La prédiction d'un événement clinique à l'aide d'outils pronostiques est une question centrale en oncologie. L'émergence des biomarqueurs mesurés au cours du temps permet de proposer des outils incorporant les données répétées de ces biomarqueurs pour mieux guider le clinicien dans la prise en charge des patients. L'objectif de ce travail est de développer et valider des outils pronostiques dynamiques de rechute de cancer de la prostate, chez des patients traités initialement par radiothérapie externe, en prenant en compte les données répétées du PSA, l'antigène spécifique de la prostate, en plus des facteurs pronostiques standard. Ces outils sont dynamiques car ils peuvent être mis à jour à chaque nouvelle mesure disponible du biomarqueur. Ils sont construits à partir de modèles conjoints pour données longitudinales et de temps d'événement. Le principe de la modélisation conjointe est de décrire l'évolution du biomarqueur à travers un modèle linéaire mixte, décrire le risque d'événement à travers un modèle de survie et lier ces deux processus à travers une structure latente. Deux approches existent, les modèles conjoints à effets aléatoires partagés et les modèles conjoints à classes latentes. Dans un premier travail, nous avons tout d'abord comparé, en terme de qualité d'ajustement et de pouvoir prédictif, des modèles conjoints à effets aléatoires partagés différant par leur forme de dépendance entre le PSA et le risque de rechute clinique. Puis nous avons évalué et comparé ces deux approches de modélisation conjointe. Dans un deuxième travail, nous avons proposé un outil pronostique dynamique différentiel permettant d'évaluer le risque de rechute clinique suivant l'initiation ou non d'un second traitement (un traitement hormonal) au cours du suivi. Dans ces travaux, la validation de l'outil pronostique a été basée sur deux mesures de pouvoir prédictif: le score de Brier et l'entropie croisée pronostique. Dans un troisième travail, nous avons enfin décrit la dynamique des PSA après un second traitement de type hormonal chez des patients traités initialement par une radiothérapie seule. / The prediction of a clinical event with prognostic tools is a central issue in oncology. The emergence of biomarkers measured over time can provide tools incorporating repeated data of these biomarkers to better guide the clinician in the management of patients. The objective of this work is to develop and validate dynamic prognostic tools of recurrence of prostate cancer in patients initially treated by external beam radiation therapy, taking into account the repeated data of PSA, the Prostate-Specific Antigen, in addition to standard prognostic factors. These tools are dynamic because they can be updated at each available new measurement of the biomarker. They are built from joint models for longitudinal and time-to-event data. The principle of joint modelling is to describe the evolution of the biomarker through a linear mixed model, describe the risk of event through a survival model and link these two processes through a latent structure. Two approaches exist, shared random-effect models and joint latent class models. In a first study, we first compared in terms of goodness-of-fit and predictive accuracy shared random-effect models differing in the form of dependency between the PSA and the risk of clinical recurrence. Then we have evaluated and compared these two approaches of joint modelling. In a second study, we proposed a differential dynamic prognostic tool to evaluate the risk of clinical recurrence according to the initiation or not of a second treatment (an hormonal treatment) during the follow-up. In these works, validation of the prognostic tool was based on two measures of predictive accuracy: the Brier score and the prognostic cross-entropy. In a third study, we have described the PSA dynamics after a second treatment (hormonal) in patients initially treated by a radiation therapy alone.

信用評等機構對於企業與市場投資者而言有著重要且無可取代的功能，其所提供之信用評等資訊應當是許多市場投資者所仰賴的重要決策依據，但近年來，卻有許多外界聲浪質疑信評機構評等之準確性，本論文之研究目的即是希望從錨定效應此一行為偏誤之觀點切入，探討國內信評機構在對企業評等時，是否會受到錨定效應影響，導致評等調整不正確或是評等落後其他財務指標等現象發生。 研究結果顯示，國內信評機構對受評企業之過往財務資訊存在錨定現象，但不至於大幅影響整體違約風險之準確性，且信評機構對受評企業財務之惡化較為敏感。另一方面，本研究也發現，信評機構對於非上市櫃公司、金融業以及初次評等等級在「twAA」以上企業之評等，存在較為明顯之錨定現象。 / Credit Rating Agencies (CRAs) play a major role in the financial market. Credit rating purport to provide investors with valuable information they need to make decisions about investing, but the accuracy of the rating itself has been called into question by many investors in recent years. The purpose of this study is to examine the anchoring effect on CRAs while the rating is being given. The results indicate that domestic CRAs tend to be anchored on the past financial information of the issuers. But the impacts are very slightly. Besides this, CRAs seem relatively sensitive to the financial deterioration. Moreover, the anchoring effect are much more significant when the debt issuers are private firms, financial institutions or the companies with greater or equal to twAA initial credit rating.

