Submitted by Renata Lopes (renatasil82@gmail.com) on 2016-01-26T12:33:12Z No. of bitstreams: 1 vivianassisfayer.pdf: 1716966 bytes, checksum: 6112a4fa69673c197984cf60d4f75fa6 (MD5) / Rejected by Adriana Oliveira (adriana.oliveira@ufjf.edu.br), reason: Renata, verifique se a palavra do título "coorte" está correta on 2016-01-27T10:54:17Z (GMT) / Submitted by Renata Lopes (renatasil82@gmail.com) on 2016-01-27T11:10:08Z No. of bitstreams: 1 vivianassisfayer.pdf: 1716966 bytes, checksum: 6112a4fa69673c197984cf60d4f75fa6 (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2016-01-27T13:56:30Z (GMT) No. of bitstreams: 1 vivianassisfayer.pdf: 1716966 bytes, checksum: 6112a4fa69673c197984cf60d4f75fa6 (MD5) / Made available in DSpace on 2016-01-27T13:56:30Z (GMT). No. of bitstreams: 1 vivianassisfayer.pdf: 1716966 bytes, checksum: 6112a4fa69673c197984cf60d4f75fa6 (MD5) Previous issue date: 2014-02-20 / CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / O câncer de mama é um dos principais problemas de saúde pública em diversas partes do mundo, uma vez que apresenta alta incidência sendo considerado uma das principais causas de morte por câncer entre mulheres em diversos países. Este estudo teve como objetivo avaliar a sobrevida de 10 anos e fatores prognósticos em mulheres com diagnóstico de câncer de mama entre 2000 e 2001, assistidas em centro de referência em assistência oncológica da região Sudeste do país. A coorte foi identificada a partir de busca no Registro Hospitalar de Câncer do serviço de saúde incluído no estudo. O seguimento dos casos foi realizado mediante retorno aos prontuários, complementado por busca no banco do Sistema de Informação sobre Mortalidade (SIM) e contato telefônico. As principais variáveis analisadas foram: natureza do serviço, idade ao diagnóstico, cor da pele, tamanho do tumor, comprometimento de linfonodos, estadiamento, marcadores tumorais, hormonioterapia e radioterapia, entre outras. As funções de sobrevida foram calculadas por meio do método de Kaplan-Meier. O modelo de riscos proporcionais de Cox foi utilizado para avaliação dos fatores prognósticos. Os resultados deste estudo indicam que a sobrevida de dez anos foi de 61,2% (IC95%: 55,0-66,8). Os fatores prognósticos mais importantes que estiveram associados a pior sobrevida, de forma independente, foram: presença de metástase ganglionar (HR=2,64; IC95%: 1,73-4,02), tumores maiores que dois centímetros (HR=1,88; IC95%: 1,15-3,09) e não uso de hormonioterapia (HR=1,53; IC95%: 1,01-2,30). Na análise multivariada estratificada por natureza do serviço de saúde, a presença de metástase ganglionar (HR=2,36; IC95%: 1,42-3,90), tumores com mais de dois centímetros (HR=1,76; IC95%: 0,95-3,25) e utilização de radioterapia (HR=0,67; IC95%: 0,42-1,09) foram os principais fatores prognósticos para pior sobrevida entre as pacientes assistidas pelo serviço público. Já para o serviço privado, estadiamentos mais avançados e ausência de marcadores tumorais associaram-se aos melhores prognósticos (HR=2,82; IC95%: 1,24-6,42). No Brasil, ainda existem poucas publicações a respeito de análise de sobrevida em pacientes com câncer de mama com seguimento de dez anos. Este trabalho foi importante para fornecer informações sobre o seguimento de longo prazo do centro de referência oncológica avaliado e para caracterizar o perfil das pacientes diagnosticadas com câncer de mama no município e regiões adjacentes. Os resultados deste estudo fortalecem a importância do diagnóstico precoce através do rastreamento por meio do exame clínico das mamas e de mamografia, e da disponibilidade do exame imuno-histoquímico para todas as pacientes, independente da natureza do serviço, para orientar a abordagem terapêutica. Estas informações são extremamente úteis para os gestores de saúde adotarem medidas voltadas para a prevenção e controle da doença, assim como para avaliações da qualidade do cuidado prestado e acessibilidade ao sistema de saúde; além de viabilizar avaliações dos avanços diagnósticos e terapêuticos empregados pelo centro de atenção oncológica. / Breast cancer is a major public health problem in many parts of the world, since it has high incidence is considered one of the leading causes of cancer death among women in many countries. This study aimed to evaluate the survival of 10 years and prognostic factors in women diagnosed with breast cancer between 2000 and 2001, assisted at a referral center for cancer care in the Southeast region of the country. Recruitment of cases was made from the search in the Hospital Cancer Registry of the health service included in the study. The follow-up of cases was achieved by returning to the medical records, supplemented by searching the database of the Mortality Information System (MIS) and telephone contact. The main variables analyzed were: type of service, age at diagnosis, race, tumor size, lymph node involvement, staging, tumor