Chez les patients porteurs d'un cancer bronchique avancé non éligible à un traitement curatif, l'objectif du traitement n'est pas la guérison mais le soulagement des symptômes et l'amélioration de la survie. Le pronostic est lié aux caractéristiques du patient et de la tumeur, à la réponse tumorale après chimiothérapie et à la prise en charge globale décidée conjointement avec l'équipe médicale, le patient et son entourage. Il convient donc d'évaluer précocement ces facteurs pronostiques pour optimiser le rapport bénéfice-risque des traitements entrepris, dans le cadre des thérapeutiques anti-cancéreuses ou de façon plus globale.Dans le cadre de ce travail, des déterminants éthiques, cliniques et paracliniques de la prise en charge et du pronostic des patients porteurs d'un cancer bronchique localement avancé ou métastatique ont été recherchés. Dans une 1ère partie, les déterminants du pronostic de patients traités par chimiothérapie (étude « Pharmacogenoscan-poumon ») ont été étudiés : association entre le niveau d'expression de biomarqueurs tumoraux protéiques et la réponse à la chimiothérapie et intérêt de l'évaluation de la réponse tumorale selon les critères RECIST. Nous avons, dans une seconde partie, analysé les éléments influençant la prise en charge et le pronostic de patients porteurs d'un cancer bronchique développant une défaillance d'organe. Trois situations ont été décrites : celle de patients dès l'apparition de cette défaillance d'organe, celle de l'ensemble des patients porteurs d'un cancer bronchique admis en réanimation, et uniquement de ceux porteurs d'une mutation tumorale oncogénique. / In patients with advanced lung cancer non-eligible for curative treatment, the goal of care is not healing but improving symptoms and survival. The prognosis is related to patient and tumor characteristics, response to chemotherapy and its overall management decided jointly with the medical team, the patient and his family. It is therefore necessary to assess early these prognostic factors to optimize the risk-benefit ratio of treatment undertaken.As part of this work, ethical, clinical and laboratory determinants of management and prognosis of locally advanced or metastatic lung cancer patients were investigated. In the first part, the prognosis' determinants of patients treated with chemotherapy (« Pharmacogenoscan-poumon » study) were studied: association between the level of expression of protein biomarkers and tumor response to chemotherapy and assessment of interest of tumor response according to RECIST criteria. We have, in the second part, analyzed the factors influencing the management and prognosis of lung cancer patients developing organ failure. Three situations were described: patients at the time of the development of this organ failure, lung cancer patients admitted to intensive care units, the overall population and only those carrying an oncogenic mutation.

Notícias de escassez no mundo inteiro, dada a crescente demanda e o rigor na triagem clínica, levaram a necessidade de investigar métodos que mensurem o comportamento de retorno do doador de sangue, sobretudo o indivíduo que manifesta a intenção voluntária em doar. Curvas de Sobrevivência entre outros métodos estatísticos são amplamente estudados na literatura com o intuito de obter uma estimativa da chance de um doador vir a realizar uma subseqüente doação, associado ao seu perfil. O objetivo do presente estudo é identificar modelos estatísticos capazes de descrever esse comportamento utilizan do os registros do Centro Regional de Hemoterapia de Ribeirão Preto. A cons trução de modelos de longa-duração, por exemplo, pode ser um meio de evidenciar possíveis subgrupos mais propensos a retornar, além de estimar a proporção de doadores que jamais retornarão. Entre os resultados, obser vamos que apenas 40% dos doadores voluntários retornaram após um ano decorrido da primeira doação, e 20% destes jamais retornarão. O ajuste do modelo longaduração possibilitou ainda indicar alguns subgrupos de doadores prováveis e improváveis de retornar, porém tais resultados reforçam as evidências de que a motivação intrínseca é o que leva o individuo a retornar. / Reports of worldwide shortages due the increased demand and rigor of clinical screening have led to the necessity to investigate methods that measure blood donor return behavior, mainly regarding individuals who manifest the voluntary intention to donate. Survival curves, among others statistical methods, have been extensively studied in the literature in order to estimate the likelihood of a donor to make another donation, associated with his profile. The aim of the present study was to identify statistical models describing this behavior using information from the Regional Hemotherapy Center of Ribeirão Preto. The construction of long-term survival model can be a useful instrument for determining the groups more likely to donate, as well as the proportion of donors who will never return. The results obtained revealed that only 40% of the volunteer primary donors return for a new donation one year after the first, with the estimate that 20% will never return. The construction of long-term survival model still facilitated to indicate some groups likely and unlike ly donors to donate, even so such re sults reinforce the evidences that the intrinsic motivation is what prompts a donor to return.

