Este trabalho visa apresentar métodos de categorização de variáveis explicativas contínuas em Análise de Sobrevivência. Do ponto de vista clínico, agrupar pacientes em grupos de risco distintos é importante para agilizar tomadas de decisões; entretanto, perda de informação e outros problemas estatísticos podem ocorrer. Portanto, métodos para seleção de pontos de corte e correção dos possíveis problemas gerados pela categorização são criticamente avaliados. Para a aplicação e comparação dos métodos são utilizados dados do Instituto do Coração do Hospital das Clínicas da Faculdade de Medicina da Universidade de São Paulo (InCor - FMUSP), em que a variável fração de ejeção é dicotomizada e tricotomizada. / This dissertation aims to present methods of categorization for continuous variables in Survival Analysis. From a clinical point of view, grouping patients into distinct risk groups is important for accelerating decision-making; however, loss of information and other statistical problems may occur. Therefore, methods for selecting cutpoints and correcting problems generated by categorization are critically evaluated. For the application and comparison of the methods, the dataset from Heart Institute - University of Sao Paulo Medical School (InCor FMUSP) is used, in which the variable ejection fraction is dichotomized and trichotomized.

Objetivo. Uma coorte de crianças foi observada sob duas abordagens: coorte prospectiva e coorte retrospectiva (histórica) e o objetivo foi comparar as estimativas da função de riscos do modelo de Cox entre as duas abordagens e a mesma comparação com o modelo complemento log-log. Além disso, comparar as estimativas da função de riscos pelos modelos de Cox e complemento log-log para cada uma das abordagens, separadamente. Métodos. Quando se estuda o tempo de sobrevida da amamentação sem as covariáveis, foram obtidas as estimativas pela técnica atuarial e modelo complemento log-log da informação retrospectiva e a técnica de Kaplan-Meier para a informação prospectiva. Os modelos de Cox e complemento log-log foram utilizados para estimar a razão de riscos (HR) com covariáveis para as duas abordagens. Resultados. Sem as covariáveis, a comparação entre as duas abordagens mostrou que as estimativas de S(t) pela informação retrospectiva estão um pouco defasadas em relação à medida prospectiva. Com as covariáveis, os resultados dos modelos de Cox e complemento log-log são semelhantes tanto para os dados prospectivos quanto para os retrospectivos. Foram semelhantes, também, os resultados de um mesmo modelo para cada uma das fontes de informação. Conclusões. Em geral, as estimativas foram bastante próximas em quaisquer das comparações. Somente a variável hábito de fumar da mãe durante a gravidez permaneceu nos modelos finais para todas as técnicas utilizadas, com estimativas próximas, reforçando semelhança entre as várias abordagens. / Objective. A cohort of children was observed under two approaches: prospective cohort and retrospective (historical) cohort and the objective was to compare the estimates of the hazard ratio from Cox model between the two approaches and do the same comparison using the complementary log-log model. Moreover, to compare the estimates of the hazard ratio from the two models to each one of the approaches, separately. Methods. When the survival time for breastfeeding is studied without the covariates, the comparison between the results of actuarial technique and the complementary log-log model has been made for the recorded information. The Kaplan-Meier technique has been used with the daily notebook measures. The Cox and complementary log-log models can estimate the risk rate of covariates categories to both approaches. Results. Without the covariates, the comparison between the two information resources showed that retrospective measures give lower estimates than that from the prospective measures. With covariates, the estimates are not so different and led to the same results. Conclusions. The estimates of each one of the comparisons were too close. Just the variable mothers smoking during the pregnancy stayed in the final models for every techniques used, with close estimates, reinforcing likeness between the several approaches.