[pt] Os modelos GAS (generalized autoregressive score) são modelos de séries temporais com parâmetros variantes no tempo, os quais possuem sua evolução ditada pelo vetor score ponderado da função de verossimilhança. A avaliação da verossimilhança nestes modelos é bastante simples, bem como incorporação de efeitos como assimetria, memória longa e outras dinâmicas. Por serem baseados no score da verossimilhança, exporta-se a estrutura completa da distribuição preditiva para o mecanismo de atualização dos parâmetros, e não apenas a média ou momentos de ordem superior. Estas características, somadas á capacidade de lidar com processos multivariados e não estacionários, tornam a classe em estudo uma nova alternativa para a construção de modelos com parâmetros variantes, particularmente para séries temporais não gaussianas. Nesta dissertação, foram desenvolvidos modelos GAS univariados para a análise das séries mensais de vazão do Rio Paraibuna (MG) e de fator de capacidade de uma usina é olica não divulgada do Nordeste, utilizando as distribuições gama e beta, respectivamente. Além disso, foi derivado um novo modelo GAS bivariado com marginais gama e beta para a modelagem conjunta dos processos de vazão e vento, de modo a explorar a complementaridade sazonal entre as séries. / [en] The GAS models (generalized autoregressive score) are time series models with time-varying parameters, which have their update mechanism drived by the scaled score of the likelihood function. The likelihood evaluation in these models is quite simple, as well as the incorporation of effects like asymmetry, long memory and other dynamics. Because they are based in the scaled score of the likelihood, it exploits the full structure of the predictive distribution to the update mechanism of the parameters, and not just mean or higher order moments. These characteristics, coupled with the ability to handle with multivariate and non-stationary processes, make the studied class a new alternative to the construction of models with time-varying parameters, particularly for non-Gaussian time series. In this dissertation, univariate GAS models were developed to analyze monthly series of streamflow of Paraibuna river (MG) and of capacity factor of a wind farm undisclosed in Northeast, using the gamma and beta distributions, respectively. In addition, a new bivariate GAS model with gamma and beta marginals was derived for the joint modeling of the streamflow and wind processes, in order to explore the seasonal complementarity between the series.

O cerne dos programas de auxílio-alimentação no Brasil foi a necessidade de combater a desnutrição que acometia uma considerável parte da população brasileira. A desnutrição foi superada no país, que atualmente enfrenta um outro problema relacionado à alimentação, que é o crescente número de pessoas com sobrepeso e obesidade. As transformações nos padrões nutricionais e alimentares, influenciadas pelas mudanças sociais, econômicas e demográficas que ocorreram ao longo das últimas décadas, contribuíram com o aumento do número de casos de pessoas com excesso de peso e de doenças associadas. Trata-se de problemas de saúde pública que geram perdas econômicas e de bem-estar. Diante desse cenário e da escassa literatura que avaliou os efeitos dos programas de auxílio-alimentação ao longo dos anos, esta pesquisa buscou analisar os impactos desses programas nos índices antropométricos, especificamente no excesso de peso e na obesidade, e também no consumo de grupos de alimentos e componentes alimentares dos trabalhadores beneficiados. Para tanto, utilizou-se os dados da Pesquisa de Orçamentos Familiares 2008/2009 e o método Propensity Score Matching. Os resultados apontaram que receber auxílio-alimentação aumenta as chances de obesidade entre as mulheres e de sobrepeso entre os homens. A análise por estratos de renda evidenciou que as chances de excesso de peso e obesidade são maiores entre os homens mais pobres, do primeiro estrato, e de obesidade para as mulheres do segundo estrato. Em relação aos impactos nos grupos de alimentos, foi observado um aumento no consumo de alimentos pertencentes ao grupo de Frutas; Farinhas e massas, Bebidas e Preparações Mistas e uma redução no consumo de Grãos e Legumes, entre as mulheres beneficiárias. Já para os homens que recebem o benefício, foi observado aumento no consumo de alimentos dos grupos das Farinhas e massas, Óleos e gorduras, Bebidas, Pizzas e salgados e de Oleaginosas, em comparação aos não beneficiários. Os resultados referentes à ingestão de nutrientes revelaram que as beneficiárias consumiram mais: Energia, Proteína; Carboidrato, Colesterol; Cálcio; Magnésio, Fósforo; Ferro; Selênio; Retinol; Vitaminas: A, B1, B2, B3, Equivalente a B3, B6; B12 e Folato; Açúcar