The side-effects of chemotherapy treatment are an increasingly important research focus as more cancer patients are reaching survivorship. While treatment allows for survival, it can also lead to problems which can significantly affect quality of life. Cognitive impairments after chemotherapy treatment are one such factor. First presented as anecdotal patient reports, over the last decade empirical evidence for this cognitive concern has been obtained. Much attention has been focused on post-chemotherapy research, yet little attention has been granted to these same patients’ cognition before treatment commences. Breast cancer (BC) patients face many obstacles before chemotherapy treatment such as: surgery and side-effects of anesthesia, increased cytokine activity, stress of a new disease diagnosis and upcoming challenges, and emotional burdens such as depression and anxiety. Many of these factors have independently been shown to affect cognitive abilities in both healthy populations as well as other patient groups. Therefore, the pre-treatment (or baseline) BC patient status warrants systematic study. This would then reduce mistakenly attributing carried-over cognitive deficits to side effects of chemotherapy. As well, it is possible that certain confounding variables may have neural manifestations at baseline that could be exacerbated by chemotherapy agents. The following thesis first presents a review paper which critically describes the current literature examining chemotherapy-related cognitive impairments (CRCIs), as well as possible confound variables affecting this population. Subsequently, three original research papers present pre-chemotherapy data showing significant neuroanatomical and neurofunctional differences in BC patients compared to controls. In particular, these neural differences are present in brain regions that have been reported in post-chemotherapy papers. This, as well as the effects of variables such as the number of days since surgery, depression and anxiety scores and more, support the initiative that research attention should increase focus on these patients at baseline in order to better understand their post-chemotherapy results.

Cette thèse étudie quelques problèmes d’identification et d’estimation dans les modèles de survie bivariée, avec présence d’hétérogénéité individuelle et des facteurs communs stochastiques.Chapitre I introduit le cadre général.Chapitre II propose un modèle pour la mortalité des deux époux dans un couple. Il permet de distinguer deux types de dépendance : l’effet de deuil et l’effet lié au facteur de risque commun des deux époux. Une analyse de leurs effets respectifs sur les primes d’assurance écrites sur deux têtes est proposée.Chapitre III montre que, sous certaines hypothèses raisonnables, on peut identifier l’évolution jointe du risque d’entrer en dépendance et du risque de mortalité, à partir des données de mortalité par cohortes. Une application à la population française est proposée.Chapitre IV étudie la queue de distribution dans les modèles de survie bivariée. