INTRODUÇÃO: As recomendações alimentares na infância são a exclusividade do leite materno (LM) até o 6º mês, sua extensão mínima até o 24º mês e introdução programada de alimentos semissólidos e sólidos até os dois anos. O padrão alimentar na infância influencia preferências sensoriais e indicadores de saúde nos ciclos de vida subsequentes. OBJETIVO: Analisar a tendência secular da alimentação de crianças brasileiras menores de cinco anos nas três últimas décadas. MÉTODOS: Os dados foram provenientes da amostra de crianças menores de cinco anos das três Pesquisas de Demografia e Saúde realizadas no Brasil em 1986, 1996 e 2006. O aleitamento materno (AM) foi descrito segundo indicadores propostos pela Organização Mundial da Saúde (2008). As medianas do AM e do aleitamento materno exclusivo (AME) e os seus fatores preditores foram estimadas por modelo de regressão de Cox. Os padrões alimentares (PA) foram identificados por análise de componentes principais (ACP) e, em seguida, foram calculados os escores de cada PA. Estes PA foram incluídos como desfechos em modelos de efeitos mistos cujos fatores preditores associados foram aqueles referentes à saúde materno-infantil e sociodemográficos. RESULTADOS: Entre 1986 e 2006, o percentual de crianças expostas ao LM foi de 91 para 97 por cento e a duração mediana, de 6 para 12 meses. Entre 1996 e 2006 a mediana do AME aumentou de 0,7 para 2 meses. A amamentação na 1ª hora de vida foi fator protetor para a duração do AME e do AM. A duração mediana do AME foi homogênea entre as regiões, associada de forma direta com relação de pré-natal, escolaridade materna e índice de riqueza. A duração mediana do AM foi associada de forma inversa com nascimento por parto cesáreo em hospital privado, intervalo interpartal inferior a 24 meses e difere segundo região geográfica, sendo superior no Norte brasileiro. Foram identificados quatro padrões alimentares para as crianças com idade entre 6 e 59 meses: PA1(iogurte, carnes, tubérculos, hortaliças, frutas), PA2(líquidos, leites não-maternos, carnes e carga fatorial negativa para leite materno e papas à base de farinhas enriquecidas), PA3(líquidos, sucos de frutas, papas à base de amido industrializado, papas à base de farinha enriquecida, iogurte e carga fatorial negativa para carnes vermelhas) e PA4(leite não-materno, fórmulas e carga fatorial negativa papas à base de farinhas enriquecidas e ovopeixefrango). A prática dos padrões PA1 e PA3 foi superior entre crianças que recebiam leite materno. O padrão PA1 foi distribuído de forma homogênea entre as regiões enquanto os outros padrões alimentares apresentaram comportamento distinto entre as regiões. A mudança mais expressiva no período analisado foi a virtual substituição do PA4 pelo PA3, cuja composição de alimentos se aproxima mais das recomendações para faixa etária. CONCLUSÃO: O quadro pró-aleitamento materno é centrado principalmente na exposição universal ao LM e na sustentação desta exposição, refletida no incremento da sua duração mediana. Entretanto, o avanço na duração do AME é menos expressivo no período. O aumento da exposição aos alimentos sólidos entre as crianças menores de 12 meses representa a principal alteração dos padrões alimentares entre 1996 e 2006. Três padrões alimentares das crianças menores de cinco anos são influenciados regionalmente e um deles socialmente. / INTRODUCTION: The infant and recommendations are exclusive breastfeeding (EBF) at six months, breastfeeding at least 24th months and the programmed introduction of solid, semi-solid or soft foods with breast milk until 24 months. Early dietary patterns can explain the sensory preferences and the health indicators throughout the life course. OBJECTIVE: To analyze the secular trend of Brazilian young child feeding practices in last three decades. METHODS: The under five children sample was from three probabilistic Brazilian Demographic Health Surveys carried out in 1986, 1996 and 2006. The breastfeeding (BF) was described according to the indicators from World Health Organization (WHO, 2008). The BF and exclusive breastfeeding (EBF) medians and the predictor factors were estimated using Cox regression model. The dietary patterns (DP) were identified by principal component analysis (PCA) and the dietary pattern scores were included as outcome variables in logistic mixed model. The predictive variables were sociodemographic, maternal and infant health. The complex design sample was considered in statistical analysis. RESULTS: In the period between 1986 and 2006, the percentage of children exposed to breast milk increased from 91 to 97 per cent and the BF median increased from 6 to 12 months and the EBF increased from 0.7 to 2.0 months. Early initiation of breastfeeding was a protective factor for BF and EBF length. The EBF length was uniform among geographic regions and it