total e de adição. Já os homens beneficiários ingeriram maiores quantidades de: Energia; Carboidrato; Magnésio, Selênio, Vitaminas: B1, B3, B6, E e Folato; Ácidos graxos: poliinsaturados, poliinsaturado 18:2 e trans total; e Açúcar total e de adição. Conclui-se, portanto, que os programas de alimentação dos trabalhadores estão contribuindo, muitas vezes, com a piora da saúde dos assistidos, em especial os mais pobres. Em partes, os programas de auxílio-alimentação reduzem a qualidade alimentar dos beneficiários e beneficiárias, pois eles consomem maior quantidade de alimentos e de componentes alimentares nocivos ao equilíbrio nutricional, ao mesmo tempo que também ingerem mais nutrientes benéficos ao correto funcionamento do organismo. / Food assistance programs in Brazil used to have as its main goal the control of malnutrition that affected large part of the population. However, malnutrition is today much reduced in Brazil, and the country is facing another problem - the increasing of overweight and obese people. Changes in nutritional and food preferences, affected by social, economic and demographic evolution in the last decades, have been collaborating to increase the number of overweight people and related diseases in the population. These problems cause economic and welfare losses. Nevertheless, the literature still lacks deep studies in this topic. In this context, this research aims to analyse the impact of a food assistance program for employees in weight and obesity indices as well as in the consumption of food and nutrients by Brazilian workers. For that, we use the database from Brazilian Expenditure Survey 2008/09 and apply the Propensity Score Matching method. Our results show that be a participant of food assistance program increases the probability of obesity between women and of overweight among men. Our results point out that the probability to be overweight is greater for poor men and the probability to be obese is greater for both poor men and women. Regarding the impact on food consumption, the results evidence, for women, an increase in the consumption of fruits, flours and pastas, drinks and regional preparations and a reduction of grains and vegetables. For men, the results point out an increase in the consumption of flours and pastas; fat food, drinks; pizzas and oilseeds. About nutrients, women participants raise their consumption of energy, carbohydrate, proteins, cholesterol, calcium, phosphor, iron, selenium, magnesium, retinol, total and additional sugar, folate and vitamins A, B1, B2, B3, and B3 equivalent. On the other hand, men participants increase their consumption of energy, carbohydrate, selenium, magnesium, total and additional sugar, vitamins B1, B3, B6, E and folate; and fatty acids: polyunsaturated, polyunsaturated 18: 2 and trans fats total. In summary, our results support the conclusion that this Brazilian food assistance program has contributed, in many cases, to deteriorate their participants health, specially the poor\'s, and the main channel for that is the reduction of the quality of consumption of food and nutrients by their participants. But the program also contributes to increase the consumption of some important nutrients.

Dans notre société de plus en plus numérique, les systèmes de recommandation ont fait leur apparition dans le but de résoudre le problème bien connu de surcharge d'information. L'adoption des réseaux sociaux a permis l'émergence de systèmes intégrant les relations sociales dans leurs recommandations. Dans cette thèse, nous proposons un système de recommandation adapté aux architectures décentralisées pouvant être déployé sur des réseaux sociaux existants. L'utilisateur conserve son profil en local et ne communique qu'avec un ensemble restreint d'utilisateurs de confiance, avec qui il accepte de partager ses données. Nous prenons en compte le réseau social de l'utilisateur afin de construire le réseau de pairs. La similarité des amis est prise en compte pour pondérer les liens. Les recommandations sont propagées dans le réseau, passant d'amis en amis jusqu'à atteindre l'utilisateur désiré. Ainsi seuls les amis directs communiquent entre eux. À partir de cette propagation, nous proposons plusieurs techniques. Tout d'abord, nous délivrons à l'utilisateur final une confiance du système dans la fiabilité de la recommandation. Ceci lui permet de choisir parmi les produits sélectionnés, lesquels semblent effectivement les plus pertinents pour lui. Cette confiance est calculée sur plusieurs critères, tels que la variation des recommandations des amis, leur nombre, la similarité et la fraîcheur de la recommandation. Ensuite, nous définissons des heuristiques adaptant notre approche aux systèmes pair-à-pair. Dans de telles architectures, le