L’intérêt principal de cette recherche porte sur la validation d’une méthode statistique en pharmaco-épidémiologie. Plus précisément, nous allons comparer les résultats d’une étude précédente réalisée avec un devis cas-témoins niché dans la cohorte utilisé pour tenir compte de l’exposition moyenne au traitement : – aux résultats obtenus dans un devis cohorte, en utilisant la variable exposition variant dans le temps, sans faire d’ajustement pour le temps passé depuis l’exposition ; – aux résultats obtenus en utilisant l’exposition cumulative pondérée par le passé récent ; – aux résultats obtenus selon la méthode bayésienne. Les covariables seront estimées par l’approche classique ainsi qu’en utilisant l’approche non paramétrique bayésienne. Pour la deuxième le moyennage bayésien des modèles sera utilisé pour modéliser l’incertitude face au choix des modèles. La technique utilisée dans l’approche bayésienne a été proposée en 1997 mais selon notre connaissance elle n’a pas été utilisée avec une variable dépendante du temps. Afin de modéliser l’effet cumulatif de l’exposition variant dans le temps, dans l’approche classique la fonction assignant les poids selon le passé récent sera estimée en utilisant des splines de régression. Afin de pouvoir comparer les résultats avec une étude précédemment réalisée, une cohorte de personnes ayant un diagnostique d’hypertension sera construite en utilisant les bases des données de la RAMQ et de Med-Echo. Le modèle de Cox incluant deux variables qui varient dans le temps sera utilisé. Les variables qui varient dans le temps considérées dans ce mémoire sont iv la variable dépendante (premier évènement cérébrovasculaire) et une des variables indépendantes, notamment l’exposition / The main interest of this research is the validation of a statistical method in pharmacoepidemiology. Specifically, we will compare the results of a previous study performed with a nested case-control which took into account the average exposure to treatment to : – results obtained in a cohort study, using the time-dependent exposure, with no adjustment for time since exposure ; – results obtained using the cumulative exposure weighted by the recent past ; – results obtained using the Bayesian model averaging. Covariates are estimated by the classical approach and by using a nonparametric Bayesian approach. In the later, the Bayesian model averaging will be used to model the uncertainty in the choice of models. To model the cumulative effect of exposure which varies over time, in the classical approach the function assigning weights according to recency will be estimated using regression splines. In order to compare the results with previous studies, a cohort of people diagnosed with hypertension will be constructed using the databases of the RAMQ and Med-Echo. The Cox model including two variables which vary in time will be used. The time-dependent variables considered in this paper are the dependent variable (first stroke event) and one of the independent variables, namely the exposure.

Resumo: No Brasil, existem poucos estudos sobre longevidade de vacas de corte, principalmente utilizando ferramentas de análise de sobrevivência na estimação de parâmetros genéticos. Todavia, os critérios para avaliar esta característica são vários, tal que em alguns destes nem todos os registros das vacas nos rebanhos podem ser levados em consideração, devido à metodologia de análise utilizada. Uma variável que é de fácil mensuração e já faz parte da maioria dos controles zootécnicos das fazendas é a idade da vaca ao último parto (IVUP). Neste trabalho, objetivo foi avaliar a longevidade das vacas nos rebanhos utilizando a variável IVUP, por meio da metodologia de análise de sobrevivência. Esta variável foi utilizada mediante um critério para analisar a longevidade produtiva de vacas nos rebanhos. O critério adotado foi a diferença entre a data em relação ao último parto de cada vaca e a data do último parto de cada fazenda. Se esta diferença foi superior a 36 meses, a vaca falhou e foi considerada descartada. Caso contrário, esta vaca foi censurada, indicando que esta ainda poderia ter futuras parições. O critério de 36 meses foi proposto por ser período suficiente para a ocorrência de um novo parto. A metodologia de análise de sobrevivência foi utilizada por considerar dados censurados e não censurados. As variáveis: estação e ano de nascimento de cada vaca, a fazenda e a idade ao primeiro parto foram utilizadas para as análises da variável IVUP. Um estudo prévio por meio de curvas de Kaplan-Meier e o modelo de Cox, utilizando a distribuição gama para os touros, desconsiderando o parentesco entre eles, foram realizados. O software Survival Kit foi empregado para estimação dos parâmetros genéticos, levando em consideração o parentesco entre os animais. Verificou-se que a IVUP apresentou herdabilidade de 0,25, e que seu uso permite avaliar a ... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: In Brazil, there are few studies about stayability, especially using tools of survival analysis in the estimation of breeding values and heritability. The criteria for evaluating this characteristic are different, such that in some of these not all information in the herds of cows can be taken into consideration, because the method of analysis used. In addition, the definitions that consider the date of disposal of the animal affect the use of information obtained in practice, because until the cows come out of the flock may take years, thus the evaluation of their parents is impaired, since the goal is selection. A variable that is easy to measure and is already part of most controls husbandry farms are cow age at last birth (IVUP). This variable was used by one criterion to analyze the productive longevity of cows in herds. The criterion was the difference between the date from the last delivery date of each cow and the last part of each farm. If this difference was more than 36 months, the cow was considered failed and discarded. Otherwise, this cow was censored, indicating that this could still have further parities. The criterion of 36 months was proposed to be sufficient time for the occurrence of a new birth. The methodology of survival analysis was used. The variables season and year of birth of each cow, farm and age at first birth were used for analysis of variable IVUP. The objective of this study was to evaluate the performance of the variable IVUP through estimates of genetic and fixed effect, to study the longevity of cows in the herd. A previous study by Kaplan-Meier and Cox model using the gamma distribution for the bulls, disregarding the relationship between them, were performed. The Survival Kit software was used to estimate the genetic parameters, taking into account the relationship between the animals. It was found that the heritability of 0.25 for IVUP ... (Complete abstract click electronic access below) / Orientador: Danísio Prado Munari / Coorientador: Claudia Cristina Paro da Paz / Coorientador: Raysildo Barbosa Lobo / Banca: João Ademir de Oliveira / Banca: Henrique Nunes De Oliveira / Banca: Roberto Carvalheiro / Banca: Lenira El Faro Zadra / Doutor

Made available in DSpace on 2014-06-11T19:32:15Z (GMT). No. of bitstreams: 0 Previous issue date: 2011-07-05Bitstream added on 2014-06-13T18:43:23Z : No. of bitstreams: 1 caetano_sl_dr_jabo.pdf: 775583 bytes, checksum: a57f6e271b7d34656d680987f933ba7d (MD5) / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) / No Brasil, existem poucos estudos sobre longevidade de vacas de corte, principalmente utilizando ferramentas de análise de sobrevivência na estimação de parâmetros genéticos. Todavia, os critérios para avaliar esta característica são vários, tal que em alguns destes nem todos os registros das vacas nos rebanhos podem ser levados em consideração, devido à metodologia de análise utilizada. Uma variável que é de fácil mensuração e já faz parte da maioria dos controles zootécnicos das fazendas é a idade da vaca ao último parto (IVUP). Neste trabalho, objetivo foi avaliar a longevidade das vacas nos rebanhos utilizando a variável IVUP, por meio da metodologia de análise de sobrevivência. Esta variável foi utilizada mediante um critério para analisar a longevidade produtiva de vacas nos rebanhos. O critério adotado foi a diferença entre a data em relação ao último parto de cada vaca e a data do último parto de cada fazenda. Se esta diferença foi superior a 36 meses, a vaca falhou e foi considerada descartada. Caso contrário, esta vaca foi censurada, indicando que esta ainda poderia ter futuras parições. O critério de 36 meses foi proposto por ser período suficiente para a ocorrência de um novo parto. A metodologia de análise de sobrevivência foi utilizada por considerar dados censurados e não censurados. As variáveis: estação e ano de nascimento de cada vaca, a fazenda e a idade ao primeiro parto foram utilizadas para as análises da variável IVUP. Um estudo prévio por meio de curvas de Kaplan-Meier e o modelo de Cox, utilizando a distribuição gama para os touros, desconsiderando o parentesco entre eles, foram realizados. O software Survival Kit foi empregado para estimação dos parâmetros genéticos, levando em consideração o parentesco entre os animais. Verificou-se que a IVUP apresentou herdabilidade de 0,25, e que seu uso permite avaliar a... / In Brazil, there are few studies about stayability, especially using tools of survival analysis in the estimation of breeding values and heritability. The criteria for evaluating this characteristic are different, such that in some of these not all information in the herds of cows can be taken into consideration, because the method of analysis used. In addition, the definitions that consider the date of disposal of the animal affect the use of information obtained in practice, because until the cows come out of the flock may take years, thus the evaluation of their parents is impaired, since the goal is selection. A variable that is easy to measure and is already part of most controls husbandry farms are cow age at last birth (IVUP). This variable was used by one criterion to analyze the productive longevity of cows in herds. The criterion was the difference between the date from the last delivery date of each cow and the last part of each farm. If this difference was more than 36 months, the cow was considered failed and discarded. Otherwise, this cow was censored, indicating that this could still have further parities. The criterion of 36 months was proposed to be sufficient time for the occurrence of a new birth. The methodology of survival analysis was used. The variables season and year of birth of each cow, farm and age at first birth were used for analysis of variable IVUP. The objective of this study was to evaluate the performance of the variable IVUP through estimates of genetic and fixed effect, to study the longevity of cows in the herd. A previous study by Kaplan-Meier and Cox model using the gamma distribution for the bulls, disregarding the relationship between them, were performed. The Survival Kit software was used to estimate the genetic parameters, taking into account the relationship between the animals. It was found that the heritability of 0.25 for IVUP ... (Complete abstract click electronic access below)