was directly associated with antenatal care, maternal schooling and wealth index. The EF median was inversely associated with cesarean delivery in private hospital, birth interval less than 24 months and it was different among regions (higher in North). Four DP were identified for children age 6 to 59 months: PA1(composed of yogurt, red meat, chicken, eggs, tubers, vegetables and fruits), PA2(composed of liquids, non-breast milk, red meat, chicken, eggs and negative loadings for breast milk and enriched starch porridge), PA3(composed of liquids, fruit juices, industrialized starch porridge, yogurt and negative factor loadings for red meat) and PA4( composed of non-breast milk, formula and negative factor loadings for enriched starch porridge and egg/fish/chicken). The PA1 and PA3 practices were higher among breastfed children. The PA1 dietary pattern was uniform among geographic regions otherwise, another patterns were differently distribute. The most important DP change was the virtual replacement PA4 for PA3, that was more appropriated for children aged higher than 12 months. CONCLUSION: The pro-breastfeeding scenarium is focused mainly in children universally exposed to BF and the upkeeping of BF, assessed by BF median duration. However, in this survey periods, the increase of EBF duration is lower than the BF duration. The increased solids exposure in children under 12 months expresses the main dietary pattern change between 1996 and 2006. Three dietary patterns of under five children are predicted by geographic region factors and one of them by socioeconomic factors.

O papel da metanálise de sumarizar estudos publicados de mesmo objetivo, por meio da estatística, torna-se cada dia mais fundamental em razão do avanço da ciência e do desejo de usar o menor número de seres humanos em ensaios clínicos, desnecessários, em vários casos. A síntese das informações disponíveis facilita o entendimento e possibilita conclusões robustas. O aumento de estudos clínicos, por exemplo, promove um crescimento da necessidade de metanálises, fazendo com que seja necessário o desenvolvimento de técnicas sofisticadas. Desse modo, o objetivo deste trabalho foi propor uma metodologia bayesiana para a realização de metanálises. O procedimento proposto consiste na mistura das distribuições a posteriori do parâmetro de interesse de cada estudo pertencente à metanálise; ou seja, a medida metanalítica proposta foi uma distribuição de probabilidade e não uma simples medida-resumo. A metodologia apresentada pode ser utilizada com qualquer distribuição a priori e qualquer função de verossimilhança. O cálculo da medida metanalítica pode ser utilizado, desde problemas simples até os mais sofisticados. Neste trabalho, foram apresentados exemplos envolvendo diferentes distribuições de probabilidade e dados de sobrevivência. Em casos, em que se há uma estatística suficiente disponível para o parâmetro em questão, a distribuição de probabilidade a posteriori depende dos dados apenas por meio dessa estatística e, assim, em muitos casos, há a redução de dimensão sem perda de informação. Para alguns cálculos, utilizou-se o método de simulação de Metropolis-Hastings. O software estatístico utilizado neste trabalho foi o R. / The meta-analysis role of using Statistics to summarize published studies that have the same goal becomes more essential day by day, due to the improvement of Science and the desire of using the least possible number of human beings in clinical trials, which in many cases is unnecessary. By match the available information it makes the understanding easier and it leads to more robust conclusions. For instance, the increase in the number of clinical researches also makes the need for meta-analysis go higher, arising the need for developing sophisticated techniques. Then our goal in this work is to propose a Bayesian methodology to conduct meta-analysis. The proposed procedure is a blend of posterior distributions from interest parameters of each work we are considering when doing meta-analysis. As a consequence, we have a probability distribution as a meta-analytic measure, rather than just a statistical summary. The methodology we are presenting can be used with any prior probability distribution and any likelihood function. The calculation of the meta-analytic measure has its uses from small to more complex problems. In this work we present some examples that consider various probability distributions and also survival data. There is a sufficient statistic available for the parameter of interest, the posterior probability distribution depends on the data only through this statistic and thus, in many cases, we can reduce our data without loss of information. Some calculations were performed through Metropolis-Hastings simulation algorithm. The statistical software used in this work was the R.

