Thèses sur le sujet « Correlated Exposure »

La ricerca epidemiologica mira a generare una migliore comprensione dei meccanismi sanitari e dell'eziologia della malattia. L'abbondanza di informazioni raccolte tramite cartelle cliniche elettroniche ha generato un problema di Big Data, rendendo l'identificazione di modelli predittivi dei rischi per la salute e dei possibili outcome climici ancora più complessa. Pertanto, sono aumentati in importanza e popolarità gli approcci statistici di selezione delle variabili. Tuttavia, i metodi di selezione delle variabili ancora oggi più utilizzati in epidemiologia si basano su conoscenze pregresse o utilizzano approcci automatizzati (es. regressione Stepwise). Quando però si combinano contemporaneamente più variabili d’esposizione con una complessa struttura di correlazione e con potenziali relazioni di tipo non lineare e non additive con l’outcome studiato, questi approcci risultano limitati nel tenere conto della multicollinearità. In questo contesto vengono utilizzati metodi di regressione penalizzati, come il least absolute shrinkage and selection operator (LASSO) e le sue numerose varianti, ma ricerche recenti hanno evidenziato che funzionano male quando sussiste un qualche tipo di struttura complessa di dipendenza tra le covariate. Negli ultimi anni, sono stati sviluppati metodi di selezione delle variabili appartenenti al framework Bayesiano, capaci di gestire congiuntamente più predittori, anche altamente correlati tra loro e con legami non lineari e non additivi con l’outcome. Tra questi metodi Bayesiani spiccano quelli che utilizzano priori spike-and-slab o priori shrinkage. Questi metodi sono ora ampiamente studiati e impiegati all'interno della letteratura epidemiologica ambientale e clinica. In questo lavoro abbiamo considerato e confrontato tre modelli di selezione delle variabili di tipo Bayesiano: il Bayesian kernel machine regression (BKMR), il Bayesian semiparametric regression (BSR) e il Bayesian LASSO (BLASSO). Il BKMR modellizza le associazioni tra predittori e outcome attraverso l'uso di una funzione kernel dei predittori, il BSR utilizzando spline naturali, mentre il BLASSO attraverso una funzione lineare dei predittori, considerando un parametro di shrinkage che mira ad effettuare selezione delle variabili ponendo a zero il coefficiente di regressione. In primo luogo, abbiamo valutato la bontà di adattamento e la capacità di selezione dei modelli attraverso uno studio di simulazione, generando diversi scenari con complessa struttura di correlazione e con relazioni disparate tra predittori e outcome. Successivamente, sulla base dei risultati ottenuti nelle simulazioni, abbiamo applicato questi metodi a due casi di studio reali. Il primo è lo studio ROCAV (Rischio di malattie cardiovascolari e aneurisma dell'aorta addominale a Varese), in collaborazione con l'Università dell'Insubria di Varese, con l'obiettivo di studiare l'impatto dell'esposizione a lungo termine a inquinanti ambientali con funzione respiratoria valutata mediante spirometria. I risultati hanno mostrato come non ci sia un inquinante che predomini sugli altri in termini di effetto, ma che tutti contribuiscono all’impatto sulla funzionalità respiratoria. Nel secondo studio (Follow Up OBese patients at AUXOlogico institute (FUOBAUXO)), in collaborazione con l'Istituto Auxologico Italiano, abbiamo studiato la relazione tra variabili antropometriche, cliniche, biochimiche, infiammatorie e metaboliche sulla perdita di peso percentuale in pazienti affetti da obesità in un periodo di ospedalizzazione di 40 giorni. Il risultato di questo lavoro ha portato all'individuazione di subsets di variabili potenzialmente associate con l’outcome di perdita di peso per diverse tipologie di paziente in base all’età, il genere e livello di BMI.<br>Epidemiological research aims to generate a better understanding of the health mechanisms and disease etiology. Nowadays, the abundance of clinical information collected via electronic medical records have made high-dimensional data extremely popular and the identification of predictive modeling of health risks and possible outcomes has become even more challenging. Therefore, many variable selection approaches have increased in importance and popularity. Nevertheless, the most frequently variable selection methods still used in epidemiology are based on prior knowledge or using stepwise automated selection approaches. However, when multiple exposures co-occur and have a strong complex correlation structure, traditional statistical approaches are limited in accounting for multi-collinearity or standard error inflation. To reduce this problem, dimensionality reduction methods –such as principal component and factor analyses– are very valuable. However, those approaches focus on the transformation of the original variables thus leading to an interpretability issue. In addition, multiple co-occurring predictors can have non-linear and non-additive relationships with the health outcome and most of the statistical methods fail to model properly those relationships. Penalized regression methods are used in this context, such as least absolute shrinkage and selection operator (LASSO) and its numerous variants, but recent research highlighted that they perform poorly when there exists some type of dependence structure between the covariates. In