More than 270,000 km of rivers and streams are impaired due to fecal pathogens, creating an economic and public health burden. Fecal indicator organisms such as Escherichia coli are used to determine if surface waters are pathogen impaired, but they fail to identify human health risks, provide source information, or have unique fate and transport processes. Statistical and machine learning models can be used to overcome some of these weaknesses, including identifying ecological mechanisms influencing fecal pollution. In this study, canonical correlation analysis (CCorA) was performed to select parameters for the machine learning model, Maxent, to identify how chemical and microbial parameters can predict E. coli impairment and F+-somatic bacteriophage detections. Models were validated using a bootstrapping cross-validation. Three suites of models were developed; initial models using all parameters, models using parameters identified in CCorA, and optimized models after further sensitivity analysis. Canonical correlation analysis reduced the number of parameters needed to achieve the same degree of accuracy in the initial E. coli model (84.7%), and sensitivity analysis improved accuracy to 86.1%. Bacteriophage model accuracies were 79.2, 70.8, and 69.4% for the initial, CCorA, and optimized models, respectively; this suggests complex ecological interactions of bacteriophages are not captured by CCorA. Results indicate distinct ecological drivers of impairment depending on the fecal indicator organism used. Escherichia coli impairment is driven by increased hardness and microbial activity, whereas bacteriophage detection is inhibited by high levels of coliforms in sediment. Both indicators were influenced by organic pollution and phosphorus limitation.

Background.The leading cause of surface water impairment in United States’ rivers and streams is pathogen contamination. Although use of fecal indicators has reduced human health risk, current approaches to identify and reduce exposure can be improved. One important knowledge gap within exposure assessment is characterization of complex fate and transport processes of fecal pollution. Novel modeling processes can inform watershed decision-making to improve exposure assessment.

Utilizing data collected by the University of Kentucky Lung Cancer Research Initiative (LCRI), this study investigated potential mining-related sources for the elevated levels of cadmium in Harlan and Letcher counties. Statistical analyses for this study were conducted utilizing SAS. A number of linear regression models and logarithmic models were used to evaluate the significance of the data. The linear regression models consisted of both simple and multivariate types, with the simple models seeking to establish significance between the potential sources and urine cadmium levels and the multivariate models seeking both to identify any statistically significant linear relationships between source types as well as establish a relationship between the potential source and the urine cadmium levels. The analysis began by investigating which ingestion method caused the increased levels of cadmium exposure. The analysis included ingestion through water sources and inhalation of dust. Of these two, dust showed the higher level of correlation. The second step was to analyze a number of sources of dust, particularly those related to mining practices in the area. These included the proximity to the Extended Haul Road System, secondary haul roads, rail roads, and processing plants. Of the variables in the analysis, Extended Haul Roads, secondary haul roads, and rail roads showed no correlation, and only the proximity to processing plants showed statistical significance.

The relationship between the number of persons per room in the home and the relative occurrence of stress-related morbidities of hospitalized patients is investigated. The social factors of age, social class, and level of social support also are examined as they relate to both density and morbidity type occurrence. The population under study is composed of all Kaiser-Permanente Medical Care Program members twenty years of age and older, discharged from Bess Kaiser Hospital during 1974, and whose diagnosed morbidity was included in specific morbidity categories. All data were obtained through the facilities of the Health Services Research Center, Kaiser Foundation Hospital, Portland, Oregon. The data are largely ordinal in nature, and are presented in the form of contingency tables, utilizing chi square and Tau B as measures of degree of association and strength of relationship. An interval level of measurement is obtained through dichotomization of the variables and the derivation of Pearsonian correlation coefficients, and regression coefficients which are utilized in path analysis. The findings indicate that higher residential population density is related to disproportionate occurrence of some, but not all, stress-related morbidities, and that some stress-related morbidities are associated with low population density. However, in comparing stress-related morbidities with non-stress related morbidities, it is found that more stress-related morbidity categories tend to be over-represented in the higher density regimes than those morbidities categorized as non-stress related. In addition, it is found that those stress-related morbidities that exhibit such a pattern of over-representation maintain it, in general, regardless of the controls applied. Age is found to be negatively related to density and to the occurrence of stress-related morbidities, while being positively related to socioeconomic status. Socioeconomic status is found to be negatively related to both stress-related morbidity occurrence and to density. Social support is positively related to both density and to the occurrence of stress-related morbidities, but appears to have no relationship to either age or class. Hospital during 1974, and whose diagnosed morbidity was included in specific morbidity categories" All data were obtained through the facilities of the Health Services Research Center, Kaiser Foundation Hospital, Portland, Oregon.