Submitted by (ana.araujo@ufrpe.br) on 2016-05-25T15:49:21Z No. of bitstreams: 1 Carlos Sergio Araujo dos Santos.pdf: 1655656 bytes, checksum: 7687ae013a71aadd4e9e9bf4783a172f (MD5) / Made available in DSpace on 2016-05-25T15:49:21Z (GMT). No. of bitstreams: 1 Carlos Sergio Araujo dos Santos.pdf: 1655656 bytes, checksum: 7687ae013a71aadd4e9e9bf4783a172f (MD5) Previous issue date: 2010-01-15 / Conselho Nacional de Pesquisa e Desenvolvimento Científico e Tecnológico - CNPq / Box and Cox (1964) developed a numerical procedure to transform the response variable such that the transformed variable should be as closed as possible to the normal distribution. The introduction of a new class of non linear symetric transformed models aims to extended the Box and Cox models to a general class of symetric models. The new class of models inclued all the continuos symmetric distributions with a possible non linear structure to the mean, making possible the use of the new class of regression models. It was applied in the estimate of volumes of the Eucalyptus tereticornis clones, with 7,5 years, planted in the Experimental Station of Araripe of the Agronomic Institute of Pernambuco (IPA), in the municipality of Araripina, in the semiarid of Pernambuco. The non linear model used as pattern was the Schumacher and Hall model. The results indicates that the transformed model with t-Student erros with two degrees of freedon adjusted better to the data set. / Box e Cox (1964) desenvolveram um procedimento numérico para escolher uma transformação da resposta tal que a distribuição da variável transformada esteja o mais próximo possível da distribuição normal. A introdução de uma nova classe de modelos simétricos transformados não lineares visa estender os modelos de Box e Cox para uma classe geral dos modelos simétricos. Esta nova classe de modelos inclui todas as distribuições contínuas simétricas com uma possível estrutura não linear para a média e capacitando o ajustamento de uma larga extensão de modelos para vários tipos de dados. Para ilustrar a utilidade dessa nova classe de modelos de regressão foi realizada uma aplicação na estimativa dos volumes de clones de Eucalyptus tereticornis com 7,5 anos oriundos de um experimento que está sendo realizado no Campo Experimental do Araripe do Instituto Agronômico de Pernambuco (IPA), localizado no Município de Araripina, no semiárido Pernambucano. O modelo não-linear utilizado para explicar os dados foi o modelo Schumacher-Hall. Diante dos resultados obtidos se concluí que o modelo transformado com erros t-Student com dois graus de liberdade foi o que melhor se ajustou os dados.

Objetivo. Uma coorte de crianças foi observada sob duas abordagens: coorte prospectiva e coorte retrospectiva (histórica) e o objetivo foi comparar as estimativas da função de riscos do modelo de Cox entre as duas abordagens e a mesma comparação com o modelo complemento log-log. Além disso, comparar as estimativas da função de riscos pelos modelos de Cox e complemento log-log para cada uma das abordagens, separadamente. Métodos. Quando se estuda o tempo de sobrevida da amamentação sem as covariáveis, foram obtidas as estimativas pela técnica atuarial e modelo complemento log-log da informação retrospectiva e a técnica de Kaplan-Meier para a informação prospectiva. Os modelos de Cox e complemento log-log foram utilizados para estimar a razão de riscos (HR) com covariáveis para as duas abordagens. Resultados. Sem as covariáveis, a comparação entre as duas abordagens mostrou que as estimativas de S(t) pela informação retrospectiva estão um pouco defasadas em relação à medida prospectiva. Com as covariáveis, os resultados dos modelos de Cox e complemento log-log são semelhantes tanto para os dados prospectivos quanto para os retrospectivos. Foram semelhantes, também, os resultados de um mesmo modelo para cada uma das fontes de informação. Conclusões. Em geral, as estimativas foram bastante próximas em quaisquer das comparações. Somente a variável hábito de fumar da mãe durante a gravidez permaneceu nos modelos finais para todas as técnicas utilizadas, com estimativas próximas, reforçando semelhança entre as várias abordagens. / Objective. A cohort of children was observed under two approaches: prospective cohort and retrospective (historical) cohort and the objective was to compare the estimates of the hazard ratio from Cox model between the two approaches and do the same comparison using the complementary log-log model. Moreover, to compare the estimates of the hazard ratio from the two models to each one of the approaches, separately. Methods. When the survival time for breastfeeding is studied without the covariates, the comparison between the results of actuarial technique and the complementary log-log model has been made for the recorded information. The Kaplan-Meier technique has been used with the daily notebook measures. The Cox and complementary log-log models can estimate the risk rate of covariates categories to both approaches. Results. Without the covariates, the comparison between the two information resources showed that retrospective measures give lower estimates than that from the prospective measures. With covariates, the estimates are not so different and led to the same results. Conclusions. The estimates of each one of the comparisons were too close. Just the variable mothers smoking during the pregnancy stayed in the final models for every techniques used, with close estimates, reinforcing likeness between the several approaches.