MSc (Statistics) / Department of Statistics / The baseline parameter vector in a discrete-time survival model is determined by the number of time points. The larger the number of the time points, the higher the dimension of the baseline parameter vector which often leads to biased maximum likelihood estimates. One of the ways to overcome this problem is to use a simpler parametrization that contains fewer parameters. A simulation approach was used to compare the accuracy of three variants of penalised regression spline methods in smoothing the baseline hazard function. Root mean squared error (RMSE) analysis suggests that generally all the smoothing methods performed better than the model with a discrete baseline hazard function. No single smoothing method outperformed the other smoothing methods. These methods were also applied to data on age at rst alcohol intake in Thohoyandou. The results from real data application suggest that there were no signi cant di erences amongst the estimated models. Consumption of other drugs, having a parent who drinks, being a male and having been abused in life are associated with high chances of drinking alcohol very early in life. / NRF

Thesis (Ph.D. (Statistics)) -- University of Limpopo, 2022 / This thesis is based on application of frailty and copula models to under five child mortality data set in South Africa. The main purpose of the study was to apply sample splitting techniques in a survival analysis setting and compare clustered survival models considering left truncation to the under five child mortality data set in South Africa. The major contributions of this thesis is in the application of the shared frailty model and a class of Archimedean copulas in particular, Clayton-Oakes copula with completely monotone generator, and introduction of sample splitting techniques in a survival analysis setting. The findings based on shared frailty model show that clustering effect was sig nificant for modelling the determinants of time to death of under five children, and revealed the importance of accounting for clustering effect. The conclusion based on Clayton-Oakes model showed association between survival times of children from the same mother. It was found that the parameter estimates for the shared frailty and the Clayton-Oakes models were quite different and that the two models cannot be comparable. Gender, province, year, birth order and whether a child is part of twin or not were found to be significant factors affect ing under five child mortality in South Africa. / NRF-TDG Flemish Interuniversity Council Institutional corporation (VLIR-IUC) VLIR-IUC Programme of the University of Limpopo

Neste trabalho propomos a distribuição Weibull-Binomial-Negativa (WBN) considerando uma estrutura de ativação latente para explicar a ocorrência do evento de interesse, em que o número de causas competitivas é modelado pela distribuição Binomial Negativa, e os tempos não observados devido às causas seguem a distribuição Weibull. Em geral, as causas competitivas podem ter diferentes mecanismos de ativação, sendo assim os casos de primeira ativação, última ativação e ativação aleatória foram considerados no estudo. Desse modo o modelo proposto inclui uma ampla distribuição, tais como Weibull-Geométrico (WG) e Exponencial-Poisson Complementar (EPC), introduzidas por Barreto-Souza et al. (2011) e G. et al. (2011), respectivamente. Baseando-nos na mesma estrutura, consideramos o modelo de regressão locação-escala baseado na distribuição proposta (WBN) e o modelo para dados de sobrevivência com fração de cura. Os principais objetivos deste trabalho é estudar as propriedades matemáticas dos modelos propostos e desenvolver procedimentos de inferências desde uma perspectiva clássica e Bayesiana. Além disso, as medidas de diagnóstico Bayesiana baseadas na \'psi\'-divergência (Peng & Dey, 1995; Weiss, 1996), que inclui como caso particular a medida de divergência Kullback-Leibler (K-L), foram consideradas para detectar observações influentes / In this work we propose the Weibull-Negative-Binomial (WNB) considering a latent activation structure to explain the occurrence of an event of interest, where the number of competing causes are modeled by the Negative Binomial distribution and the no observed time due to the causes following the Weibull distribution. In general, the competitive causes may have different activation mechanisms, cases of first, last and random activation were considered in the study. Thus, the proposed model includes a wide distribution such as Weibull-Geometric distribution (WG) and Exponential-Poisson complementary (EPC) introduced by (Barreto-Souza et al., 2011) and (G. et al., 2011) respectively. Based on the same structure, we propose a location-scale regression model based on the proposed distribution (WNB) and the model for survival data with cure fraction. The main objectives of this work is to study the mathematical properties of the proposed models and develop procedures inferences from a classical and Bayesian perspective. Moreover, the Bayesian diagnostic measures based on the \'psi\'-divergence (Peng & Dey, 1995; Weiss, 1996), which includes Kullback-Leibler (K-L) divergence measure as a particular case, were considered to detect influential observations

Health problems of the human beings in a society are one of the main components of the social security systems due to the dimension of the financial burden it might bring on individuals, employers, insurance companies and governments. Morbidity measures, such as incidence and prevalence of a specific disease in a certain population enable researchers to estimate for individuals the probability of being diagnosed or being prone to the diseases. This information is usually not tractable because of the non-availability of the convenient data or recordings for many countries as well as Turkey. Even if it is available, it is commonly limited with largely varying characteristics about the type and coverage of the diseases. In this regard, the pattern that a population follows for an acute disease may not be the same for chronic diseases. Having those indicators determined for a group of insureds will enable underwriters to have more profitable and economical premium calculation and precision on required reserve estimation. v Based on their characteristics such as acute or chronic behaviour, the gender, and the location of residency of people, the diseases show different behaviour on their occurrences. From the insurer