réseau est une ressource critique et ne doit pas être constamment surchargé. Ces heuristiques limitent la consommation réseau de notre approche tout en fournissant des recommandations pertinentes à l'utilisateur. Enfin, nous proposons plusieurs stratégies de score par défaut, dans le cas où aucun score n'est calculable, prenant en compte les contraintes en terme d'accès à l'information par le système. Nous comparons notre approche avec des approches classiques de recommandation, de filtrage collaboratif ou basées sur la confiance, en utilisant plusieurs jeux de données existants, tels qu'Epinions et Flixster, ainsi que deux jeux de données que nous avons construits nous-même. Nous montrons qu'une approche purement locale, associée à des stratégies de score par défaut, offre de meilleurs résultats que la plupart des autres approches, notamment en ce qui concerne les "cold start users". / Recommender systems are widely used to achieve a constantly growing variety of services. Alongside with social networks, recommender systems that take into account friendship or trust between users have emerged. In this thesis, we propose an evolution of trust-based recommender systems adapted to decentralized architectures that can be deployed on top of existing social networks. Users profiles are stored locally and are exchanged with a limited, user-defined, list of trusted users. Our approach takes into account friends' similarity and propagates recommendation to direct friends in the social network in order to prevent ratings from being globally known. Moreover, the computational complexity is reduced since calculations are performed on a limited dataset, restricted to the user's neighborhood. On top of this propagation, our approach investigates several aspects. Our system computes and returns to the final user a confidence on the recommendation. It allows the user to tune his/her choice from the recommended products. Confidence takes into account friends' recommendations variance, their number, similarity and freshness of the recommendations. We also propose several heuristics that take into account peer-to-peer constraints, especially regarding network flooding. We show that those heuristics decrease network resources consumption without sacrificing accuracy and coverage. We propose default scoring strategies that are compatible with our constraints. We have implemented and compared our approach with existing ones, using multiple datasets, such as Epinions and Flixster. We show that local information with default scoring strategies are sufficient to cover more users than classical collaborative filtering and trust-based recommender systems. Regarding accuracy, our approach performs better than others, especially for cold start users, even if using less information.

Introdução: Os escores CHADS2 e CHA2DS2VASc são, comumente, utilizados para a estratificação de risco para eventos tromboembólicos em pacientes com fibrilação atrial. Recentemente, demonstrou-se a utilidade desses escores em pacientes sem fibrilação atrial, particularmente naqueles portadores de doença coronariana, predizendo mortalidade e eventos tromboembólicos nessa população. A adição de alterações ecocardiográficas, por meio das técnicas transtorácica e transesofágica, a tais escores, na presença de fibrilação atrial, pode aumentar a acurácia dessa estratificação, identificando os indivíduos que estão sujeitos a maiores riscos. Objetivos: O objetivo primário desse estudo é demonstrar que indivíduos coronariopatas sem fibrilação atrial apresentam alterações ecocardiográficas que podem justificar maior risco de complicações tromboembólicas e que estas são mais frequentes quanto maiores os escores CHADS2. Como objetivo secundário, demonstrar que essas alterações são mais frequentes quanto maior o escore CHA2DS2VASc. Metodologia: Foram incluídos 111 pacientes com doença coronariana, em ritmo sinusal, os quais foram estratificados pelos escores CHADS2 e CHA2DS2VASc, e divididos em dois grupos de acordo com o risco de eventos tromboembólicos (escores CHADS2 < 2 vs CHADS2 >= 2 e CHA2DS2VASc < 3 vs CHA2DS2VASc >= 3). Após, foram submetidos à ecocardiografia transtorácica e transesofágica, sendo avaliadas: dimensões de átrio esquerdo (diâmetro e volume indexado), massa ventricular esquerda indexada pela superfície corpórea, funções sistólica e diastólica, e a contratilidade miocárdica de ventrículo esquerdo, velocidade de esvaziamento de apêndice atrial esquerdo, presença de contraste espontâneo em átrio esquerdo, presença de trombos em átrio e/ou ventrículo esquerdo. Análise estatística: Dimensões de átrio esquerdo, massa ventricular esquerda, fração de ejeção de ventrículo esquerdo e velocidade de esvaziamento de apêndice atrial esquerdo foram analisadas