markers, hormone therapy and radiotherapy, among others. Survival functions were calculated using the Kaplan-Meier method. The model of Cox proportional hazards regression was used to evaluate prognostic factors.The results of this study indicate that the ten-year survival was 61.2% (95% CI= 55.0 to 66.8). The most important prognostic factors that were associated with worse survival, independently, were: presence of lymph node metastasis (HR = 2.64, 95% CI= 1.73 to 4.02), larger than two centimeters (HR = 1 tumors , 88, 95% CI= 1.15 to 3.09) and no use of hormone therapy (HR = 1.53, 95% CI= 1.01 to 2.30). In multivariate analysis stratified by type of health service, the presence of lymph node metastasis (HR = 2.36, 95% CI= 1.42 to 3.90), tumors larger than two centimeters (HR = 1.76, 95% CI= 0.95 to 3.25) and use of radiotherapy (HR = 0.67, 95% CI= 0.42 to 1.09) were the main prognostic factors for poor survival among patients attended by the public service. As for the private service, more advanced staging and absence of tumor markers were associated with the best prognosis (HR = 2.82, 95% CI= 1.24 to 6.42). In Brazil, there are few publications about survival analysis in patients with breast cancer followed up for ten years. This work was important to provide information on the outcome of long-term oncology referral center assessed and to characterize the profile of patients diagnosed with breast cancer in the city and surrounding regions. The results of this study reinforce the importance of early diagnosis through screening by clinical breast examination and mammography, and availability of immunohistochemical examination for all patients, regardless of the nature of the service, to guide the therapeutic approach. This information is extremely useful for health managers adopt measures aimed at the prevention and control of disease, as well as evaluations of the quality of care delivery and accessibility to health care, as well as providing reviews of diagnostic and therapeutic advances employed by the center cancer care.

O principal objetivo deste trabalho é desenvolver procedimentos inferências em uma perspectiva bayesiana para modelos de sobrevivência com (ou sem) fração de cura baseada na distribuição exponencial por partes. A metodologia bayesiana é baseada em métodos de Monte Carlo via Cadeias de Markov (MCMC). Para detectar observações influentes nos modelos considerados foi usado o método bayesiano de análise de influência caso a caso (Cho et al., 2009), baseados na divergência de Kullback-Leibler. Além disso, propomos o modelo destrutivo binomial negativo com fração de cura. O modelo proposto é mais geral que os modelos de sobrevivência com fração de cura, já que permitem estimar a probabilidade do número de causas que não foram eliminadas por um tratamento inicial / The main objective is to develop procedures inferences in a bayesian perspective for survival models with (or without) the cure rate based on piecewise exponential distribution. The methodology is based on bayesian methods for Markov Chain Monte Carlo (MCMC). To detect influential observations in the models considering bayesian case deletion influence diagnostics based on the Kullback-Leibler divergence (Cho et al., 2009). Furthermore, we propose the negative binomial model destructive cure rate. The proposed model is more general than the survival models with cure rate, since the probability to estimate the number of cases which were not eliminated by an initial treatment

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

O objetivo desta pesquisa foi avaliar o desempenho clínico de restaurações bulk fill em indivíduos com molares decíduos submetidos à RSTC (remoção seletiva de tecido cariado) durante 12 meses de acompanhamento. 62 indivíduos (idade média de 5.9 anos ± 1.74) com no mínimo um molar com lesão cavitada ativa profunda em dentina foram incluídos na pesquisa. 144 molares decíduos foram randomicamente divididos em 2 grupos de material restaurador: Resina Filtek Bulk Fill (grupo teste=FBF) ou Vitremer (grupo controle= CIVMR). Os procedimentos restauradores foram realizados por 2 dentistas especialistas em odontopediatria treinados para realização da técnica e calibrados para avaliação do diagnóstico da atividade da lesão e do desempenho da restauração. Um terceiro examinador também treinado, calibrado e contemplando princípio de cegamento, avaliou as radiografias durante o período de acompanhamento. Os critérios avaliados para determinar o sucesso clínico e radiográfico do tratamento de RSTC foram: ausência de dor espontânea e/ ou sensibilidade à pressão, ausência de sinais de pulpite irreversível, mobilidade anormal incompatível com o período de reabsorção radicular. O desempenho da restauração foi categorizado em cada grupo seguindo o critério da FDI para descoloração marginal, forma anatômica, fratura de material e retenção, adaptação