Made available in DSpace on 2016-06-02T20:04:52Z (GMT). No. of bitstreams: 1 4552.pdf: 1449121 bytes, checksum: 5d02e70bde72ea9ef3c257c80ceed1dc (MD5) Previous issue date: 2012-05-03 / Financiadora de Estudos e Projetos / In this dissertation we propose new models for survival with cure fraction to describe the biological mechanism of the event of interest (cancer) in studies of carcinogenesis in the presence of competing causes latent independent or correlated. The formulation of new models is based on stochastic modeling of the occurrence of tumors through three stages: initiation of a tumor not detectable, promotion and progression of the tumor to a detectable cancer. These models allow a simple pattern of the dynamics of tumor growth, and incorporate into the analysis features of the stage of tumor progression that is not possible in most survival models with cure fraction commonly used. For the proposed models, the inferential process was discussed in terms of classical and Bayesian point of view. Simulations studies were conducted in order to analyze the asymptotical properties of the classical estimation procedure. Real data applications demonstrate of use of the models. / Neste trabalho propomos modelos de sobrevivência com fração de cura para descrever o mecanismo biológico da ocorrência do evento de interesse (câncer) em estudos da carcinogênese na presença de causas competitivas latentes independentes ou correlacionadas. A formulação dos novos modelos é baseada na modelagem estocástica da ocorrência dos tumores através de três estágios: iniciação de um tumor não detectável, promoção e a progressão do tumor até um câncer detectável. Estes modelos permitem um padrão simples da dinâmica de crescimento do tumor, além de incorporarem características do estágio de progressão do tumor, que não é possível na maioria dos modelos de sobrevivência com fração de cura comumente utilizados. Para os modelos propostos, discutimos o processo inferencial do ponto de vista clássico e bayesiano. Estudos de simulações foram feitos com o objetivo de analisar as propriedades assintóticas do processo de estimação clássico. Aplicações a conjuntos de dados reais mostraram a aplicabilidade dos modelos.

Neste trabalho propomos a distribuição Weibull-Binomial-Negativa (WBN) considerando uma estrutura de ativação latente para explicar a ocorrência do evento de interesse, em que o número de causas competitivas é modelado pela distribuição Binomial Negativa, e os tempos não observados devido às causas seguem a distribuição Weibull. Em geral, as causas competitivas podem ter diferentes mecanismos de ativação, sendo assim os casos de primeira ativação, última ativação e ativação aleatória foram considerados no estudo. Desse modo o modelo proposto inclui uma ampla distribuição, tais como Weibull-Geométrico (WG) e Exponencial-Poisson Complementar (EPC), introduzidas por Barreto-Souza et al. (2011) e G. et al. (2011), respectivamente. Baseando-nos na mesma estrutura, consideramos o modelo de regressão locação-escala baseado na distribuição proposta (WBN) e o modelo para dados de sobrevivência com fração de cura. Os principais objetivos deste trabalho é estudar as propriedades matemáticas dos modelos propostos e desenvolver procedimentos de inferências desde uma perspectiva clássica e Bayesiana. Além disso, as medidas de diagnóstico Bayesiana baseadas na \'psi\'-divergência (Peng & Dey, 1995; Weiss, 1996), que inclui como caso particular a medida de divergência Kullback-Leibler (K-L), foram consideradas para detectar observações influentes / In this work we propose the Weibull-Negative-Binomial (WNB) considering a latent activation structure to explain the occurrence of an event of interest, where the number of competing causes are modeled by the Negative Binomial distribution and the no observed time due to the causes following the Weibull distribution. In general, the competitive causes may have different activation mechanisms, cases of first, last and random activation were considered in the study. Thus, the proposed model includes a wide distribution such as Weibull-Geometric distribution (WG) and Exponential-Poisson complementary (EPC) introduced by (Barreto-Souza et al., 2011) and (G. et al., 2011) respectively. Based on the same structure, we propose a location-scale regression model based on the proposed distribution (WNB) and the model for survival data with cure fraction. The main objectives of this work is to study the mathematical properties of the proposed models and develop procedures inferences from a classical and Bayesian perspective. Moreover, the Bayesian diagnostic measures based on the \'psi\'-divergence (Peng & Dey, 1995; Weiss, 1996), which includes Kullback-Leibler (K-L) divergence measure as a particular case, were considered to detect influential observations