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.

O estudo teve como objetivo analisar os determinantes sociais da saúde associados à mortalidade por TB e verificar as mortes precoces ocorridas por TB e TB/HIV e seus fatores associados, por meio de duas abordagens: uma de base ecológica e uma de base individual, utilizando análise espacial e de sobrevida. O estudo foi realizado na capital do Paraná, Curitiba; e, para análise espacial, foram consideradas as 148 unidades de desenvolvimento humano (UDH). A população de estudo foi composta dos casos de mortes por TB como causa básica (CID 15-19). Para análise de sobrevida, foram acrescidos os casos de mortes pela coinfecção TB/HIV (CID 20.0). Os dados foram obtidos do Sistema de Informação sobre Mortalidade (SIM) e do Sistema de Informações sobre Doenças de Notificação (SINAN) referentes ao período 2008 a 2015. As Unidades de Desenvolvimento Humano foram caracterizadas de acordo com a mortalidade por TB e com as variáveis dos determinantes sociais da saúde. Inicialmente os casos de óbitos por TB foram geocodificados e calculadas taxas de mortalidade bruta, taxa Bayesiana e investigados quanto à autocorrelação espacial e existência de aglomerados de risco por meio da técnica de varredura espacial e obtidos riscos relativos espaciais. Para correlacionar as áreas de risco espacial para mortalidade por TB foi utilizada a regressão logística, tendo como variável dependente área de risco: sim e não, e após avaliado com uso da curva ROC, também foi elaborado um mapa de sobreposição de áreas de risco dos determinantes sociais da saúde e correlacionados com aglomerados de risco para mortalidade. Por último, para investigar as mortes precoces por TB e TB/HIV e fatores associados foi utilizada a técnica de Kaplan-Meier e Regressão de Cox. Foram identificados 131 óbitos por TB, dos quais 126 (96,2%) foram geocodificados e 05 (4,8%) foram excluídos devido a endereços incompletos. Para a primeira fase, foram calculadas as taxas resultado em taxa média bruta de 1,07/100.000 habitantes. As mortes estiveram distribuídas de maneira difusa, porém, com maior concentração nas regiões periféricas e sul do município. Foi detectado um aglomerado espacial de risco na região sul para mortalidade por TB e para variáveis dos determinantes sociais da saúde, sendo onde as piores condições foram detectadas. O estudo confirmou a relação entre os determinantes sociais e as áreas de risco de mortes por TB quando relacionados com a Dimensão 1 extraída com (OR= 0,093; IC95% 0,34-0,25). O mapa de sobreposição dos aglomerados de risco relacionados com aglomerado para mortalidade por TB resultou em um OR= 5,98 (IC95%: 2,41-14,49) e curva ROC= 0,865; IC95%=0,796-0,934. Na segunda fase, ao analisar as mortes precoces por TB, foi encontrada uma mediana de dias sobrevividos de 22 dias, sendo que 88 (59,1%) dos pacientes morreram até 30 dias após o diagnóstico e 107 (72,5%) após 60 dias (mínimo: 1, máximo: 349, D.P: 68,8 e média: 50 dias). Dentre os 179 óbitos analisados, 105 (58,6%) óbitos tinham diagnóstico de TB (CID 15.0-19) e 74 (41,4%) óbitos a coinfecção TB/HIV (CID 20.0). A maioria dos casos ocorreu em pessoas do sexo masculino, 138 (77,1%), da raça/ cor branca predominante 120 (67%) e a média de idade foi de 47 anos (mínimo:20, máximo: 94, mediana: 44, DP: 14). Os resultados corroboram com a necessidade de melhorias múltiplas nas condições de vida da população, com enfoque nas regiões mais vulneráveis (áreas de aglomerados de risco espacial) identificadas e políticas específicas para prevenção do uso de álcool, diante da identificação deste fator associado às mortes precoces / The objective of the study was to analyze the social determinants of health associated with TB mortality and to verify the early deaths caused by TB and TB / HIV and their associated factors, through two approaches: one based on an ecological basis and an individual basis using analysis spatial and survival The study was carried out in the capital of Paraná, Curitiba; and