marginal e recorrência de cárie, sendo a falha registrada. Os dentes foram reavaliados clinicamente e radiograficamente quanto o tratamento e desempenho da restauração em 6 e 12 meses. Características sociodemográficas foram registradas no baseline e características clínicas como índice ceod/CPOD, índice de placa visível e índice de sangramento gengival foram registrados durante os períodos de acompanhamento. Os dados foram analisados considerando o tipo de restauração, sintomas, número de faces envolvidas, tipo dente, experiência do operador e padrão de sucesso das restaurações. Para determinar as taxas de sucesso das restaurações foram geradas curvas de sobrevida com o estimador Kaplan-Meyer para cada grupo avaliado. O modelo de regressão de Cox com falhas compartilhadas foi realizado para avaliar diferenças nas taxas de sobrevida das restaurações de acordo com o tratamento e características clínicas e demográficas da amostra. Resultados: A taxa de sucesso das restaurações foi de 83,9% (85,9% para FBF e 81,9% para CIVMR, p = 0,675). Não houve diferenças no risco de falha de acordo com material restaurador. Das 23 falhas, três foram exclusivamente pulpar e uma pulpar e restauradora, e 100% da falha foram causadas por fratura de material. As restaurações oclusoproximais demonstraram a menor taxa de sucesso, aumentando em 4,12 vezes a probabilidade de ter uma falha restauradora em comparação com a restauração oclusal. Os pacientes com perfil “cárie ativos” foram associados a mais falhas do que pacientes “cárie controlados”. O grau de experiência do operador mostrou diferença significativa, e o menos experiente teve mais probabilidade de falhas restauradoras. As restaurações de resina filtek bulk fill realizadas em molares decíduos após a remoção seletiva da cárie mostraram uma sobrevida satisfatória de 85.9% após 12 meses de acompanhamento. / The purpose of this study was to evaluate the clinical performance of bulk-fill restorations in primary molars submitted to SCR (selective caries removal) at 12 months of follow-up. 62 (average age of 5.9 years ± 1.74) were included with at least one molar with active cavitated lesion in deep dentin.144 primary molars were randomly divided into 2 groups of restorative material: Filtek Bulk Fill Resin (test group = FBF) or Vitremer (control group = RMGIC). Restorative procedures were performed by 2 specialists in pediatric dentistry trained to perform the technique and calibrated to evaluate the diagnosis of lesion activity and restoration performance. A third examiner also trained, calibrated and contemplating the principle of blinding, evaluated the radiographs during the follow-up period. The criteria evaluated to determine the clinical and radiographic success of the SCR were: absence of spontaneous pain and / or sensitivity to pressure, absence of signs of irreversible pulpitis, abnormal mobility incompatible with the period of root resorption. Restoration performance was categorized in each group following the FDI criteria for marginal staining, anatomical form, fracture of material and retention, marginal adaptation and recurrence of caries, and the failure was recorded. The teeth were reassessed clinically and radiographically regarding the treatment and performance of the restoration at 6 and 12 months. Sociodemographic characteristics were recorded in the baseline and clinical characteristics such as DMFT index, visible plaque index and gingival bleeding index were recorded during the follow - up. The data were analyzed considering the type of restoration, symptoms, number of faces involved, tooth type, operator experience and success rate of the restorations. To determine the success rates of the restorations, survival curves were generated with the Kaplan-Meyer estimator for each group evaluated. The Cox regression model with shared failures was performed to assess differences in survival rates of restorations according to treatment and clinical and demographic characteristics of the sample. Results: Success rate of restorations was 83.9% (85.9% for FBF and 81.9% for RMGIC, p=0.675). There were no differences in the risk of failure according to the material restoration. Of the 23 failures, three failures were exclusive pulp and one pulp and restorative, and 100% of the failure were to due to fracture of material. Occlusoproximal restorations demonstrated the lowest success rate, increases 4.12 times the probability of having a restorative failure compared to occlusal restoration. Patients with "active caries" profile were associated with more failures than "controlled caries" patients. The degree of operator experience showed significant difference, and the less experienced operator had more restorative failures. Bulk Fill composite restorations performed in molar primary teeth after selective caries removal showed satisfactory survival of 85.9% after 12 months of follow-up.