como variáveis contínuas e, também, como categóricas (normais e anormais). As variáveis contínuas foram comparadas por meio do teste de Mann-Whitney e as categóricas, pelo teste exato de Fisher. Modelos de regressão logística foram aplicados para determinar variáveis de risco independente e verificar quais componentes dos escores CHADS2 e CHA2DSVASc influenciam essas variáveis. Resultados: Da população estudada, 72,1% eram do sexo masculino, idade de 62,4±8,7 anos, 96 (86,5%) eram hipertensos, 44 (39,6%) diabéticos, 17 (15,3%) portadores de insuficiência cardíaca e 12 (10,8%) apresentavam AVC/AIT prévio. A avaliação multivariada demonstrou que pacientes com escore CHADS2 >= 2 (n = 57), mais frequentemente, apresentam maior volume indexado de átrio esquerdo, menor velocidade de esvaziamento de apêndice atrial esquerdo, maior massa ventricular esquerda indexada pela superfície corpórea, menor fração de ejeção de ventrículo esquerdo e, mais frequentemente, apresentavam disfunção diastólica de ventrículo esquerdo. Com relação ao escore CHA2DS2VASc, pacientes com risco elevado apresentam, mais frequentemente, maior volume indexado de átrio esquerdo, menor velocidade de esvaziamento de apêndice atrial esquerdo e maior massa ventricular esquerda indexada pela superfície corpórea. Conclusão: Pacientes com escores CHADS2 e CHA2DS2VASc elevados apresentam alterações ecocardiográficas que podem justificar maior risco para a ocorrência de acidente vascular cerebral em indivíduos com doença arterial coronariana na presença de ritmo sinusal. / CHADS2 and CHA2DS2VASc score are commonly used to stratify risk for thromboembolic events in patients with atrial fibrillation. Recently, it has been demonstrated the usefulness of these scores in patients without atrial fibrillation, particularly in those patients with coronary disease, predicting mortality and thromboembolic events in this population. The addition of echocardiographic changes, through the transthoracic and transesophageal techniques, to those scores, in the presence of atrial fibrillation can increase the accuracy of this stratification, identifying individuals who are subject to greater risks. Objectives: The primary objective of this study is to demonstrate that coronary disease patients without atrial fibrillation have echocardiographic changes that may justify increased risk of thromboembolic complications and that these are more frequent the higher the CHADS2 scores. As a secondary objective, demonstrate that those changes are more frequent the higher the score CHA2DS2VASc. Methods: 111 patients with coronary artery disease in sinus rhythm were included, which were stratified by CHADS2 and CHA2DS2VASc scores and divided into two groups according to the risk of thromboembolic events (CHADS2 < 2 vs CHADS2 >= 2 and CHA2DS2VASc < 3 vs CHA2DS2VASc >= 3). After that, they were subjected to transthoracic and transesophageal echocardiography, being analyzed: left atrial size (diameter and volume index), left ventricular mass indexed by body surface, systolic and diastolic function and myocardial contractility of left ventricle, left atrial appendage emptying velocity, presence of spontaneous echo contrast in the left atrium, presence of thrombi in the atrium and / or left ventricle. Statistical analysis: left atrial size, left ventricular mass, left ventricular ejection fraction and left atrial appendage emptying velocity were analyzed as continuous variables and as categorical (normal and abnormal). Continuous variables were compared using the Mann-Whitney test, and the categorical, using Fisher\'s exact test. Logistic regression models were applied to determine independent risk variables and determine which components of CHADS2 and CHA2DSVASc scores influence these variables. Results: In the study population, 72.1% were male, age 62.4±8.7 years, 96 (86.5%) were hypertensive, 44 (39.6%) diabetic patients, 17 (15.3 %) heart failure patients and 12 (10.8%) had prior stroke / transient ischemic attack. Multivariate evaluation showed that patients with scores CHADS2 >= 2 (n = 57) most often have greater left atrial volume index, lower left atrial appendage emptying velocity, higher left ventricular mass indexed by body surface area, lower left ventricular ejection fraction and more often had diastolic dysfunction of the left ventricle. Regarding CHA2DS2VASc score, patients at high risk have more often greater left atrial volume index, lower left atrial appendage emptying velocity and greater left ventricular mass indexed by body surface area. Conclusion: Patients with high CHADS2 and CHA2DS2VASc scores present echocardiographic changes that could justify a higher risk for the occurrence of stroke in patients with coronary artery disease in the presence of sinus rhythm.