recent years, Bayesian variable selection methods have been developed to handle jointly multiple correlated predictors and both non-linear and non-additive relationships, allowing the inclusion of prior information. Among these Bayesian methods stand out those employing spike-and-slab priors or shrinkage priors for features selection. These methods are now widely studied and employed within the environmental and clinical epidemiological literature. In this thesis work, we considered and compared three Bayesian variable selection models: the Bayesian kernel machine regression (BKMR), the Bayesian semiparametric regression (BSR) and the Bayesian LASSO (BLASSO). BKMR models the outcome-predictors associations using a kernel function of predictors, BSR employing natural splines, while BLASSO using a linear function of predictors and considering a shrinkage parameter that aims to perform variable selection by setting the regression coefficient to zero. We firstly evaluated the models’ goodness of fit and selection ability simulating several predictors with a complex correlation structure and with disparate relationships with a continuous outcome, considering data with different sample sizes. Subsequently, based on the results obtained in the simulations we applied this methods to two real case studies. The first is the ROCAV study (Risk Of Cardiovascular diseases and abdominal aortic Aneurism in Varese), in collaboration with the University of Insubria of Varese, with the aim of studying the impact of long-term exposure to environmental pollutants with respiratory function assessed by spirometry. Results showed that no pollutant overcome the others in terms of effect, but rather all may contribute to the relationship with the respiratory outcome. The second, in collaboration with the Italian Auxological Institute, leveraged data on weight loss in hospitalized obesity affected women from the Follow Up OBese patients at AUXOlogico institute (FUOBAUXO) cohort, with the aim to select most important predictors in explaining the association between biochemical, anthropometric and clinical variables on weight loss percentage in these patients over a period of 40 days. The result of this work led to the identification of subsets of variables potentially associated with the weight loss outcome for different types of patients based on age, gender and BMI level.

First Nations peoples experience many benefits from eating locally-harvested wild foods, but these benefits must be considered along with the potential risks associated with exposure to environmental contaminants. Unlike store-bought foods, wild foods are an important traditional resource and a significant source of dietary protein, essential minerals and polyunsaturated fatty acids, believed to help in the prevention and treatment of obesity and obesity-related diseases such as type-2 diabetes mellitus. Wild foods continue to be an important and healthy food choice for First Nations peoples; however, they are also a primary source of dietary mercury, polychlorinated biphenyls (PCBs) and other persistent organic pollutants (POPs). To assess the effects of wild food consumption on dietary markers and contaminant accumulation, we grouped individuals from two remote Oji-Cree First Nations communities of north-western Ontario (n=71) according to their level of wild food consumption. In this study, I observed significantly higher organic contaminants in blood and higher mercury concentrations in hair for individuals consuming greater amounts of wild food. Age-adjusted contaminant concentrations were on average 3.5-times higher among high-frequency wild food consumers, with many exceeding federal and international health guidelines for mercury and PCB exposures. Contaminants in these populations approach, and in some cases exceed, threshold levels for adverse effects with potential consequences especially for prenatal development. Here, I also investigated the potential for stable isotope ratios of carbon and nitrogen (δ13C and δ15N) to serve as dietary markers and found strong positive correlations between stable isotopes and frequency of wild food and fish consumption. Frequency of fish consumption and δ15N was also shown to be positively correlated with mercury concentrations in hair and PCB concentrations in plasma. The results of this thesis demonstrate that known differences in dietary behaviour are clearly reflected in stable isotope ratios and contaminant concentrations. The data also show that contaminant exposures to those consuming wild foods in remote Boreal ecosystems is comparable to those associated with serious health effects in industrialized areas, and the problem of contaminants in wild foods is more widespread than the available literature would have led us to believe. These results affect our appreciation of contaminant exposures to First Nations peoples and will have implications for dietary choices, particularly if individuals are encouraged to consume greater amounts of wild foods for their proposed health benefits. We recommend further attention be given to the risks of contaminants in locally-harvested wild foods when promoting the benefits of their consumption to First Nations people as the problem of contaminants in remote communities practicing traditional lifestyles is often underreported and underplayed.