Objetivo: Avaliar o impacto do nascimento pré-termo sobre a mortalidade neonatal numa série temporal de 2000 a 2014 no município de Porto Alegre. Métodos: Estudo de coorte retrospectivo de base populacional com a utilização dos registros oficiais de nascimento e de morte ligados de 2000 a 2014 de recém-nascidos com menos de 32 semanas de idade gestacional de Porto Alegre. Foram utilizadas como variáveis independentes idade e escolaridade maternas, número de consultas pré-natal, tipo de hospital, via de parto, idade gestacional (IG), sexo e peso do recém-nascido e ano de nascimento. O desfecho primário foi morte neonatal (morte ocorrida de 0 a 27 dias de vida). Foram excluídos recém-nascidos duplicados, com menos de 500 gramas ou com peso inconsistente, com IG menor de 22 semanas, com anomalias congênitas, gemelares e de partos extra hospitalares. Foi calculado razão de risco (hazard ratio-HR) ajustado para o risco de morte neonatal para todas as variáveis independentes através de análise de sobrevivência pela regressão de Cox para riscos proporcionais com nível de significância p<0,05. Posteriormente, foi realizado análise por quintil de peso de nascimento. Resultados: Foram analisados os registros de 3282 recém-nascidos com IG menor que 32 semanas de 2000 a 2014 dos quais, 643 foram ao óbito neonatal e 2639 sobreviveram. O risco de morte neonatal absoluto diminuiu de 25% no triênio 2000 a 2002 para 17% no período de 2012 a 2014. O mesmo risco ajustado foi significativamente menor para os recém-nascidos de menor peso (média de 673 ± 86 gramas) de parto cesáreo [HR 0,57 (IC95% 0,45-0,73)] enquanto que, para os de maior peso (média 1.834 ± 212 gramas) este risco inverteu-se e foi significativamente maior para esta via de parto [HR 8,44 (IC95% 1,86-38,22)]. Conclusão: Houve diminuição do risco absoluto de morte neonatal entre os recém-nascidos com IG menor de 32 semanas nos últimos anos em Porto Alegre e o aprimoramento do uso racional do parto cesáreo nos hospitais do município pode contribuir para uma redução ainda maior desse indicador. / Objective: Assess impact of prematurity on neonatal mortality from 2000 to 2014 in Porto Alegre through official information systems. Methods: Populational base retrospective cohort study with record linkage of birth and death database certificates. There were included records of birth and death from 2000 to 2014 of infants with less than 32 weeks of gestational age of Porto Alegre. There were used mother age and schooling, number of antenatal visits, delivery type, hospital type, gestational age, sex and birth weight and birth year of infant as independent variables. The primary outcome examined was neonatal death (death at 0-27 days of age). There were excluded infant records duplicate, with less than 500g or inconsistent birthweight, with gestational age less than 22 weeks, with congenital anomalies, twins and out-of-hospital births. Adjusted Hazard Ratio (HR) were calculated for the risk of neonatal death for all independent variables through Cox regression for survival analysis with p-value<0,05 for statistical significance. The analysis also was performed at quintiles of birthweight. Results: There were 3282 infant records of infants with less than 32 weeks of gestational age from 2000 to 2014 who progress to 643 neonatal deaths or 2639 survival. The neonatal death absolut risk decline from 25% at 2000-2002 period to 17% at 2012-2014 period. The adjusted neonatal death risk was significantly reduced for lightest preterm (mean birthweight 673g ± 86) born by C-section [HR 0.57 (CI95% 0.45-0.73)], while, for the heaviest ones (mean birthweight 1.834g ± 212) the risk was significantly increased for that delivery route [HR 8.44 (CI95% 1.86-38.22)]. Conclusion: The absolut risk of neonatal death in infants with less than 32 weeks of gestational age has been declining over the years and more rational use of C-section can contribute to further improving the neonatal survival.