for spatial analysis, the 148 human development units (UDH) were considered. The study population was composed of cases of TB deaths as the underlying cause (ICD 15-19). Survival analysis included cases of TB / HIV coinfection deaths (ICD 20.0). Data were obtained from the Mortality Information System (SIM) and the Notification Disease Information System (SINAN) for the period 2008 to 2015. The Human Development Units were characterized according to TB mortality and the variables determinants of health. Initially the cases of TB deaths were geocoded and gross mortality rates, Bayesian taxa were calculated and investigated for spatial autocorrelation and existence of clusters of risk by means of the spatial scanning technique and obtained relative spatial risks. In order to correlate spatial risk areas with mortality from TB, logistic regression was used as a risk variable: yes and no and after being evaluated using the ROC curve, a map of overlapping risk areas of social determinants correlated with clusters of risk for mortality. A total of 131 TB deaths were identified, 126 (96.2%) of which were geocoded and 05 (4%) of the deaths were TB and HIV and associated factors were Kaplan-Meier and Cox Regression. A total of 131 TB deaths were identified, of which 126 (96.2%) were geocoded and 05 (4.8%) were excluded due to incomplete addresses. For the first phase, the results were calculated at a gross average rate of 1.07 / 100,000 inhabitants. The deaths were distributed in a diffuse way, however, with greater concentration in the peripheral and southern regions of the municipality. It was detected a spatial cluster of risk in the southern region for mortality by TB and for variables of the social determinants of health and where the worst conditions were detected. The study confirmed the relationship between social determinants and risk areas for TB deaths when related to Dimension 1 extracted with (OR= 0,093; IC95% 0,34-0,25). The map of overlapping cluster-related risk clusters for TB mortality resulted in an OR= 5.95 IC95%=2.41-14.49 and ROC curve= 0.865 (CI95%= 0.796-0.934). In the second phase, when analyzing the early TB deaths, a median number of surviving days of 22 days was found, of which 88 (59.1%) died within 30 days after diagnosis and 107 (72.5%) after 60 days days (minimum: 1, maximum: 349, SD: 68.8 and average: 50 days). Among the 179 deaths analyzed, 105 (58.6%) deaths had a diagnosis of TB (ICD 15.0-19) and 74 (41.4%) had TB / HIV co-infection (ICD 20.0). The majority of the cases occurred in males, 138 (77.1%), the predominant white race (67%) and the mean age was 47 years (minimum: 20, maximum: 94, median: 44, DP: 14). The results corroborate the need for multiple improvements in the living conditions of the population, with a focus on the most vulnerable regions (areas of agglomerates of spatial risk) identified and specific policies to prevent alcohol use, in view of the identification of this factor associated with early deaths

Nuclear magnetic resonance (NMR) is a highly versatile analytical technique for studying molecular configuration, conformation, and dynamics, especially of biomacromolecules such as proteins. However, due to the intrinsic properties of NMR experiments, results from the NMR instruments require a refencing step before the down-the-line analysis. Poor chemical shift referencing, especially for 13C in protein Nuclear Magnetic Resonance (NMR) experiments, fundamentally limits and even prevents effective study of biomacromolecules via NMR. There is no available method that can rereference carbon chemical shifts from protein NMR without secondary experimental information such as structure or resonance assignment. To solve this problem, we constructed a Bayesian probabilistic framework that circumvents the limitations of previous reference correction methods that required protein resonance assignment and/or three-dimensional protein structure. Our algorithm named Bayesian Model Optimized Reference Correction (BaMORC) can