Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Desde o início dos estudos lingüísticos das afasias, muito se discute a validade de se agruparem indivíduos para a realização de pesquisas. Alguns autores criticam os estudos de grupo por considerarem haver muita variação entre os indivíduos estudados, e serem os critérios utilizados para os agruparem pré-teóricos e subjetivos, visto que, geralmente, são baseados em impressões clínicas e não são lingüisticamente estruturados. Esses autores acreditam que apenas estudos de caso sejam metodologicamente adequados (Badecker e Caramazza, 1985, 1991). Em oposição, os defensores dos estudos de grupo consideram que apenas esse tipo de metodologia pode dar conta da citada variação, pois estudos de caso não permitem identificar as exceções no desempenho dos indivíduos (Caplan, 1986; Zurif, Swinney e Fodor, 1991). A despeito de todo o debate, poucos autores procuram soluções para a principal causa do desacordo citado, que pode ser considerada uma das maiores deficiências da área: a má seleção de sujeitos de pesquisa. Os critérios de seleção de pacientes comumente utilizados são teoricamente ultrapassados, podendo mesmo variar dependendo dos objetivos de cada pesquisador (Caplan, 1995). A presente dissertação tem como objetivo demonstrar a necessidade de se reverem esses critérios. Por meio de uma pesquisa realizada com quatro afásicos de Broca selecionados pelos critérios clássicos, procuramos demonstrar como a má seleção de pacientes pode prejudicar o desenvolvimento da área por tornar tendenciosos tanto estudos de caso quanto de grupo, aumentando a falta de consenso teórico entre os autores. Para tal, os pacientes foram submetidos a dois experimentos que investigavam seus desempenhos na compreensão de sentenças ativas, passivas e interrogativas QU, de sujeito, objeto e objeto in situ. Ainda, foi realizada uma avaliação desses pacientes sem que se assumissem tais critérios de classificação, partindo-se de uma caracterização lingüística inicial obtida a partir dos resultados nas sentenças ativas e passivas, ficando esta análise isenta dos problemas de seleção de pacientes comuns tanto aos estudos de caso quanto aos estudos de grupo. Como resultado, observou-se que a utilização de uma forma diferente de análise dos desempenhos pode informar a respeito do déficit, bem como sobre a organização do sistema lingüístico, sem a necessidade de se vincular a uma metodologia de caso ou de grupo no sentido tradicional / Since the beginning of linguistic aphasiology studies, the validity of grouping aphasics for research purposes has been widely debated. Some authors criticize group studies because they believe there is too much variation among individuals performance and consider the criteria used for grouping patients together to be subjective and pre-theoretical, since those are usually defined by clinical impressions and are not linguistically grounded. Under these assumptions they judge the case studies to be the only adequate methodology (Badecker e Caramazza, 1985, 1991). On the opposite direction, there are some authors who support group studies as they believe that solely this methodology could account for the variation, which would be impossible on case studies, in which the identification of exceptionalities on individuals performance is unattainable (Caplan, 1986; Zurif, Swinney e Fodor, 1991). In spite of this debate, there are only a few authors who actually look for solutions to the main cause of the conflict, which may also be considered as one of the major flaws in this field of study. The criteria of patients selection commonly applied in the studies are theoretically antiquated and can even vary depending on each authors objectives (Caplan, 1995). The goal of this dissertation is to demonstrate the need for a reevaluation of these criteria. With this intention in mind, we undertook an investigation of four (traditionally classified) Brocas aphasics comprehension on two experimental tasks which measured their performance on active and passive sentences, as well as in subject, object and in situ object WH questions. The results of these experiments allowed us to demonstrate how a problematic patients selection hinders the development of this field of study, as it makes both case and group studies biased, increasing the theoretical disagreements among the authors. Thus, conversely to what is generally done, we carried out an evaluation of the patients performances without assuming those problematic, pre-established classification criteria, which relieves our work from the problems found on case or group studies, since we assumed an initial linguistic characterization of the patients, considering their results on active and passive sentences, instead of the clinical one. As a result, we noticed that through this different way of analyzing patients performance it is possible to get information concerning the deficit and the organization of the linguistic system, without needing to be compromised with case or group views in a traditional way