Sabe-se que um bom fluxo sanguíneo colateral (FSC) leptomeníngeo está associado a melhor desfecho funcional em paciente com oclusão vascular da circulação anterior. Entretanto, poucos estudos avaliaram os preditores de FSC e seu impacto prognósticos em países em desenvolvimento. O objetivo deste trabalho é avaliar os preditores e o impacto prognóstico do FSC leptomeníngeo nas oclusões do primeiro segmento (M1) da Artéria Cerebral Média (ACM). Métodos: Foi avaliado de maneira retrospectiva, uma coorte prospectiva de pacientes com acidente vascular cerebral (AVC), admitidos em um centro terciário acadêmico de referência para o tratamento de AVC entre janeiro de 2014 e junho de 2017. Foram incluídos pacientes que apresentavam oclusão do segmento M1 da ACM e realizaram angiotomografia de crânio (AngioTC) até 12 horas após o ictus. O FSC foi analisado utilizando uma escala previamente validada na literatura, o rLMC score. Foi avaliado a relação do rLMC score com variáveis clínicas e com desfecho. Na análise univariada foi utilizado o teste de Fisher para variáveis categóricas e t de Student para variáveis quantitativas, também foi utilizado o teste de Spearman para correlação de variáveis contínuas. Características associadas com rLMC score e com desfecho funcional foram incluídas da regressão linear (rLMC não categorizado) e regressão logística binária (rLMC categorizado). Foi utilizado um valor de significância estatística convencional (p=0,05) Resultados: Dos 1559 pacientes com AVC isquêmico admitidos, foram analisados 121 pacientes que preencheram critérios de inclusão/exclusão do estudo. A idade média foi de 67.1±15 anos, NIHSS 13 [IQR 11-22]. Na regressão linear, menor idade (? = -0.2; p=0.005), Doença de14 Chagas (? = 0,22; p=0,008), Tabagismo (? = 0,22; p=0,009), menor NIHSS (? = - 0,25; p=0,002) e história de doença coronariana (? = 0,23; p=0,005) foram associados com melhor rLMC score. Menor idade (? = 0,29; p<0,001), menor NIHSS (? = 0,36; p<0,001) e rLMC>= 17 (? = -0,16; p=0,04) foram associados com melhor desfecho clínico. Conclusão: Pacientes com bom FSC leptomeníngeo apresentam menor NIHSS, menor idade, maior frequência de doença de Chagas, história de tabagismo e doença coronariana prévia. Um bom FSC leptomeníngeo está associado a melhor desfecho funcional entre paciente com AVC isquêmico por oclusão de artérias proximais da circulação anterior no Brasil. / A Good leptomeningeal collateral flow has been associated with better outcomes in patients with proximal occlusions of the anterior circulation. Nevertheless, only few studies have assessed the predictors of good leptomeningeal collateral flow and its impact on clinical outcome in developing countries. We aim to evaluate predictors and prognostic impact of leptomeningeal collateral flow among patients with proximal middle cerebral artery (MCA) occlusion. Method: We retrospectively analyzed a prospective stroke registry of AIS patients admitted to a tertiary Stroke Center in Brazil, with MCA-M1 occlusion within twelve hours of symptoms onset, admitted from January 2014-June 2017. Leptomeningeal score was evaluated using a previously validated Regional Leptomeningeal (rLMC) score. We evaluated association between rLMC score with clinical variables and outcomes. Univariate analysis was performed using Fisher exact test for categorical variables and t-Student for continuous variables. Spearman\'s correlation analysis was performed to find the relationship between continuous variables and rLMC. Variables were included in linear regression model (non categorized rLMC) or binary logistic regression model (categorized rLMC). Conventional levels of statistical significance were used (p=0.05). Results: From 1559 AIS patients, 121 were analyzed. Demographics: age 67.1±15 years, NIHSS 13 [IQR 11-22]. In linear regression, lower age (? = -0.2; p=0.021), Chagas Disease (? = 0.22; p=0.008), smoking (? = 0.22; p=0.009), lower NIHSS (? = -0.25; p=0.002) and history of coronary artery disease (? = 0.23; p=0,005) were associated with better rLMC score. Lower age (? = 0.29; p<0.001), lower NIHSS (? = 0.36; p<0.001) and rLMC>= 17 (? = -0.16; p=0.04) were associated with better functional outcomes. Conclusion: Patients with good leptomeningeal collateral flow had lower age, lower NIHSS, Chagas disease and history of smoking. Good leptomeningeal collaterals on CT angiography are associated to good outcome among AIS patients with proximal arterial occlusion of the anterior circulation in Brazil.

A compreensão da dinâmica de persistência e evasão da atividade exportadora é fundamental para o desenho de incentivos adequados às firmas estreantes no mercado internacional e encontra respaldo nos modelos da nova teoria do comércio internacional. O propósito deste trabalho é investigar os determinantes do desempenho de firmas industriais brasileiras estreantes no mercado internacional, em termos de probabilidade de sobrevivência e evolução do valor exportado, com especial atenção aos impactos dos instrumentos de apoio Drawback, BNDES Exim e Proex. Para tanto, serão analisadas empresas estreantes no comércio exterior entre os anos de 1998 e 2003, configurando um painel desbalanceado com 8,5 mil firmas. Por meio de análise econométrica para dados em painel e estimação de modelos de duração, verificou-se que a função de sobrevivência e crescimento do valor médio exportado no tempo difere claramente entre firmas que utilizam um dos benefícios e aquelas que não o fazem. Também se pode afirmar que custos irreversíveis de entrada no comércio internacional não sejam desprezíveis entre as firmas industriais analisadas, o que indica que os programas públicos de promoção de exportações devam se concentrar na (i) redução da taxa de abandono das recém-exportadoras e (ii) na minimização dos custos fixos associados aos investimentos para entrada no mercado exportador. / Understanding the dynamics of persistence and evasion of export activity is essential for the design of export promotion for new exporters and international trade policies. Several results point to the importance of taking into account the specific sector of the industry in the implementation of public policies, which is supported by the new models of international trade theory. The purpose of this work is to investigate the determinants of the performance of Brazilian industrial new exporters, with particular attention to the impacts of Drawback, BNDES Exim and Proex. For this, we analyzed firms between the 1998 and 2003, constituting an unbalanced panel with 8500 firms. By panel data analysis and estimation of duration models, we found that the survival function and the growth of exports clearly differs among companies that use one of the programs and those that do not. One can also say that sunk costs are not negligible among the industrial firms analyzed, which indicates that public export promotion should focus on (i) reducing the dropout rate of new exporters and (ii) minimize the fixed sunk costs related with initial investments to begin the export activty.