Environmental epidemiologists are increasingly interested in establishing causality between exposures and health outcomes. A popular model for causal inference is the Rubin Causal Model (RCM), which typically seeks to estimate the average difference in study units' potential outcomes. If the exposure Z is binary, then we may express this as E[Y(Z=1)-Y(Z=0)]. An important assumption under RCM is no interference; that is, the potential outcomes of one unit are not affected by the exposure status of other units. The no interference assumption is violated if we expect spillover or diffusion of exposure effects based on units' proximity to other units and several other causal estimands arise. For example, if we consider the effect of other study units on a unit in an adjacency matrix A, then we may estimate a direct effect, E[Y(Z=1,A)-Y(Z=0,A)], and a spillover effect, E[Y(Z,A)=Y(Z,A`)]. This thesis presents novel methods for estimating causal effects under interference. We begin by outlining the potential outcomes framework and introducing the assumptions necessary for causal inference with no interference. We present an association study that assesses the relationship of animal feeding operations (AFOs) on groundwater nitrate in private wells in Iowa, USA. We then place the relationship in a causal framework where we estimate the causal effects of AFO placement on groundwater nitrate using propensity score-based methods. We proceed to causal inference with interference, which we motivate with examples from air pollution epidemiology where upwind events may affect downwind locations. We adapt assumptions for causal inference in social networks to causal inference with spatially structured interference. We then use propensity score-based methods to estimate both direct and spillover causal effects. We apply these methods to estimate the causal effects of the Environmental Protection Agency’s nonattainment regulation for particulate matter on lung cancer incidence in California, Georgia, and Kentucky using data from the Surveillance, Epidemiology, and End Results Program. As an alternative causal method, we motivate use of wind speed as an instrumental variable to define principal strata based on which study units are experiencing interference. We apply these methods to estimate the causal effects of air pollution on asthma incidence in the San Diego, California, USA region using data from the 500 Cities Project. All our methods are proposed in a Bayesian setting. We conclude by discussing the contributions of this thesis and the future of causal analysis in environmental epidemiology.

The overall objective of the research presented in this dissertation was to investigate molecular risk factors of susceptibility to estrogenic chemicals, polychlorinated biphenyls (PCBs), hormone replacement therapy, and oral contraceptives and how that leads to the development of pulmonary arterial hypertension (PAH). Environmental and molecular risk factors for PAH are not clearly understood. This is a major hurdle for the development of new therapy against PAH as well as understanding individual susceptibility to this disease. Gender has been shown to impact the prevalence of PAH. Although controversial, estrogens have been implicated to be a risk factor for PAH. Thus, we hypothesize that women exposed to estrogenic chemicals are at increased risk of developing PAH when endocrine disrupting chemicals interact with unopposed estrogen to worsen pulmonary arterial disease. In support of this hypothesis, we have accomplished the following: Microarray data on PAH were collected and subsequent meta-analysis was conducted using genome-wide association and environment-wide association approaches on published studies as well as GEO and NHANES data. All PCB geometric mean concentrations found higher levels in people at risk of PAH than people not at risk of PAH. The sum of non-dioxin-like PCBs and the sum of dioxin-like PCBs were significantly higher in people at risk of PAH than people not at risk of PAH. Also, different levels of LOD (including PCBs concentration >LOD, > 50th percentile, 50th-75th percentile, and ≥75th percentile) were significantly higher in people at risk of PAH than people not at risk of PAH. We reported that females used estrogen pills and oral contraceptive were associated with risk of PAH. However, females used progestin and estrogen/progestin pills were not at risk of PAH. Molecular risk factor analysis using machine learning approaches revealed that VAMP2, LAMA5, POLR2C, VEGFB, and PRKCH genes are causal genes of PAH pathogenesis. Gene ontology and pathway analysis of PAH showed that genes involved in the apoptosis pathway, p53 pathway, Ras Pathway, T-cell activation, TGF-beta pathway, VEGF pathway, and Wnt pathway appear to be significantly associated with PAH. Documenting the exposure to estrogenic chemicals among the general U.S. population, and identifying agents and molecular risk factors associated with PAH have the potential to fill research gaps and facilitate our understanding of the complex role environmental chemicals play in producing toxicity in the lungs.