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Desde o início dos estudos lingüísticos das afasias, muito se discute a validade de se agruparem indivíduos para a realização de pesquisas. Alguns autores criticam os estudos de grupo por considerarem haver muita variação entre os indivíduos estudados, e serem os critérios utilizados para os agruparem pré-teóricos e subjetivos, visto que, geralmente, são baseados em impressões clínicas e não são lingüisticamente estruturados. Esses autores acreditam que apenas estudos de caso sejam metodologicamente adequados (Badecker e Caramazza, 1985, 1991). Em oposição, os defensores dos estudos de grupo consideram que apenas esse tipo de metodologia pode dar conta da citada variação, pois estudos de caso não permitem identificar as exceções no desempenho dos indivíduos (Caplan, 1986; Zurif, Swinney e Fodor, 1991). A despeito de todo o debate, poucos autores procuram soluções para a principal causa do desacordo citado, que pode ser considerada uma das maiores deficiências da área: a má seleção de sujeitos de pesquisa. Os critérios de seleção de pacientes comumente utilizados são teoricamente ultrapassados, podendo mesmo variar dependendo dos objetivos de cada pesquisador (Caplan, 1995). A presente dissertação tem como objetivo demonstrar a necessidade de se reverem esses critérios. Por meio de uma pesquisa realizada com quatro afásicos de Broca selecionados pelos critérios clássicos, procuramos demonstrar como a má seleção de pacientes pode prejudicar o desenvolvimento da área por tornar tendenciosos tanto estudos de caso quanto de grupo, aumentando a falta de consenso teórico entre os autores. Para tal, os pacientes foram submetidos a dois experimentos que investigavam seus desempenhos na compreensão de sentenças ativas, passivas e interrogativas QU, de sujeito, objeto e objeto in situ. Ainda, foi realizada uma avaliação desses pacientes sem que se assumissem tais critérios de classificação, partindo-se de uma caracterização lingüística inicial obtida a partir dos resultados nas sentenças ativas e passivas, ficando esta análise isenta dos problemas de seleção de pacientes comuns tanto aos estudos de caso quanto aos estudos de grupo. Como resultado, observou-se que a utilização de uma forma diferente de análise dos desempenhos pode informar a respeito do déficit, bem como sobre a organização do sistema lingüístico, sem a necessidade de se vincular a uma metodologia de caso ou de grupo no sentido tradicional / Since the beginning of linguistic aphasiology studies, the validity of grouping aphasics for research purposes has been widely debated. Some authors criticize group studies because they believe there is too much variation among individuals performance and consider the criteria used for grouping patients together to be subjective and pre-theoretical, since those are usually defined by clinical impressions and are not linguistically grounded. Under these assumptions they judge the case studies to be the only adequate methodology (Badecker e Caramazza, 1985, 1991). On the opposite direction, there are some authors who support group studies as they believe that solely this methodology could account for the variation, which would be impossible on case studies, in which the identification of exceptionalities on individuals performance is unattainable (Caplan, 1986; Zurif, Swinney e Fodor, 1991). In spite of this debate, there are only a few authors who actually look for solutions to the main cause of the conflict, which may also be considered as one of the major flaws in this field of study. The criteria of patients selection commonly applied in the studies are theoretically antiquated and can even vary depending on each authors objectives (Caplan, 1995). The goal of this dissertation is to demonstrate the need for a reevaluation of these criteria. With this intention in mind, we undertook an investigation of four (traditionally classified) Brocas aphasics comprehension on two experimental tasks which measured their performance on active and passive sentences, as well as in subject, object and in situ object WH questions. The results of these experiments allowed us to demonstrate how a problematic patients selection hinders the development of this field of study, as it makes both case and group studies biased, increasing the theoretical disagreements among the authors. Thus, conversely to what is generally done, we carried out an evaluation of the patients performances without assuming those problematic, pre-established classification criteria, which relieves our work from the problems found on case or group studies, since we assumed an initial linguistic characterization of the patients, considering their results on active and passive sentences, instead of the clinical one. As a result, we noticed that through this different way of analyzing patients performance it is possible to get information concerning the deficit and the organization of the linguistic system, without needing to be compromised with case or group views in a traditional way

Introduction: La procédure d'implantation de valve aortique transcathéter (TAVI) a été développée comme une alternative à la chirurgie cardiaque conventionnelle de remplacement de la valve aortique (RVA) chez des patients à plus haut risque («Society of Thoracic Surgery» (STS) ≥3%) présentant une sténose aortique sévère symptomatique. Cependant, la sélection de potentiels candidats peut demeurer un défi, en partie dû à la paucité de données sur des métriques de performance déterminées pour ce genre de procédure. Les objectifs de cette thèse étaient: 1) l'évaluation du taux de réadmissions hospitalières (une métrique importante de performance) post procédure de TAVI et l'identification des prédicteurs qui y sont associés; 2) l'évaluation de la performance de 4 marqueurs de fragilité communément utilisés comme prédicteurs de résultats cliniques précoces et tardifs chez des patients qui ont eu une procédure de TAVI et l'identification de seuils optimaux pour la stratification du risque; et 3) l'évaluation des résultats cliniques au cours des années de patients avec sténose aortique sévère symptomatique discutés par un «Heart Team» afin d'évaluer si les résultats cliniques TAVI versus RVA standard s'améliorent sur différentes périodes de temps et quelles en sont les raisons. Méthode: Une étude de cohorte rétrospective a été menée afin d'évaluer la valeur des comorbidités de base ainsi que des complications procédurales comme prédicteurs de réadmissions hospitalières précoces et tardives suivant une procédure de TAVI. L'extension de Fine et Gray des modèles de régression de Cox a été utilisée afin de considérer dans l'analyse les risques compétitifs. Quatre variables déterminantes de la fragilité (i.e., albumine sérique, le test de marche de 5 