O câncer de próstata (CaP) é a segunda causa mais comum de câncer em homens. De acordo com o INCA, no Brasil, em 2016, estimam-se aproximadamente 61.200 novos casos de câncer de próstata. Objetivo: Identificar as características demográficas e epidemiológicas, bem como dados do estadiamento tumoral dos pacientes com CaP atendidos na Unidade de Oncologia do Hospital Marcio Cunha na cidade de Ipatinga-MG nos anos de 2012, 2013 e 2014. Metodologia: Trata-se de um estudo retrospectivo e descritivo onde foram analisados 668 prontuários de pacientes, com registro do diagnóstico anatomopatológico, atendidos nos anos de 2012, 2013 e 2014, conforme lista fornecida pela instituição, com diagnóstico de CaP cadastrados com CID-10 - C 61. As variáveis analisadas foram: procedência, ano do diagnóstico, faixa etária, raça autodeclarada, fatores de risco como tabagismo, etilismo, história familiar de CaP, PSA total ao diagnóstico, tipo histológico da biópsia, score de Gleason da biópsia, tipo histológico da peça cirúrgica, score de Gleason da peça cirúrgica. Os dados foram analisados empregando-se estatística descritiva e inferencial, utilizando o software SPSS, versão 19.0. Resultados: A maior incidência de casos de CaP foram provenientes das cidades mais populosas da microrregião de saúde analisada e faixa etária mais prevalente foi entre 61 e 80 anos com prevalência em pardos e brancos e com histórico familiar de 17,2% de parentes de primeiro grau; com o pai em 37,3%, o irmão em 60,8% e filho em 1,9%. Apenas 165 (25,9 %) eram fumantes e 20,8% etilistas. Os níveis de PSA ficaram entre 4,1ng/ e 10ng/ml (49,5%) e quanto maior a faixa etária maiores os valores do PSA. Pacientes pardos apresentaram PSA total mais elevado. Ao avaliarmos se existia relação entre os níveis de PSA total com fatores de risco como tabagismo, etilismo e histórico familiar, somente houve relação estatisticamente significativa com o etilismo. Houve concordância do score de Gleason entre biópsia e peça cirúrgica em 70%, subgraduação em 18,7% e supergraduação em 11,3%. Comparando a idade dos pacientes com Score de Gleason, quanto maior a idade do paciente maior foi o Score de Gleason do material obtido pela biópsia via transretal Pacientes tabagistas e etilistas apresentaram Score de Gleason da peça cirúrgica mais elevados. Conclusão: A concordância entre o Score de Gleason da biópsia e o Score de Gleason da peça cirúrgica foi de 70%; etilistas apresentaram PSA mais elevados; quanto maior foi a faixa etária, mais indiferenciado foi o tumor ( biópsia); pacientes tabagistas e etilistas apresentaram tumores mais indiferenciados na peça cirúrgica; este é o primeiro estudo epidemiológico de CaP desenvolvido na região do Vale do Aço, a caracterização sócio demográfica e as associações aqui encontradas podem contribuir com programas para desenvolver ações de controle do CaP nesta região. / Prostate cancer (PCa) is the second most common cause of cancer in men. According to INCA, in Brazil, in 2016, approximately 61,200 new cases of prostate cancer are estimated. Objective: To identify the demographic and epidemiological characteristics, as well as data on the tumor staging of patients with PCa treated at the Oncology Unit of Hospital Marcio Cunha in the city of Ipatinga-MG in the years of 2012, 2013 and 2014. Methodology: This is a retrospective and descriptive study where 668 patients\' records, with a diagnosis of pathological diagnosis, were analyzed in the years 2012, 2013 and 2014, according to the list provided by the institution, with a diagnosis of PCa registered with ICD-10-C 61. The analyzed variables were: origin, year of diagnosis, age group, self-reported race, risk factors such as smoking, alcoholism, family history of PCa, total PSA at diagnosis, histological type of biopsy, Gleason score of biopsy, histological type of the surgical specimen, Gleason score of the surgical specimen. Data were analyzed using descriptive and inferential statistics, using SPSS software, version 19.0. Results: The highest incidence of PCa cases came from the most populated cities of the analyzed health micro-region and the most prevalent age group was between 61 and 80 years old, with prevalence in brown and whites and with a family history of 17.2% of first-degree relatives degree; With father in 37.3%, brother in 60.8% and son in 1.9%. Only 165 (25.9%) were smokers and 20.8% were alcoholics. PSA levels ranged from 4.1ng / e to 10ng / ml (49.5%) and the higher the age group the higher the PSA values. Brown patients had higher total PSA. When we evaluated whether there was a relationship between total PSA levels and risk factors such as smoking, alcohol consumption and family history, there was