Cette thèse est consacrée à l'inférence de certains modèles statistiques mal spécifiés. Chaque résultat obtenu trouve son application dans une étude sur les facteurs pronostiques dans le cancer du sein, grâce à des données collectées par l'Institut Curie. Dans un premier temps, nous nous intéressons au modèle à risques non proportionnels, et exploitons la connaissance de la survie marginale du temps de décès. Ce modèle autorise la variation dans le temps du coefficient de régression, généralisant ainsi le modèle à hasards proportionnels. Dans un deuxième temps, nous étudions un modèle à hasards non proportionnels ayant un coefficient de régression constant par morceaux. Nous proposons une méthode d'inférence pour un modèle à un unique point de rupture, et une méthode d'estimation pour un modèle à plusieurs points de rupture. Dans un troisième temps, nous étudions l'influence du sous-échantillonnage sur la performance des forêts médianes et essayons de généraliser les résultats obtenus aux forêts aléatoires de survie à travers une application. Enfin, nous présentons un travail indépendant où nous développons une nouvelle méthode de recherche de doses, dans le cadre des essais cliniques de phase I à ordre partiel. / The thesis focuses on inference of statistical misspecified models. Every result finds its application in a prognostic factors study for breast cancer, thanks to the data collection of Institut Curie. We consider first non-proportional hazards models, and make use of the marginal survival of the failure time. This model allows a time-varying regression coefficient, and therefore generalizes the proportional hazards model. On a second time, we study step regression models. We propose an inference method for the changepoint of a two-step regression model, and an estimation method for a multiple-step regression model. Then, we study the influence of the subsampling rate on the performance of median forests and try to extend the results to random survival forests through an application. Finally, we present a new dose-finding method for phase I clinical trials, in case of partial ordering.

Le stade terminal de l’insuffisance cardiaque peut nécessiter l’inscription en liste d’attente pour une greffe cardiaque. L’offre en greffon étant faible, l’accès à cette thérapeutique est priorisé pour les malades les plus graves dans le système d’allocation actuel, faisant de l’équité un enjeu éthique et sociétal important. L’objectif de cette thèse, grâce à des méthodes originales et de nouvelles sources de données, était d’analyser les déterminants d’accès à la greffe cardiaque liés aux candidats et aux équipes de greffe en France sur une période récente et d’apporter des éléments nouveaux sur le parcours de soin de ces malades. L’analyse du devenir des candidats à une greffe cardiaque inscrits entre 2010 et 2013 en intégrant la méthode de risques compétitifs a permis de dissocier le risque médical du risque induit par le système d’allocation actuel. Nous avons ainsi trouvé 7 facteurs relatifs au candidat associés à un risque de mortalité élevé dont 4 reliés à la sévérité de l’insuffisance cardiaque et 3 non spécifiques de l’insuffisance cardiaque mais associés à un faible accès à la greffe. La prise en compte de l’effet centre sur les inscrits entre 2010 et 2014 grâce à un modèle mixte de survie a permis de déterminer les facteurs équipe associés à l’accès à la greffe. Parmi les 23 équipes de greffe en France, l’ajustement sur les facteurs candidat et équipe, permettait d’observer que 5 équipes avaient des résultats différents des autres dont 3 avec un accès défavorisé. La mise en évidence de disparités médicales, géographiques et structurelles, révélatrices de failles dans le système actuel d’allocation des greffons cardiaques, nous a permis d’étayer la discussion sur la mise en place d’un nouveau système d’attribution en France. L’attribution au patient plutôt qu’à l’équipe de greffe devrait être plus pertinente. Les méthodologies utilisées permettront d’évaluer précisément ces répercussions. Enfin, l’étude du parcours hospitalier un an avant greffe (2010-2015) à l’aide des données du PMSI a permis de constater un nombre élevé d’hospitalisations pré-inscription majoritairement liées à leur défaillance cardiaque et de longue durée, confirmant le caractère réfractaire de ces insuffisants cardiaques. Ce travail sera poursuivi par une étude approfondie de la consommation de soins de ces malades grâce aux données du SNDS, indispensable étape pour évaluer la prise en charge et estimer les besoins en greffe cardiaque. / Heart transplantation (HTx) is the preferred option for medically refractory advanced heart failure. Due to the small number of available grafts, current allocation policy in France, as in many other countries, is based on the severity of the candidate’s heart disease. This Ph. D thesis was designed to determine candidate and center factors associated with access to heart transplantation in France and in-hospital care pathways one year before heart transplantation using appropriate methodologies and the national hospital database. We first analyzed 1-year mortality in patients listed for HTx in France from 2010 to 2013 using competing risk models in order to distinguish patient-related predictors and the influence of allocation policy. We then distinguished the proportions explained by candidate characteristics and center factors with the persistent between-center variability on 1-year access to transplantation (2010-2014). These disparities are mediated by the severity of the candidate’s heart disease, the allocation system and listing practices rather than by transplant activity. These findings provide a new contribution to improve the heart transplant allocation scoring system in France. The study based on the nationwide administrative database overcomes a major limitation of the national transplantation registry by shedding light on the healthcare pathway of heart transplanted recipients (2010-2015) during the year prior to transplantation. These findings will be useful to assess the medical benefits and criteria for registration on the heart transplant waiting list. This study will be continued by a detailed analysis of the healthcare consumption of these patients based on French national health insurance (SNDS) data.

Dengue fever affects over 390 million people annually worldwide and is of particu- lar concern in Southeast Asia where it is one of the leading causes of hospitalization. Modeling trends in dengue occurrence can provide valuable information to Public Health officials, however many challenges arise depending on the data available. In Thailand, reporting of dengue cases is often delayed by more than 6 weeks, and a small fraction of cases may not be reported until over 11 months after they occurred. This study shows that incorporating data on Google Search trends can improve dis- ease predictions in settings with severely underreported data. We compare penalized regression approaches to seasonal baseline models and illustrate that incorporation of search data can improve prediction error. This builds on previous research show- ing that search data and recent surveillance data together can be used to create accurate forecasts for diseases such as influenza and dengue fever. This work shows that even in settings where timely surveillance data is not available, using search data in real-time can produce more accurate short-term forecasts than a seasonal baseline prediction. However, forecast accuracy degrades the further into the future the forecasts go. The relative accuracy of these forecasts compared to a seasonal average forecast varies depending on location. Overall, these data and models can improve short-term public health situational awareness and should be incorporated into larger real-time forecasting efforts.

No description available.

Pour les patients diagnostiqués avec un cancer au stade précoce, les décisions de traitement dépendent de l’évaluation du risque de rechute métastatique. Les outils de pronostic actuels sont fondés sur des approches purement statistiques, sans intégrer les connaissances disponibles sur les processus biologiques à l’oeuvre. L’objectif de cette thèse est de développer des modèles prédictifs du processus métastatique en utilisant une approche de modélisation mécaniste et la modélisation à effets mixtes. Dans la première partie, nous étendons un modèle mathématique du processus métastatique pour décrire la croissance de la tumeur primaire et de la masse métastatique totale chez des souris traitées avec le sunitinib (un inhibiteur de tyrosine kinase ayant une action anti-angiogénique) administré comme traitement néoadjuvant (i.e. avant exérèse de la tumeur primaire). Le modèle est utilisé pour tester des hypothèses expliquant les effets différentiels du sunitinib sur la tumeur primaire et les métastases. Des algorithmes d’apprentissage statistique sont utilisés pour évaluer la valeur prédictive des biomarqueurs sur les paramètres du modèle.Dans la deuxième partie de cette thèse, nous développons un modèle mécaniste pour la prédiction du temps de rechute métastatique et le validons sur des données cliniques des patientes atteintes d’un cancer du sein localisé. Ce modèle offre des prédictions personnalisées des métastases invisibles au moment du diagnostic, ainsi que des simulations de la croissance métastatique future, et il pourrait être utilisé comme un outil de prédiction individuelle pour aider à la gestion des patientes atteintes de cancer du sein. / For patients diagnosed with early-stage cancer, treatment decisions depend on the evaluation of the risk of metastatic relapse. Current prognostic tools are based on purely statistical approaches that relate predictor variables to the outcome, without integrating any available knowledge of the underlying biological processes. The purpose of this thesis is to develop predictive models of the metastatic process using an established mechanistic modelling approach and the statistical mixed-effects modelling framework.In the first part, we extend the mathematical metastatic model to describe primary tumour and metastatic dynamics in response to neoadjuvant sunitinib in clinically relevant mouse models of spontaneous metastatic breast and kidney cancers. The calibrated model is then used to test possible hypothesis for the differential effects of sunitinib on primary tumour and metastases, and machine learning algorithms are applied to assess the predictive power of biomarkers on the model parameters.In the second part of this thesis, we develop a mechanistic model for the prediction of the time to metastatic relapse and validate it on a clinical dataset of breast cancer patients. This model offers personalised predictions of the invisible metastatic burden at the time of diagnosis, as well as forward simulations of metastatic growth, and it could be used as a personalised prediction tool to assist in the routine management of breast cancer patients.