This work characterized burn-off emissions from automobiles. After an exhaustive literature review, engine temperatures were determined to reach a maximum temperature of approximately 110oC, while exhaust system components reached a maximum temperature around 600oC. Metal-drawing fluids were used to bend the exhaust system components during manufacturing. Because these components were not rinsed prior to incorporation into a vehicle, residues could be left on the surfaces. An experimental test chamber was constructed to conduct controlled testing of three metalworking fluids of various types to mimic real-world conditions. Real-time particle number measurements were made using a condensation particle counter and an optical particle counter. The temperature at which burn-off begins to occur was found to be around 120 to 150oC. This burn-off was found to be an evaporation-condensation phenomenon when metalworking fluid residues vaporize and condense forming fine (0.1µm to 2.5µm) and ultrafine (<0.1µm) aerosols. The temperature dependency of this phenomenon was observed to follow the Clausius-Clapeyron equation that states as temperature increases, vapor pressure increases. Most aerosol particles were observed to be in the range of less than 0.01µm to approximately 2.0µm.

The purpose of this research is to demonstrate the ability to generate and characterize a nanometer sized aerosol using solutions, suspensions, and a bulk nanopowder, and to research the viability of using an acoustic dry aerosol generator/elutriator (ADAGE) to aerosolize a bulk nanopowder into a nanometer sized aerosol. The research compares the results from a portable scanning mobility particle sizer (SMPS) to the more traditional method of counting and sizing particles on a filter sample using scanning electron microscopy (SEM). Sodium chloride aerosol was used for the comparisons. The sputter coating thickness, a conductive coating necessary for SEM, was measured on different sizes of polystyrene latex spheres (PSLS). Aluminum oxide powder was aerosolized using an ADAGE and several different support membranes and sound frequency combinations were explored. A portable SMPS was used to determine the size distributions of the generated aerosols. Polycarbonate membrane (PCM) filter samples were collected for subsequent SEM analysis. The particle size distributions were determined from photographs of the membrane filters. SMPS data and membrane samples were collected simultaneously. The sputter coating thicknesses on four different sizes of PSLS, range 57 nanometers (nm) to 220 nm, were measured using transmission electron microscopy and the results from the SEM and SMPS were compared after accounting for the sputter coating thickness. Aluminum oxide nanopowder (20 nm) was aerosolized using a modified ADAGE technique. Four different support membranes and four different sound frequencies were tested with the ADAGE. The aerosol was collected onto PCM filters and the samples were examined using SEM. The results indicate that the SMPS and SEM distributions were log-normally distributed with a median diameter of approximately 42 nm and 55 nm, respectively, and geometric standard deviations (GSD) of approximately 1.6 and 1.7, respectively. The two methods yielded similar distributional trends with a difference in median diameters of approximately 11 - 15 nm. The sputter coating thickness on the different sizes of PSLSs ranged from 15.4 - 17.4 nm. The aerosols generated, using the modified ADAGE, were low in concentration. The particles remained as agglomerates and varied widely in size. An aluminum foil support membrane coupled with a high sound frequency generated the smallest agglomerates. A well characterized sodium chloride aerosol was generated and was reproducible. The distributions determined using SEM were slightly larger than those obtained from SMPS, however, the distributions had relatively the same shape as reflected in their GSDs. This suggests that a portable SMPS is a suitable method for characterizing a nanoaerosol. The sizing techniques could be compared after correcting for the effects of the sputter coating necessary for SEM examination. It was determined that the sputter coating thickness on nano-sized particles and particles up to approximately 220 nm can be expected to be the same and that the sputter coating can add considerably to the size of a nanoparticle. This has important implications for worker health where nanoaerosol exposure is a concern. The sputter coating must be considered when SEM is used to describe a nanoaerosol exposure. The performance of the modified ADAGE was less than expected. The low aerosol output from the ADAGE prevented a more detailed analysis and was limited to only a qualitative comparison. Some combinations of support membranes and sound frequencies performed better than others, particularly conductive support membranes and high sound frequencies. In conclusion, a portable SMPS yielded results similar to those obtained by SEM. The sputter coating was the same thickness on the PSLSs studied. The sputter coating thickness must be considered when characterizing nanoparticles using SEM. Finally, a conductive support membrane and higher frequencies appeared to generate the smallest agglomerates using the ADAGE technique.

