in California, have been the worst in recorded history. These massive fires have been driven largely in part by the anthropogenically caused climate crisis. As the climate crisis worsens, the fire season will as well. This has massive implications for human health as air pollution from each fire can cause disease and hospitalizations. Wildfires are a large source of particulate matter that is less than 2.5 microns in diameter (PM2.5). PM2.5 has been known to cause extremely severe respiratory and cardiovascular issues. Recently, studies have shown that PM2.5 generated by wildfires could have a different and more potent toxicity than PM2.5 generated from other pollutant sources. This project studied the impact of PM2.5 on hospitalizations for respiratory diseases in California between 2006-2013, through a health impact assessment (HIA). We quantified the burden of respiratory hospitalizations related to PM2.5 exposure among California communities through two different approaches: (i) naïve (considering the same toxicity for all PM2.5) and (ii) nuanced (considering the higher toxicity of PM2.5 due to wildfires). The results of the HIA displayed higher attributable numbers of respiratory hospitalizations when accounting for the larger health burden wildfire PM2.5 (i.e., nuanced approach). The delta between the naïve and nuanced approach was higher in northern California. By not considering the differential toxicity of wildfire PM2.5, we underestimate the attributable number of respiratory hospitalizations in California related to PM2.5 exposure. This study can be useful for future air pollution guideline recommendations.

Background: While the detrimental health effects of extreme heat and wildfire smoke are well established in high-income countries, there is substantially less evidence in low and middle-income settings. This dissertation research aimed to quantify the effects of extreme heat and wildfire smoke through a social vulnerability approach, using the San Diego-Tijuana border as a context to study differential impacts.Methods: This dissertation includes three studies examining the differential impacts of extreme heat and wildfire smoke in a binational context. The first study evaluates the spatial variation in the effect of extreme heat across municipalities in Mexico using a within-community-matched design with a Bayesian Hierarchical model extension and meta-regression to explore socioeconomic drivers. In the second study, we used the San Diego-Tijuana region as a unique context to study the differential effects of wildfire smoke on cardio-respiratory hospitalizations across the border using synthetic control methods (SCM). The third study also used SCM to explore the potential role of wildfire smoke in driving COVID-19 mortality in the San Diego-Tijuana border region. Results: In the first study, we found substantial spatial heterogeneity in the effects of heat across Mexico at the municipality level, and disadvantaged social conditions such as low education, poor housing conditions, and higher marginalization were important predictors of these differences. In the second study, wildfire smoke increased cardio-respiratory hospitalizations in the San Diego-Tijuana border region, with a higher, albeit imprecise, relative change in Tijuana likely driven by a higher poverty rate and increased social vulnerability. Lastly, in the third study, no strong effect of wildfire smoke on COVID-19 mortality was observed in either San Diego or Tijuana. Conclusion: The results of this dissertation indicate that social vulnerability is an important factor in understanding the health risks of extreme heat and wildfire smoke. We hope that highlighting the differential susceptibility of populations to these effects can help inform binational efforts to protect those that are most vulnerable to extreme weather events which will be increasingly prevalent and severe in the context of climate change.

As average life-expectancy increases worldwide, there is a heightened public health concern about impaired cognitive function and dementia with advancing age. There are an estimated 10 million new cases of dementia each year and the worldwide prevalence is expected to triple by 2050. Without current effective treatment, research has expanded to identify modifiable risk factors for dementia to promote healthy cognitive aging.

Air pollution is a unique modifiable risk factor as levels can be shaped by individual behaviors and population-level environmental policies and regulations. Ambient air pollution is a mixture of particulate matter, gases, and other organic and metallic components. Emerging evidence suggests chronic exposure may affect diseases of the central nervous system, including dementia and cognitive impairment. It is biologically plausible that increased exposure to air pollution can produce a neuroinflammatory response, which in turn can result in structural changes in the brain. Although epidemiologic studies have observed a link among some populations, the causal pathway between air pollution and dementia remains unclear and these studies are subject to specific methodological challenges.

The first chapter of this dissertation reviews the epidemiological evidence regarding the relationship between dementia, cognitive impairment, and air pollution and outlines some methodological challenges my research addresses. The second chapter demonstrates different methods to account for competing events in a cohort of older adults in France and provides recommendations for future research in studies of air pollution and dementia. The third chapter examines the causal pathway between air pollution and dementia, evaluating cardiovascular disease as a potential intermediate in a population-based cohort in Ontario, Canada. The fourth chapter expands the generalizability of the current research by examining the relationship between air pollution and cognitive impairment in a US cohort of Hispanic/Latino adults, an understudied ethnic group with well-documented disparities in both air pollution exposure and chronic health outcomes. The final chapter of this dissertation summarizes key findings and highlights future directions to advance epidemiologic research of air pollution and cognitive aging.

Recent estimates show significant human population expansion in the Amazon region that in turn implies dramatic changes in natural landscapes and also in human behaviors such as human mobility which may further expose these communities to malaria risk. The current path toward eliminating malaria is severely threatened by these environmental and human activity changes, yet there is scarce evidence of how changing land cover affects human mobility dynamics and ultimately malaria transmission, especially in the Amazon region.

