Many observational studies assessing the effects of treatments or exposures are limited to comparisons between treatment users and nonusers or exposed and unexposed participants at study entry. However, the underlying and etiologically relevant exposure may gradually increase over time before reaching some plateau. This amount of time required for this latent cumulative exposure to reach a maximum hazard will be referred to as the "lag", coming from the concept that the association between exposure and outcome is lagged or delayed. Accounting for the lag is essential when analyzing exposure-response associations adequately. My challenge was to simultaneously estimate the lag-time and the exposure's lagged-association with the outcome at plateau. In this dissertation, I draw an analogy with the pharmacokinetic one-compartment model (OCM). OCM describes the accumulation of a medication in the body based on an exponential cumulative density function whose rate of increase is defined by a half-life parameter. Upon discontinuation, the OCM assumes that a medication will eliminate at the same half-life rate. The decline, for my purposes, can be interpreted as the time to return to a null effect of exposure, which occurs at roughly 4-5 half-lives. My methods model the association of a latent exposure and dichotomous outcome using a half-life of effect, similar to the OCM, in longitudinal analyses of single and repeated exposures. I derive profile likelihood-based algorithms to estimate of the upper limit of association simultaneously with the rate of latent exposure growth towards or away from plateau. Lastly, I extend this approach to allow different half-life parameters for incline and decline. Using simulations, I analyze the performance of my approach by comparing bias and coverage of the estimates for the half-life and effect parameters. With data from the Black Women's Health Study Cohort (a prospective cohort of 59,000 women followed 1995-2015), I show that prolonged cigarette smoking is associated with a maximum hazard of cardiovascular disease (CVD) at 2.5 times the hazard of never smokers. Additionally, I estimate that it takes about 7 years of smoking cessation for an individual's hazard of CVD to decrease by 50%. / 2020-11-13T00:00:00Z

Asthma is the most common chronic disease in children and the effects of air pollution exposure on asthma and respiratory health in children have been a growing concern over recent decades. Although a number of epidemiological studies have been carried out in this field, these have produced conflicting results. The aim of this study was to assess the effects of long term exposure to nitrogen dioxide (NO2) and particulate matter (PM10) on asthma prevalence and lung function in children. To achieve this, a novel exposure model was developed and evaluated, which allowed retrospective exposure assessment of children participating in a population based birth cohort study – the Manchester Asthma and Allergy Study (MAAS). MAAS is a prospective birth cohort study comprising 1185 children specifically designed to study asthma and allergies. Clinical follow up took place at ages 3, 5, 8 and 11 years. At each follow up parents completed questionnaires on asthma diagnosis and symptoms and children underwent skin prick tests for common allergens. Children’s specific airways resistance (sRaw, at ages 3, 5, 8, 11) and forced expiratory volume in one second (FEV1, at ages 5, 8, 11) were measured. At ages 5 and 11 years FEV1 was measured at baseline and after bronchodilator treatment. The exposure model developed during this study incorporated outdoor and indoor air pollution, spatio-temporal variation in air pollution and time-activity patterns of children. The model was based on the concept of microenvironmental exposure. It modelled personal exposure based on PM10 and NO2 concentrations in children’s home, school and journey microenvironments (MEs) and the length of time they spend in these MEs. Land use regression (LUR) models were used to model PM10 and NO2 concentrations in outdoor MEs. These LUR models were specifically developed for the Greater Manchester area. A novel method was used to develop the LUR models, which used the output from an air dispersion model as dependent variables in the regression analysis. Furthermore, a novel approach was used to obtain annual concentration of PM10 and NO2 from 1996 to 2010, which involved the recalibration of the LUR models for each year. A mass balance model and indoor to outdoor ratios were used to model concentrations in indoor MEs. The performance of the exposure model was evaluated through a personal monitoring study in schoolchildren attending a local secondary school. Children wore personal NO2 monitors for two consecutive days in four seasons. Parental questionnaires and time-activity diaries were used to obtain information for the exposure model and to model NO2 exposure for the same time period. The results showed good agreement between monitored and modelled NO2 concentrations (Normalised mean bias factor=-0.04). Multiple linear regression and generalised estimating equations (GEE) were used to assess the cross-sectional and longitudinal effect of modelled exposure on sRaw and FEV1 (as % predicted). Multiple logistic regression and GEE were used to assess the effect of modelled exposure on the prevalence of asthma and current wheeze.The longitudinal analyses showed significant associations between PM10 and NO2 exposure and % predicted FEV1 (PM10: B=-1.37, p=0.019; NO2: B=-0.83, p=0.003), but no association with sRaw (PM10: B=0.009, p=0.37; NO2: B=-0.007, p=0.16). The cross-sectional analyses showed no association between pollutant exposure during the summer or winter prior to age 11 and any of the lung function measures (p>0.05). Long term PM10 or NO2 exposure were not associated with asthma or current wheeze (p>0.05).This study developed and evaluated a novel air pollution exposure model for epidemiological research. The results of this study suggest a negative impact of long term exposure to NO2 and PM10 on growth in FEV1 during primary school age. However, no evidence of an association between long term exposure to NO2 and PM10 and childhood asthma was found.

Soil quality standards (SQSs) are one of the most important elements of management system for contaminated sites. In order to manage risks associated with soil contamination, risk based SQSs are used worldwide. However, in Turkey, the Soil Pollution Control Regulation in force was focusing mainly on the use of stabilized sludge on soil and was including standards for a limited number of parameters, mainly metals and some organic chemicals. Thus, existing SQSs were far away from providing common criteria for assessment of the soil quality. In this study, the aim was to develop human health risk based SQSs for Turkey. For derivation of risk based SQSs, the conceptual framework and technical infrastructure were established. SQSs were derived for 151 chemical substances and for three different land use types by incorporating generic site characteristics for Turkey. Since SQSs are highly sensitive to site conditions and chemical-specific data used in calculations, a Microsoft Excel based exposure model was developed as a technical tool. This tool serves for calculation of generic and site-specific SQSs and maintenance of the currency of the standards by allowing periodic update of data used in calculations. Besides, a hydrogeologic database was developed to provide information on the general soil and hydrogeologic characteristics that are used in derivation of SQSs. This database is ultimately, expected to serve for development of conceptual site models, sampling strategies, and derivation of dilution factors during risk assessment studies. As a result, this study presents a general perspective and approach for derivation of human health risk based SQSs. It is believed that the developed conceptual and technical infrastructure will contribute to contaminated site management and risk assessment studies conducted by the regulatory authorities and the other stakeholders in Turkey.

Human health risk assessments are used by environmental regulatory agencies to determine risk from Hazardous Air Pollutants (HAPs). In this study, the Human Exposure Model (HEM-3) was used to compare the cancer and non-cancer inhalation health effects of a single organic chemical manufacturing facility in Geismar, Louisiana prior to and after Maximum Achievable Control Technologies (MACT) were implemented. The results indicate significant reductions in both cancer risk and non-cancer hazards. The analysis also indicated that the equivalent cancer risk reduction could have been achieved by addressing MACT in only one production process and one single pollutant (ethylene dichloride) within that process. This demonstrates the value that these risk assessments have at evaluating emissions at the facility level, and how they could be used in the control strategy decision making process.

Researchers conduct life testing on objects of interest in an attempt to determine their life distribution as a means of studying their reliability (or survivability). Determining the life distribution of the objects under study helps manufacturers to identify potential faults, and to improve quality. Researchers sometimes conduct accelerated life tests (ALTs) to ensure that failure among the tested units is earlier than what could result under normal operating (or environmental) conditions. Moreover, such experiments allow the experimenters to examine the effects of high levels of one or more stress factors on the lifetimes of experimental units. Examples of stress factors include, but not limited to, cycling rate, dosage, humidity, load, pressure, temperature, vibration, voltage, etc. A special class of ALT is step-stress accelerated life testing. In this type of experiments, the study sample is tested at initial stresses for a given period of time. Afterwards, the levels of the stress factors are increased in agreement with prefixed points of time called stress-change times. In practice, time and resources are limited; thus, any experiment is expected to be constrained to a deadline which is called a termination time. Hence, the observed information may be subjected to Type-I censoring. This study discusses maximum likelihood inferential methods for the parameters of multiple step-stress models from a generalized Birnbaum-Saunders distribution under time constraint alongside other inference-related problems. A couple of general inference frameworks are studied; namely, the observed likelihood (OL) framework, and the expectation-maximization (EM) framework. The last-mentioned framework is considered since there is a possibility that Type-I censored data are obtained. In the first framework, the scoring algorithm is used to get the maximum likelihood estimators (MLEs) for the model parameters. In the second framework, EM-based algorithms are utilized to determine the required MLEs. Obtaining observed information matrices under both frameworks is also discussed. Accordingly, asymptotic and bootstrap-based interval estimators for the model parameters are derived. Model discrimination within the considered generalized Birnbaum-Saunders distribution is carried out by likelihood ratio test as well as by information-based criteria. The discussed step-stress models are illustrated by analyzing three real-life datasets. Accordingly, establishing optimal multiple step-stress test plans based on cost considerations and three optimality criteria is discussed. Since maximum likelihood estimators are obtained by numerical optimization that involves maximizing some objective functions, optimization methods used, and their software implementations in R are discussed. Because of the computational aspects are in focus in this study, the benefits of parallel computing in R, as a high-performance computational approach, are briefly addressed. Numerical examples and Monte Carlo simulations are used to illustrate and to evaluate the methods presented in this thesis. / Thesis / Doctor of Science (PhD)

Population exposure to traffic pollution is a rapidly developing, multi-disciplinary scientific field. While the link between long-term exposure and respiratory issues is well-established, there are probable links to a number of more serious health effects, which are still not fully understood. In the interests of protecting human health, it is prudent that we take a cautionary approach and actively seek to reduce exposure levels, especially in the home environment where people spend a significant portion of their time. In many large cities, a substantial number of homes are situated on land immediately adjacent to busy freeways and other heavily-trafficked roads. Characterising exposures of local residents is incredibly challenging but necessary for advancing epidemiological understandings. While existing studies are plentiful, the results are mixed and generally not transferable to other urban areas due to the localised nature of the built environment and meteorological influences. This thesis aimed to employ a variety of methods to develop a holistic understanding of the influence of traffic emissions on near-highway residents' exposure in two communities of South Auckland, New Zealand, where Annual Average Daily Traffic (AADT) is as high as 122,000 vehicles. First, ultrafine particles (UFPs), nitrogen oxides (NOx), carbon monoxide (CO) and particulate matter ≤ 10 μm (PM₁₀) were continuously monitored using a series of fixed stations at different distances from the highways, over several months during the winters of 2010 and 2011. Emissions modelling output (based on traffic composition), was used within a dispersion model to compare modelled concentrations with monitored levels. In addition, community census meshblock units were mapped by level of social deprivation in order to assess potential inequities in highway emissions exposure. The second layer of local air quality investigation involved using a bicycle platform to systematically measure concentrations of UFPs, CO and PM₁₀ using the entire street-grid network throughout each community. This was done forty times - five times at four times of day (07:00, 12:00, 17:00 and 22:00), for each study area, with the aim of mapping the diurnal fluctuation of microspatial variation in concentrations. Using global positioning system (GPS) data and geographical information system (GIS) software, spatially-resolved pollutant levels were pooled by time of day and the median values mapped, providing a visualisation of the spatial extent of the influence of emissions from the highways compared to minor roads. The third layer involved using data from multiple ambient monitors, both within the local areas and around the city, to simulate fifty-four residents' personal exposure for the month of June, 2010. This required collecting timeactivity information which was carried out by door-to-door surveying. The time-activity data were transformed into microenvironment and activity codes reflecting residents movements across a typical week, which were then run through the US-EPA's Air Pollution Exposure Model (APEX). APEX is a probabilistic population exposure model for which the user sets numerous microenvironmental parameters such as Air Exchange Rates (AERs) and infiltration factors, which are used in combination with air pollutant concentrations, meteorological, and geospatial data, to calculate individuals' exposures. Simulated exposure outputs were grouped by residents' occupations and their home addresses were artificially placed at varying distances from the highways. The effects of residential proximity to the highway, occupation, work destination and commute distance were explored using a Generalised Linear Model (GLM). Surveyed residents were also asked a series of Likert-type, ordered response questions relating to their perceptions and understandings of the potential impacts of living near a significant emissions source. Their response scores were explored as a function of proximity to the highway using multivariate linear regression. This formed the final layer of this investigation into air quality throughout these South Auckland communities of Otahuhu and Mangere Bridge. Results show that concentrations of primary traffic pollutants (UFPs, NOx, CO) are elevated by 41 - 64% within the roadside corridor compared to setback distances approximately 150 m away and that the spatial extent of UFPs can reach up to 650 m downwind early in the morning and late in the evening. Further, social deprivation mapping revealed that 100% of all census meshblocks within 150 m either side of both highways are at the extreme end of the deprivation index (NZDep levels 8 - 10). Simulations for residents dispersed across the community of Otahuhu estimated daily NOx and CO exposure would increase by 32 and 37% (p<0.001) if they lived immediately downwind of the highway. If they were to shift 100 m further downwind, daily exposure would decline by 56 - 70% (p<0.001). The