The input collected from warranty claims data links customer feedback with product quality. Results from warranty claim analysis can potentially improve product quality, customer relationships and positively affect business. However working on warranty claims data holds many challenges that requires a significant share of time devoted to data cleaning and data processing. The purpose of warranty claims analysis is to get the comprehensive overview of the reliability, costs and quality of household appliances produced by ASKO. While there are different ways to approach this problem, we will focus on non-parametric and semi-parametric methods, by using Kaplan-Meier estimators and Cox proportional hazard model respectively. These kinds of models are time dependent and therefore used for prediction of household appliance reliability. Even though non-parametric models are quite informative they cannot handle additional characteristics about observable product hence the semi-parametric Cox proportional hazard model was proposed. Apart from the reliability analysis, we will also predict warranty costs with probit model and observe inequality in household appliances part failures as a part of quality control analysis. Described methods were selected due to the fact that the warranty claims analysis will be practiced in future by ASKO’s quality department and therefore straight forward methods with very informative results are needed.

Au cours des études de survie, certains facteurs ayant un rôle pronostique peuvent être inobservés ou indisponibles. Dans le cadre des essais cliniques randomisés où deux groupes de traitement sont comparés par un modèle de Cox, « l’oubli » de tels facteurs dans le modèle entraîne une sous-estimation en valeur absolue de l’effet du traitement. L’objectif de la thèse est de préciser le rôle de différents déterminants de ce biais et de suggérer l’utilisation de méthodes pouvant le réduire, avec un intérêt particulier pour les essais de prévention à l'infection VIH où de tels facteurs peuvent exister. L’effet des principaux facteurs pouvant influencer le biais est précisé dans une première partie de la thèse. Si certains facteurs sont connus de longue date, d’autres, comme la durée de l’essai n’ont, à notre connaissance, pas été étudiés. L’approche utilisée repose sur l’utilisation d’un « vrai » modèle de type à risques proportionnels. Dans ce modèle, l’effet des covariables « oubliées » est résumé par l’introduction dans l’expression du risque instantané d’un terme aléatoire de « fragilité » propre à chaque sujet. Le biais cherché est calculé comme la limite asymptotique, sous le modèle marginal correspondant au « vrai » modèle, du rapport des risques instantanés déduit du modèle de Cox n’incluant pas la fragilité. Les résultats montrent une nette augmentation du biais, en valeur absolue, avec la durée de l’essai. Cette augmentation est particulièrement marquée pour des distributions de fragilité continues comme celles pouvant être rencontrées en pratique, par rapport à des fragilités binaires. Par ailleurs, les résultats antérieurs de la littérature sont confirmés et précisés. Dans un second temps, les conséquences d’une variation de la fragilité au cours du temps sur le biais sont recherchées. Plus précisément, la situation envisagée est celle qui est rencontrée dans les essais de prévention contre le VIH effectués auprès de populations « instables » vis-à-vis du risque, comme les jeunes ou les prostituées en Afrique sub-saharienne. Ces populations montrent une hétérogénéité comportementale liée aux changements de partenaires sexuels dont le statut VIH est par ailleurs inconnu. Il s’agit d’évaluer le biais en présence d’une telle fragilité « intermittente » au moyen de simulations reflétant les situations réelles. Les résultats montrent que le biais dû à l’omission de la fragilité dans l’analyse, bien qu’inférieur au biais obtenu dans le cas d’une fragilité constante au cours du temps, reste significatif et doit être considéré. Les différentes fragilités générées au cours du suivi sont soit indépendantes entre elles, soit corrélées. Enfin, dans le cas d’une fragilité supposée constante au cours du temps, l’intérêt de l’utilisation d’un essai en « cross-over » est recherché. Dans un tel essai, les sujets exempts d’évènement après un temps de suivi donné changent de groupe de traitement. Dans le cadre des essais de prévention VIH, Auvert & al [2011] ont montré en particulier une diminution du biais avec un schéma en cross-over comparativement à un schéma parallèle classique, en utilisant une fragilité catégorielle dans une étude de simulation. Buyze & Goetghebeur [2011] ont également montré les avantages du cross-over, en particulier concernant l’efficacité relative d’un test de comparaison des deux groupes, en utilisant une fragilité de distribution gamma ou log-normale. Ces résultats sont précisés en calculant formellement le biais asymptotique dans l’estimation du risque relatif pour les