MSc (Statistics) / Department of Statistics / This study explores the use of exible link functions in discrete survival models through a simulation study and an application to the Swaziland Demographic and Health Survey (SDHS) data. The objective of the research study is to perform simulation exercises in order to compare the e ectiveness of di erent families of link functions and to construct a discrete multilevel survival model for age at rst marriage among women in Swaziland using a exible link function. The Pareto hazard model, Pregibon and Gosset families of link functions were considered in models with and without unobserved heterogeneity. The Pareto model where the family parameter is estimated from the data was found to outperform the other models, followed by the Pregibon and the Gosset family of link functions. The results from both simulation study and real data analysis of the SDHS data illustrated that, misspecication of the link function causes bias on the estimation of results. This demonstrates the importance of choosing the right link. The ndings of this study reveal that women who are highly educated, stay in the Manzini and Shiselweni region, those who reside in urban areas were more likely to marry later compared to their counterparts in Swaziland. The results also reveal that the proportion of early rst marriages is declining since the di erence among birth cohorts is found to be very high, with women of younger cohorts getting married later compared to older women. / NRF

Made available in DSpace on 2016-06-02T20:06:06Z (GMT). No. of bitstreams: 1 4336.pdf: 1594022 bytes, checksum: ff2370b4d516b9cdf6dd6da3be557c42 (MD5) Previous issue date: 2012-03-16 / Financiadora de Estudos e Projetos / Experiments involving chemical induction of carcinogens in animals are common in the biological area. Interest in these experiments is, in general, evaluating the chemopreventive effect of a substance in the destruction of damaged cells. In this type of study, two variables of interest are the number of induced tumors and their development times. We explored the use of statistical model proposed by Kokoska (1987) for the analysis of experimental data of chemoprevention of cancer in animals. We flexibility the Kokoska s model, subsequently used by Freedman (1993), whereas for the variable number of tumors induced Conway-Maxwell Poisson (COM-Poisson) distribution. This distribution has demonstrated efficiency due to its great flexibility, when compared to other discrete distributions to accommodate problems related to sub-dispersion and super-dispersion often found in count data. The purpose of this paper is to adapt the theory of long-term destructive model (Rodrigues et al., 2011) for experiments chemoprevention of cancer in animals, in order to evaluate the effectiveness of cancer treatments. Unlike the proposed Rodrigues et al. (2011), we formulate a model for the variable number of detected malignant tumors per animal, assuming that the probability of detection is no longer constant, but dependent on the time step. This is an extremely important approach to cancer chemoprevention experiments, because it makes the analysis more realistic and accurate. We conducted a simulation study, in order to evaluate the efficiency of the proposed model and to verify the asymptotic properties of maximum likelihood estimators. We also analyze a real data set presented in the article by Freedman (1993), to demonstrate the efficiency of the COM-Poisson model compared to results obtained by him with the Poisson and Negative Binomial distributions. / Experimentos que envolvem a indução química de substâncias cancerígenas em animais são comuns na área biológica. O interesse destes experimentos é, em geral, avaliar o efeito de uma substância quimiopreventiva na destruição das células danificadas. Neste tipo de estudo, duas variáveis de interesse são o número de tumores induzidos e seus tempos de desenvolvimento. Exploramos o uso do modelo estatístico proposto por Kokoska (1987) para a análise de dados de experimentos de quimioprevenção de câncer em animais. Flexibilizamos o modelo de Kokoska (1987), posteriormente utilizado por Freedman (1993), considerando para a variável número de tumores induzidos a distribuição Conway-Maxwell Poisson (COM-Poisson). Esta distribuição tem demonstrado eficiência devido à sua grande flexibilidade, quando comparada a outras distribuições discretas, para acomodar problemas relacionados à subdispersão e sobredispersão encontrados frequentemente em dados de contagem. A proposta deste trabalho consiste em adaptar a teoria de modelo destrutivo de longa duração (Rodrigues et al., 2011) para experimentos de quimioprevenção do câncer em animais, com o propósito de avaliar a eficiência de tratamentos contra o câncer. Diferente da proposta de Rodrigues et al. (2011), formulamos um modelo para a variável número de tumores malignos detectados por animal, supondo que sua probabilidade de detecção não é mais constante, e sim dependente do instante de tempo. Esta é uma abordagem extremamente importante para experimentos quimiopreventivos de câncer, pois torna a análise mais realista e precisa. Realizamos um estudo de simulação com o propósito de avaliar a eficiência do modelo proposto e verificar as propriedades assintóticas dos estimadores de máxima verossimilhança. Analisamos também um conjunto de dados reais apresentado no artigo de Freedman (1993), visando demonstrar a eficiência do modelo COM-Poisson em relação aos resultados por ele obtidos com as distribuições Poisson e Binomial Negativa.