detect and correct 13C chemical shift referencing errors before the protein resonance assignment step of analysis and without a three-dimensional structure. By combining the BaMORC methodology with a new intra-peaklist grouping algorithm, we created a combined method called Unassigned BaMORC that utilizes only unassigned experimental peak lists and the amino acid sequence. Unassigned BaMORC kept all experimental three-dimensional HN(CO)CACB-type peak lists tested within ± 0.4 ppm of the correct 13C reference value. On a much larger unassigned chemical shift test set, the base method kept 13C chemical shift referencing errors to within ± 0.45 ppm at a 90% confidence interval. With chemical shift assignments, Assigned BaMORC can detect and correct 13C chemical shift referencing errors to within ± 0.22 at a 90% confidence interval. Therefore, Unassigned BaMORC can correct 13C chemical shift referencing errors when it will have the most impact, right before protein resonance assignment and other downstream analyses are started. After assignment, chemical shift reference correction can be further refined with Assigned BaMORC. To further support a broader usage of these new methods, we also created a software package with web-based interface for the NMR community. This software will allow non-NMR experts to detect and correct 13C referencing errors at critical early data analysis steps, lowering the bar of NMR expertise required for effective protein NMR analysis.

The aim of the present study is to develop a statistical algorithm and model associ- ated with breast and lung cancer patients. In this study, we developed several statistical softwares, R packages, and models using our new statistical approach. In the present study, we used the five parameters logistic model for determining the optimal doses of a pharmaceutical drugs, including dynamic initial points, an automatic process for outlier detection and an algorithm that develops a graphic user interface(GUI) program. The developed statistical procedure assists medical scientists by reducing their time in determining the optimal dose of new drugs, and can also easily identify which drugs need more experimentation. Secondly, in the present study, we developed a new classification method that is very useful in the health sciences. We used a new decision tree algorithm and a random forest method to rank our variables and to build a final decision tree model. The decision tree can identify and communicate complex data systems to scientists with minimal knowledge in statistics. Thirdly, we developed statistical packages using the Johnson SB probability distribu- tion which is important in parametrically studying a variety of health, environmental, and engineering problems. Scientists are experiencing difficulties in obtaining estimates for the four parameters of the subject probability distribution. The developed algorithm com- bines several statistical procedures, such as, the Newtwon Raphson, the Bisection, the Least Square Estimation, and the regression method to develop our R package. This R package has functions that generate random numbers, calculate probabilities, inverse probabilities, and estimate the four parameters of the SB Johnson probability distribution. Researchers can use the developed R package to build their own statistical models or perform desirable statistical simulations. The final aspect of the study involves building a statistical model for lung cancer sur- vival time. In developing the subject statistical model, we have taken into consideration the number of cigarettes the patient smoked per day, duration of smoking, and the age at diagnosis of lung cancer. The response variables the survival time. The significant factors include interaction. the probability density function of the survival times has been obtained and the survival function is determined. The analysis is have on your groups the involve gender and with factors. A companies with the ordinary survival function is given.