only a statistically significant relationship with alcohol consumption. There was concordance of the Gleason score between biopsy and surgical specimen in 70%, subgrade in 18.7% and overdose in 11.3%. Comparing the age of patients with Gleason score, the greater the patient\'s age, the greater the Gleason score of the material obtained by the transrectal biopsy. Smokers and alcoholists presented the highest Gleason score of the surgical specimen. Conclusion: The agreement between the Gleason score of the biopsy and the Gleason score of the surgical specimen was 70%; Higher PSA levels; The longer the age group, the more undifferentiated was the tumor (biopsy); Smokers and alcoholics presented more undifferentiated tumors in the surgical specimen; This is the first epidemiological study of PCa developed in the Vale do Aço region, the socio-demographic characterization and the associations found here can contribute with programs to develop actions of control of PCa in this region.

La médecine fondée sur les preuves a conféré à l’essai contrôlé randomisé (ECR) le plus haut niveau de preuve dans l’évaluation de l’effet des médicaments, et par extension de toute intervention en santé. Cependant, le recours aux études observationnelles s’avère également nécessaire (i) pour conforter, en situation réelle, les résultats issus des ECR dont la validité externe est limitée, (ii) dans des situations, notamment lorsqu’il s’agit d’interventions complexes, où l’ECR n’est pas toujours réalisable pour des questions éthiques et/ou organisationnelles. Toutefois, les études observationnelles sont sujettes à différents types de biais, et notamment au biais d’indication. Ce travail de thèse explore les différentes techniques d’analyse statistique des résultats permettant de maîtriser ce biais. Dans une première partie, les aspects théoriques ont été abordés. Les différentes techniques disponibles ont été identifiées, analysées et comparées : les techniques d’ajustement multivarié, celles utilisant un score de propension (SP) et celles utilisant une variable instrumentale (VI). Pour approfondir les connaissances sur la question, une revue systématique de la littérature a été effectuée. Elle a mis en évidence la faible concordance entre les résultats obtenus en utilisant un SP et ceux obtenus en utilisant une VI, lorsque ces deux techniques étaient utilisées dans une même étude pour évaluer la même intervention. Dans une seconde partie, l’utilisation de SP et/ou VI a été testée dans trois exemples d’évaluation d’interventions complexes à partir de données de pratiques courantes recueillies dans le cadre de deux études observationnelles de cohorte : (i) l’évaluation de l’effet d’un réseau de soins spécialisé dans l’insuffisance cardiaque (IC) sur la mortalité ; (ii) l’évaluation de l’effet des stratégies médicamenteuses appropriées dans l’IC sur la mortalité ; (iii) l’évaluation de l’effet des stratégies antithrombotiques chez les patients hémodialysés sur le risque hémorragique. / Evidence-based medicine placed randomized controlled trials (RCT) at the highest level of evidence to evaluate the effects of medications and, by extension, of all health interventions. Nevertheless, observational studies are necessary (i) to support, in real-world settings, the results of RCTs, the external validity of which is limited, and (ii) in situations where RCTs are not feasible for ethical or practical reasons, particularly when evaluating complex interventions. However, observational studies are particularly prone to confounding by indication. This thesis focuses on analytical methods to reduce this bias. In its first part, the theoretical aspects were addressed. Available methods were identified, reviewed and compared: multivariate adjustment methods, methods using a propensity score (PS) and methods using an instrumental variable (IV). To further knowledge on this issue, a systematic literature review was performed. This review revealed that more and more observational studies simultaneously use PS and IV approaches to evaluate the same intervention, often leading to nonconcordant results that may be dif?cult to interpret. In a second part, the use of PS and/or VI methods was tested in three evaluations of complex interventions in real-world settings, using data from two cohort studies: (i) to evaluate the effectiveness on mortality of a community-based multidisciplinary disease management programme for heart failure (HF) patients; (ii) to evaluate the effectiveness of recommended drug prescriptions on mortality in patients with HF; (iii) to evaluate the effect of antiplatelet and anticoagulant therapies on the risk of major bleeding events in chronic hemodialysis patients