différentes distributions possibles de la fragilité omise, continue ou catégorielle. Les résultats obtenus sont nettement en faveur du cross-over, avec une diminution du biais entre 60% et 90% et une amélioration de l’efficacité du test. Le temps de changement de groupe optimal est également recherché. Il apparaît que celui-ci dépend essentiellement de la durée le l'essai et de la valeur de l'effet traitement. / In survival analysis, some prognostic factors can be unobserved or unavailable. In randomized clinical trials framework where two treatment groups are compared in the Cox model setting, the omission of such factors in the model leads to an under-estimation in absolute value of the treatment effect. The aim of the project is to better understand the determinants of this bias and to suggest the use of methods that could reduce it, with a particular interest in HIV prevention trials where such factors are likely to exist. In a first step, the role of the main determinants of the bias is highlighted. While some of them have long been identified, others like trial duration have never been considered to our knowledge. The bias was calculated as the asymptotic limit of the maximum likelihood estimator of the treatment effect when the analysis is done following a proportional hazard model which no takes into account the frailty. The results show a clear increase of the bias in absolute value with the trial duration. This increase is especially marked with continuous frailty distributions, such as those which can be encountered in practice, compared to binary frailties. Also, some previous results have been confirmed. In a second step, a frailty depending on time is considered, as it can be encountered in HIV clinical trials including “unstable” population about infection risk, as prostitutes or young people in sub-Saharan Africa. These populations present a behavioral heterogeneity linked to the change of partner whose the HIV status is unknown. The bias is estimated using a transient frailty with a simulation study reflecting real-life situations. The results show that when the frailty is regenerated during the follow-up, the bias caused by its omission, although lower than the bias obtained with a time-independent frailty, stays significant and has to be considered. The different frailties generated during the follow-up are independent or correlated. Finally, with a “stable” population whose the frailty can be supposed constant over time, the interest of the use of the “cross-over” is searched. In such trial, the subjects which have not presented the event after a given time of follow-up, change of treatment group. In HIV prevention trials framework, Auvert & al. (2011) have particularly shown a decrease of the bias with a cross-over design comparatively to a parallel design, using a categorized frailty in a simulation study. Buyze & Goetghebeur (2011) have also shown the advantages of the cross-over, particularly about the relative efficiency of the comparison test of the two treatment groups, using gamma or log-normal frailty distributions. These results are specified calculating explicitly the asymptotic bias of the hazard ratio estimate for different possible distributions of the omitted frailty, continuous or categorized. The obtained results are clearly in favor of the cross-over, with a decrease of the bias between 60% et 90% and a significant improvement of the efficiency of the comparison test. The optimal switch time and its prognostic factors are searched. It appears that it essentially depends on the trial duration and is little affected by the frailty distribution or the value of the treatment effect.

Title: Joint Models for Longitudinal and Time-to-Event Data Author: Jana Vorlíčková Department: Department of Probability and Mathematical Statistics Supervisor: doc. RNDr. Arnošt Komárek, Ph.D., Department of Probability and Mathematical Statistics Abstract: The joint model of longitudinal data and time-to-event data creates a framework to analyze longitudinal and survival outcomes simultaneously. A commonly used approach is an interconnection of the linear mixed effects model and the Cox model through a latent variable. Two special examples of this model are presented, namely, a joint model with shared random effects and a joint latent class model. In the thesis we focus on the joint latent class model. This model assumes an existence of latent classes in the population that we are not able to observe. Consequently, it is assumed that the longitudinal part and the survival part of the model are independent within one class. The main intention of this work is to transfer the model to the Bayesian framework and to discuss an estimation procedure of parameters using a Bayesian statistic. It consists of a definition of the model in the Bayesian framework, a discussion of prior distributions and the derivation of the full conditional distributions for all parameters of the model. The model's ability to...