In the development process of novel cancer drugs, one important aspect is to identify patient populations with a high risk of early death so that resources can be focused on patients with the highest medical unmet need. Many cancer types are heterogeneous and there is a need to identify patients with aggressive diseases, meaning a high risk of early death, compared to patients with indolent diseases, meaning a low risk of early death. Predictive modeling can be a useful tool for risk stratification in clinical practice, enabling healthcare providers to treat high-risk patients early and progressively, while applying a less aggressive watch-and-wait strategy for patients with a lower risk of death. This is important from a clinical perspective, but also a health economic perspective since society has limited resources, and costly drugs should be given to patients that can benefit the most from a specific treatment. Thus, the goal of predictive modeling is to ensure that the right patient will have access to the right drug at the right time. In the era of personalized medicine, Artificial Intelligence (AI) applied to high-quality data will most likely play an important role and many techniques have been developed. In particular, Graph Neural Network (GNN) is a promising tool since it captures the complexity of high dimensional data modeled as a graph. In this work, we have applied Network Representation Learning (NRL) techniques to predict survival, using pseudonymized patient-level data from national health registries in Sweden. Over the last decade, more health data of increased complexity has become available for research, and therefore precision medicine could take advantage of this trend by bringing better healthcare to the patients. However, it is important to develop reliable prediction models that not only show high performances but take into consideration privacy, avoiding any leakage of personal information. The present study contributes novel insights related to GNN performance in different survival prediction tasks, using population-based unique nationwide data. Furthermore, we also explored how privacy methods impact the performance of the models when applied to the same dataset. We conducted a set of experiments across 6 dataset using 8 models measuring both AUC, Precision and Recall. Our evaluation results show that Graph Neural Networks were able to reach accuracy performance close to the models used in clinical practice and constantly outperformed, by at least 4.5%, the traditional machine learning methods. Furthermore, the study demonstrated how graph modeling, when applied based on knowledge from clinical experts, performed well and showed high resiliency to the noise introduced for privacy preservation. / I utvecklingsprocessen för nya cancerläkemedel är en viktig aspekt att identifiera patientgrupper med hög risk för tidig död, så att resurser kan fokuseras på patientgrupper med störst medicinskt behov. Många cancertyper är heterogena och det finns ett behov av att identifiera patienter med aggressiv sjukdom, vilket innebär en hög risk för tidig död, jämfört med patienter med indolenta sjukdom, vilket innebär lägre risk för tidig död. Prediktiv modellering kan vara ett användbart verktyg för riskstratifiering i klinisk praxis, vilket gör det möjligt för vårdgivare att behandla patienter olika utifrån individuella behov. Detta är viktigt ur ett kliniskt perspektiv, men också ur ett hälsoekonomiskt perspektiv eftersom samhället har begränsade resurser och kostsamma läkemedel bör ges till de patienter som har störst nytta av en viss behandling. Målet med prediktiv modellering är således att möjliggöra att rätt patient får tillgång till rätt läkemedel vid rätt tidpunkt. Framför allt är Graph Neural Network (GNN) ett lovande verktyg eftersom det fångar komplexiteten hos högdimensionella data som modelleras som ett diagram. I detta arbete har vi tillämpat tekniker för inlärning av grafrepresentationer för att prediktera överlevnad med hjälp av pseudonymiserade data från nationella hälsoregister i Sverige. Under det senaste decennierna har mer hälsodata av ökad komplexitet blivit tillgänglig för forskning. Även om denna ökning kan bidra till utvecklingen av precisionsmedicinen är det viktigt att utveckla tillförlitliga prediktionsmodeller som tar hänsyn till patienters integritet och datasäkerhet. Den här studien kommer att bidra med nya insikter om GNNs prestanda i prediktiva överlevnadsmodeller, med hjälp av populations -baserade data. Dessutom har vi också undersökt hur integritetsmetoder påverkar modellernas prestanda när de tillämpas på samma dataset. Sammanfattningsvis, Graph Neural Network kan uppnå noggrannhets -prestanda som ligger nära de modeller som tidigare använts i klinisk praxis och i denna studie preserade de alltid bättre än traditionella maskininlärnings -metoder. Studien visisade vidare hur grafmodellering som utförs i samarbete med kliniska experter kan vara effektiva mot det brus som införs av olika integritetsskyddstekniker.