The etiology of musculoskeletal outcomes associated with the use of information technology (IT) has predominately been defined by studies of adults in their work environments. Theories explaining the causation of work related musculoskeletal disorders have identified individual user (biomechanical, physiological and psychosocial), task demand, work organization and environmental risk factors. Models based on these theories have subsequently been developed to investigate the causal relationship between IT exposure and outcomes experienced by the user.Computers are an important IT type used by children, and computer use by children is rapidly growing in both school and home environments. Recent literature demonstrates an increase in children’s reports of computer related musculoskeletal outcomes. Children’s computer use appears to be different to adult’s work related computer use. Thus, although many potential risk factors for children may be similar to those for adults, it is proposed that risk factors and models of causal relationships between computer use and musculoskeletal outcomes may vary for children.The main aims of this study were: (1) to investigate children’s computer exposure in their usual occupational environments of school and home; and (2) to develop and test a multivariable model that would assist in understanding relationships between child user correlates, computer exposure and computer related musculoskeletal outcomes.1351 students (792 boys and 559 girls) from eight primary and five secondary schools in Perth, Australia, participated in the study in 2006. Convenience sampling was undertaken within stratified groups, to ensure the sample had the required range of participants from different socioeconomic status (SES) backgrounds, both genders and school Years 1, 6, 9 and 11(approximate ages 6, 9, 14 and 16 years).The study design was cross sectional involving the completion of a questionnaire survey by participants, and for younger participants their parents. Questionnaires contained items relating to the participant and their activity exposure as an individual, within a family context, and within their neighbourhood. Physical measures of height and weight were also collected.The results showed that 100% of children had access to computers at school, and at home 98.9% of children had access to computers, with 95.9% reporting home internet access. The use of different exposure measures demonstrated that at school 97.8% of children had used a computer in the last month, for an average of 2.4 hours per week, commonly for 30-60 minutes in one sitting. At home 95.7% of children had used a computer in the last month, for an average of 7.2 hours per week, commonly for 60 - 120 minutes in one sitting. Computer activities performed more frequently at school were surfing the internet, learning programs and multimedia. At home the most frequent computer activities were surfing the internet and email. Children with bedroom computer access were found to have nearly 50% greater mean weekly hours of use. The use of a range of computer exposure measures (frequency, usual and longest duration, mean weekly hours and frequency of computer activities) provided better characterization of the amount and nature of children’s school and home computer exposure.Age and gender were associated with children’s school and home computer use. Computer use was greater with age for both boys and girls, and boys had greater use than girls across all Year levels for all exposure measures except school usual duration. Children with greater computer exposure were shown to experience less computer anxiety; reported more somatic complaints; had used a broader range of computer activities; had greater exposure to other IT activities (electronic games, TV, mobile phone) and moderate vigorous physical activity. SES was associated with computer exposure, with children from low SES backgrounds having greater home computer exposure, and children from high SES backgrounds having greater school computer exposure.Computer related musculoskeletal outcomes were reported by 10% of children for school computer use and 20% for home computers. The most commonly affected body locations were the neck and back, and 30% of those children reporting outcomes limited their activity participation, 10% took medication and 7% consulted a treating health professional. The use of a range of outcome measures allowed for a better understanding of the impact of children’s computer related musculoskeletal outcomes.Given the significant findings of different relationships between children’s computer exposure patterns at school and home, two models were developed and tested, with one model for school computer exposure and one model for home computer exposure. Path analysis modeling, accounting for user correlates, tested direct relationships and indirect relationships via computer exposure for a range of user correlates. The final school computer exposure model showed direct relationships between gender, somatic complaints, computer exposure and musculoskeletal soreness; and indirect relationships, via computer exposure, between age, computer flow, TV exposure, SES and musculoskeletal soreness. The final home computer exposure model showed direct relationships between gender, age, somatic complaints, computer exposure and musculoskeletal soreness; and indirect relationships, via computer exposure, between age, computer flow, computer anxiety, TV exposure, SES and musculoskeletal soreness.In conclusion, the child specific model tested within this study demonstrated direct relationships between children’s computer exposure and musculoskeletal outcomes. Additionally, direct and indirect relationships were also shown between a range of user correlates, the environment and musculoskeletal outcomes. These findings will assist researchers, teachers and parents to understand the range of potential risk factors for computer related musculoskeletal outcomes. This will also allow researchers to target interventions to child users and their computer environments to ensure children’s computer use is performed in a safe and productive manner.

The fetus is highly responsive to the level of glucocorticoids in the gestational environment. Perturbing glucocorticoids during fetal development could yield long-term consequences. Extending prior research about effects of prenatally exposed synthetic glucocorticoids (sGC) on brain structural development during childhood, we investigated functional brain correlates of cognitive conflict monitoring in term-born adolescents, who were prenatally exposed to sGC. Relative to the comparison group, behavioral response consistency (indexed by lower reaction time variability) and a brain correlate of conflict monitoring (the N2 event-related potential) were reduced in the sGC exposed group. Relatedly, source localization analyses showed that activations in the fronto-parietal network, most notably in the cingulate cortex and precuneus, were also attenuated in these adolescents. These regions are known to subserve conflict detection and response inhibition as well as top-down regulation of stress responses. Moreover, source activation in the anterior cingulate cortex correlated negatively with reaction time variability, whereas activation in the precuneus correlated positively with salivary cortisol reactivity to social stress in the sGC exposed group. Taken together, findings of this study indicate that prenatal exposure to sGC yields lasting impacts on the development of fronto-parietal brain functions during adolescence, affecting multiple facets of adaptive cognitive and behavioral control.