mètres, la force de préhension, et le questionnaire d'index Katz sur l'autonomie) ont été évaluées dans une seconde étude de cohorte. Les associations entre les indicateurs de fragilité et un composé d'évènements d'effets adverses majeurs ont été évaluées selon des modèles de régression logistique, avec des seuils optimaux identifiés pour chacun des indicateurs de la fragilité en utilisant la méthode statistique des courbes de ROC. Finalement, la mortalité à 30 jours et les complications post-procédurales ont été examinées chez des patients évalués en majorité par un «Heart Team» entre 2012 et 2019. Une analyse de scores de propension-appareillés a été effectuée. Résultats: Nous avons identifiés des comorbidités de base et des complications procédurales qui étaient directement associées avec des réadmissions précoces et tardives suivant une procédure de TAVI. L'anémie et un AVC post-opératoire ont été associés avec une augmentation de la mortalité. Parmi les patients à risque élevé qui vont subir un TAVI, l'albumine sérique, être autonome dans ses activités de la vie quotidienne, le test de marche de 5 mètres ont indépendamment prédit la survenue du composé d'évènements adverses à 30 jours. Un nouveau modèle de fragilité (4 indicateurs de fragilité, l'âge et le sexe) a été proposé et a démontré avoir une meilleure discrimination que le score de risque habituellement utilisé, celui de la «Society of Thoracic Surgeons (STS) ». Aussi, des seuils individuels d'augmentation du risque ont été trouvés pour chaque variable individuelle de fragilité. Finalement, une amélioration des résultats cliniques post TAVI et RVA standard a été observée au cours des années, démontrant une réduction de la mortalité à 30 jours (4,0% à 0,4%), des AVCs (2,5% à 0,9%), de l'insuffisance rénale (3.5% à 1.5%), de pneumonie (10.0% à 1.5%) et de la durée de séjour (de 7 jours à 2 jours). Nous avons identifié des complications différentes pour chacune de ces interventions et discuterons des explications possibles pour l'amélioration de ces résultats au cours des années. Conclusions: Cette thèse a identifié des métriques simples qui peuvent assister les «Heart Team» dans leur sélection de patients à plus haut risque qui présentent une sténose aortique sévère pour une procédure TAVI. Nous avons également redéfini des seuils de risque aux indicateurs de fragilité connus. Finalement, l'augmentation du nombre de procédures TAVI au cours des années, a résulté en une diminution de la mortalité et de morbidités autant chez les patients ayant eu une procédure TAVI que chirurgicale. / Objective: Transcatheter aortic valve intervention (TAVI) has emerged as an alternative to conventional surgical replacement of the aortic valve (SAVR) in high-risk patients (STS score ≥3%) with symptomatic aortic stenosis. Nevertheless, the selection of suitable candidates remains challenging, in part due to a paucity of data regarding performance metrics. The objectives of this thesis were: 1) to assess the rate of hospital readmission (an important performance metric) following TAVI and identify associated predictors; 2) to evaluate the performance of four commonly used frailty markers as predictors of early and late outcomes among patients undergoing TAVI and identify optimal cutoff values for risk stratification; and 3) to assess the clinical results over time of patients with aortic stenosis with a procedural indication who were discussed by a «Heart Team» in order to evaluate if TAVI and conventional surgical outcomes improve over different time periods and what are the reasons. Methods: A retrospective cohort study was conducted to assess the value of baseline comorbidities and procedural complications in predicting early and late hospital readmissions following TAVI. The Fine and Gray extension of Cox regression models was used to account for competing risks. Four frailty variables (i.e., serum albumin, 5-meter walk test, grip strength, and Katz index of independence) were assessed in a second cohort study. Associations between frailty indicators and a composite endpoint of major adverse events were assessed in logistic regression models, with optimal cutoff values for each frailty indicator identified using receiver-operating characteristics curves. Finally, 30-day mortality and post-procedural complications were examined in patients triaged in majority by a «Heart Team» between 2012 and 2019. A propensity-matched analysis was performed. Results: We identified baseline comorbidities and procedural complications that were directly associated with early and late readmissions following TAVI. Anemia and post-operative stroke were associated with increased mortality. Among high-risk patients undergoing TAVI, serum albumin, activities of daily living, and 5 meters walk test independently predicted the 30-day composite endpoints. A new frailty model (four frailty indicators, age, and sex) was proposed and found to provide superior discrimination compared to The Society of Thoracic Surgeons risk model. Individual frailty variable cutoff values were identified. Finally, improvement in both TAVI and conventional surgical outcomes were observed over time, with a reduction in 30-day mortality (4,0% to 0,4%), stroke (2,5% to 0,9%), renal failure (3,5% to 1,5%), pneumonia (10,0% to 1,5%), and length of stay (7 to 2 days). The two interventions (TAVI and AVR) were associated with different complications and we will discuss about the possible explanations for the improvement in outcomes. Conclusion: This thesis identified simple metrics that could assist the «Heart Team» in selecting appropriate candidates with severe symptomatic aortic stenosis for TAVI, and refined cut-off values for common frailty indicators. Finally, an increased use of TAVI over the years with broadening indications resulted in a decrease in complications and mortality in patients who received a TAVI, but also patients who received a standard AVR.