<p>A general time-to-event data analysis known as event history analysis is considered. The focus is on the analysis of time-to-event data using Cox's regression model when the time to the event may be measured from different origins giving several observable time scales and when longitudinal measurements are involved. For the multiple time scales problem, procedures to choose a basic time scale in Cox's regression model are proposed. The connections between piecewise constant hazards, time-dependent covariates and time-dependent strata in the dual time scales are discussed. For the longitudinal measurements problem, four methods known in the literature together with two proposed methods are compared. All quantitative comparisons are performed by means of simulations. Applications to the analysis of infant mortality, morbidity, and growth are provided.</p>

A frequent problem that appears in practical survival data analysis is censoring. A censored observation occurs when the observation of the event time (duration or survival time) may be prevented by the occurrence of an earlier competing event (censoring time). Censoring may be due to different causes. For example, the loss of some subjects under study, the end of the follow-up period, drop out or the termination of the study and the limitation in the sensitivity of a measurement instrument. The literature about censored data focuses on the i.i.d. case. However in many real applications the data are collected sequentially in time or space and so the assumption of independence in such case does not hold. Here we only give some typical examples from the literature involving correlated data which are subject to censoring. In the clinical trials domain it frequently happens that the patients from the same hospital have correlated survival times due to unmeasured variables like the quality of the hospital equipment. Censored correlated data are also a common problem in the domain of environmental and spatial (geographical or ecological) statistics. In fact, due to the process being used in the data sampling procedure, e.g. the analytical equipment, only the measurements which exceed some thresholds, for example the method detection limits or the instrumental detection limits, can be included in the data analysis. Many other examples can also be found in other fields like econometrics and financial statistics. Observations on duration of unemployment e.g., may be right censored and are typically correlated. When the data are not independent and are subject to censoring, estimation and inference become more challenging mathematical problems with a wide area of applications. In this context, we propose here some new and flexible tools based on a nonparametric approach. More precisely, allowing dependence between individuals, our main contribution to this domain concerns the following aspects. First, we are interested in developing more suitable confidence intervals for a general class of functionals of a survival distribution via the empirical likelihood method. Secondly, we study the problem of conditional mean estimation using the local linear technique. Thirdly, we develop and study a new estimator of the conditional quantile function also based on the local linear method. In this dissertation, for each proposed method, asymptotic results like consistency and asymptotic normality are derived and the finite sample performance is evaluated in a simulation study.

A frequent problem that appears in practical survival data analysis is censoring. A censored observation occurs when the observation of the event time (duration or survival time) may be prevented by the occurrence of an earlier competing event (censoring time). Censoring may be due to different causes. For example, the loss of some subjects under study, the end of the follow-up period, drop out or the termination of the study and the limitation in the sensitivity of a measurement instrument. The literature about censored data focuses on the i.i.d. case. However in many real applications the data are collected sequentially in time or space and so the assumption of independence in such case does not hold. Here we only give some typical examples from the literature involving correlated data which are subject to censoring. In the clinical trials domain it frequently happens that the patients from the same hospital have correlated survival times due to unmeasured variables like the quality of the hospital equipment. Censored correlated data are also a common problem in the domain of environmental and spatial (geographical or ecological) statistics. In fact, due to the process being used in the data sampling procedure, e.g. the analytical equipment, only the measurements which exceed some thresholds, for example the method detection limits or the instrumental detection limits, can be included in the data analysis. Many other examples can also be found in other fields like econometrics and financial statistics. Observations on duration of unemployment e.g., may be right censored and are typically correlated. When the data are not independent and are subject to censoring, estimation and inference become more challenging mathematical problems with a wide area of applications. In this context, we propose here some new and flexible tools based on a nonparametric approach. More precisely, allowing dependence between individuals, our main contribution to this domain concerns the following aspects. First, we are interested in developing more suitable confidence intervals for a general class of functionals of a survival distribution via the empirical likelihood method. Secondly, we study the problem of conditional mean estimation using the local linear technique. Thirdly, we develop and study a new estimator of the conditional quantile function also based on the local linear method. In this dissertation, for each proposed method, asymptotic results like consistency and asymptotic normality are derived and the finite sample performance is evaluated in a simulation study.