Certain N95 FFR models that staple the head straps directly onto filtering material are commercially available. This method of assembly can tear or reduce fiber density in the immediate area surrounding the staple punctures. Five N95 FFR models were evaluated to determine if staple punctures on the filter medium reduce the protection offered by the respirators. Total penetrations were measured with the staple punctures intact and when stretching the head straps a distance equivalent to a 95% male head circumference. Filter penetration were measured by sealing the staple punctures. Aerosols of 200, 500, and 1000 nm were used to challenge respirators at 28 and 85 Liter/min flow rates. Staple punctures were visually inspected by macrophotography with a light source on the opposing side of the punctures. Three FFR models had greater mean leakages than the remaining two. However, only two FFR models had statistically significant greater total penetrations than filter penetrations. Pulling the head straps increased total penetration, but was not statistically significant. Filter penetrations were greatest at 85 Liter/min and 200 nm, while leakages were greatest at 28 Liter/min flow rate and 1000 nm. Leakage through the staple punctures had greater contributions to total penetration than filter penetration allowing a greater percentage of 1000 nm particles into the breathing zone. Leakage was dependent on the tearing of the filter material or the reduction of fiber density near the puncture, regardless of filter efficiency. Total penetration to filter penetration ratios showed that leakage was greater than filter penetration 15 fold for 1,000 nm. This value is similar to what has been reported for face seal leaks on human subjects. Protection factors were reduced from ~930 to ~60 when the staple punctures created a tear. N95 FFR with stapled head straps that puncture the filter medium should be avoided because they can reduce protection to the user.

Study Purpose: The purpose of this study was to examine whether there were differences in patterns of care between African American (AA) and Non-Hispanic White (NHW) men newly diagnosed with prostate cancer in Florida, and how the treatment patterns compare with the National Comprehensive Cancer Network (NCCN) initial treatment recommendations. Materials and Methods: This retrospective cohort study utilized data from the Florida Cancer Data System (FCDS), to identify incident cases of prostate cancer diagnosed between 1982 and 2012. The variables of interest included: race/ethnicity, marital status, age at diagnosis, stage at diagnosis, tumor grade, year of diagnosis, and treatment modality (singular or multimodality). Adjusted odds ratios (AORs) and 95% confidence intervals were calculated to determine disparities in the receipt of treatment by age at diagnosis, stage at diagnosis and tumor grade between AA and NHW men. Results: A total of 244,449 AA (30,556 cases or 12.5%) and NHW (213,893 cases or 87.5%) men met the study inclusion/exclusion criteria. AA men were significantly less likely to receive surgery only or surgery in combination with other treatment modalities compared to NHW men, localized disease (AOR=0.66, 95% CI (0.63-0.68), regional disease (AOR=0.63, 95% CI (0.57- 0.71), distant disease (AOR=0.50, 95% CI (0.34-0.75). Comparisons of adherence to the NCCN initial treatment recommendations indicate that AA men with(5% versus 13%). Moreover, AA men in the very high risk group had a higher NCCN initial treatment adherence percentage compared to NHW men (76% versus 70%). Conclusion: After adjusting for potential demographic and clinical confounders, significant differences exist in the receipt of first course of treatment where AA men were more likely to receive radiation and/or hormone therapy and less likely to receive surgery compared to NHW men. Further research is needed to address this disparity.