difference in individuals' exposure levels by occupation varied across the same distance by a factor of eight (p<0.05), with unemployed or retired persons the most exposed due to having more free time to spend outdoors at home (recreation, gardening, etc.). Those working in ventilated offices were the least exposed, even though ambient concentrations - likely due to a strong urban street canyon effect - were higher than the nearest highway monitor (5 m downwind) by 25 - 30% for NOx and CO, respectively. Inverse linear relationships were identified for distance from highway and measures of concern for health impacts, as well as for noise (p<0.05). Positive linear relationships were identified for distance from highway and ratings of both outdoor and indoor air quality (p<0.05). Measures of level of income had no conclusive statistically significant effect on perceptions (p>0.05). The main findings within this thesis demonstrate that those living within the highway corridor are disproportionately exposed to elevated long-term average concentrations of toxic air pollutants which may impact on physical health. While the socioeconomic characteristics could also heighten susceptibility to potential health impacts in these areas, certain activity patterns can help mitigate exposure. This thesis has also shown that there may be quantifiable psychological benefits of a separation buffer of at least 100 m alongside major highways. These results enhance a very limited knowledge base on the impacts of near-roadway pollution in New Zealand. Furthermore, the results lend additional support to the international literature which is working to reduce residential exposures and population exposure disparities through better policies and improved environmental planning. Where possible, the placement of sensitive population groups within highway corridors, e.g. retirement homes, social housing complexes, schools and childcare centres, should be avoided.

In this thesis, we investigate several inferential issues regarding the lifetime data from exponential distribution under different censoring schemes. For reasons of time constraint and cost reduction, censored sampling is commonly employed in practice, especially in reliability engineering. Among various censoring schemes, progressive Type-I censoring provides not only the practical advantage of known termination time but also greater flexibility to the experimenter in the design stage by allowing for the removal of test units at non-terminal time points. Hence, we first consider the inference for a progressively Type-I censored life-testing experiment with k uniformly spaced intervals. For small to moderate sample sizes, a practical modification is proposed to the censoring scheme in order to guarantee a feasible life-test under progressive Type-I censoring. Under this setup, we obtain the maximum likelihood estimator (MLE) of the unknown mean parameter and derive the exact sampling distribution of the MLE through the use of conditional moment generating function under the condition that the existence of the MLE is ensured. Using the exact distribution of the MLE as well as its asymptotic distribution and the parametric bootstrap method, we discuss the construction of confidence intervals for the mean parameter and their performance is then assessed through Monte Carlo simulations. Next, we consider a special class of accelerated life tests, known as step-stress tests in reliability testing. In a step-stress test, the stress levels increase discretely at pre-fixed time points and this allows the experimenter to obtain information on the parameters of the lifetime distributions more quickly than under normal operating conditions. Here, we consider a k-step-stress accelerated life testing experiment with an equal step duration τ. In particular, the case of progressively Type-I censored data with a single stress variable is investigated. For small to moderate sample sizes, we introduce another practical modification to the model for a feasible k-step-stress test under progressive censoring, and the optimal τ is searched using the modified model. Next, we seek the optimal τ under the condition that the step-stress test proceeds to the k-th stress level, and the efficiency of this conditional inference is compared to the preceding models. In all cases, censoring is allowed at each change stress point iτ, i = 1, 2, ... , k, and the problem of selecting the optimal Tis discussed using C-optimality, D-optimality, and A-optimality criteria. Moreover, when a test unit fails, there are often more than one fatal cause for the failure, such as mechanical or electrical. Thus, we also consider the simple stepstress models under Type-I and Type-II censoring situations when the lifetime distributions corresponding to the different risk factors are independently exponentially distributed. Under this setup, we derive the MLEs of the unknown mean parameters of the different causes under the assumption of a cumulative exposure model. The exact distributions of the MLEs of the parameters are then derived through the use of conditional moment generating functions. Using these exact distributions as well as the asymptotic distributions and the parametric bootstrap method, we discuss the construction of confidence intervals for the parameters and then assess their performance through Monte Carlo simulations. / Thesis / Doctor of Philosophy (PhD)