This dissertation includes three studies examining the relationship of human mobility and land cover change on the epidemiology of malaria in the Peruvian Amazon. The first study assesses the connectivity structure and centrality between cities and villages as malaria transmission drivers in rural Amazonia using novel network analysis on granular passive case detection data. The second study determines the effect of out-of-village working activities on recent malaria exposure using two population-based studies and a g-computation approach to simulate multiple scenarios of mobility restrictions (by proportion of travelers, gender, and age) to quantify the impact of such restriction policies on malaria exposure reduction. The third study quantifies the effect of human population mobility on malaria risk using GPS data and fine-scale mobility metrics computed by a novel movement ecology non-parametric Bayesian framework.

The first study showed that localities with high connectivity consistently have higher malaria endemicity that was exacerbated in regions with the highest baseline malaria transmission rates. The second study presented the crucial significance of human mobility in supporting malaria transmission in the Peruvian Amazon. This study demonstrated the importance of targeting key subpopulations when creating occupational interventions by simulating the incidence of out-of-village employment activities to represent different policy scenarios. Targeting males and adults (18 years and older) groups has the greatest influence on malaria seropositivity. Finally, the third study showed that the high interaction between Amazon villages for reasons such as labor, commerce, or recreation may sustain such endemicity levels by increasing exposure to the malaria parasites and eventually increasing the importation risk.

The findings of this research can be used to inform control strategies and policies in the Amazon region to shift towards approaches that incorporate village connectivity structure and human mobility patterns to prioritize connected areas instead of single villages and intensify malaria screening in sub-populations defined by their mobility profile.

As the climate continues to change, there is a heightened public health concern about how extreme weather events can affect adverse health outcomes, especially in more vulnerable populations. Children and pregnant women living in low-to-middle-income countries (LMICs) are considered highly vulnerable populations. Acute respiratory infections have become the number one cause of under-five mortality across the globe. Additionally, preterm birth, complications from preterm birth, and stillbirths are the number one cause of neonatal deaths worldwide. These adverse health outcomes are preventable. Thus, research needs to focus on preventable environmental determinants that can help reduce the number of childhood respiratory infections and adverse birth outcomes in LMICs. Air pollution and extreme heat are preventable risk factors because exposure levels can be determined by individual behaviors and population-level policies and action plans. Wildland fire smoke and dust storms can release substantial amounts of particulate matter into the atmosphere. Prior evidence has tied these extreme weather events with adverse respiratory and cardiovascular outcomes and premature mortality in children. Additionally, exposure to extreme heat has been linked to higher risks of preterm birth and stillbirth in high-income countries. The proposed biological mechanisms of how particulate matter and extreme heat can cause these outcomes involve oxidative stress and inflammation. Previous evidence has mainly come from high-income countries, which may be because of challenges in exposure measurement. The first chapter of this dissertation reviews the epidemiological evidence regarding the relationship of three extreme weather events (wildland fire smoke, dust, and extreme heat) and childhood respiratory outcomes and adverse birth outcomes in LMICs. The second chapter of this dissertation examines acute exposure to wildland and agricultural fire smoke affecting the risk of childhood respiratory outcomes in 14 LMICs. The third chapter is a case-study of acute exposure to heightened dust days and the risk of childhood cough in Benin. The fourth chapter studies the relationship between extreme heat and aims to identify specific susceptible windows of increased risk of preterm birth and stillbirth in a comprehensive set of LMICs. The latter three chapters expand on the limited evidence base of these exposure and outcome relationships in LMICs. The final chapter summarizes key findings, describes the implications and innovations of this research, and highlights future research objectives to build upon this dissertation research.

Fine particulate matter (PM2.5) is a mixture of microscopic solid particles and liquid droplets that pose significant health risks due to their ability to penetrate deep into the respiratory system. In recent decades, several studies have examined the relationships between air pollution and various meteorological factors, such as surface temperature, relative humidity, wind, and precipitation. Among these factors, precipitation can effectively remove PM2.5 and other airborne particles from the atmosphere through processes such as scavenging and wet deposition. Most studies focusing on the relationship between precipitation and PM2.5 have been conducted in regions affected by monsoon rains. This current work investigated the impact of precipitation patterns on PM2.5 concentrations in Los Angeles (LA) County, the most populous county in the United States, over twenty years (2003 to 2022). This study analyzed daily precipitation, PM2.5 concentrations, and atmospheric rivers (AR) to look at the impact on PM2.5 pollution by extreme precipitation events in a region characterized by a Mediterranean climate. The results of a moving window correlation analysis at different locations within LA County showed that PM2.5 concentrations and precipitation are inversely related, for the most part, particularly during wet winters. A matching analysis between days with precipitation above a given threshold and control days showed that PM2.5 tended to decrease the day of and the day after a precipitation event. The findings of this study can help understand the spatial heterogeneity of the relationship between precipitation and PM2.5 pollution in Southern California and the potential implications for public health in a changing climate.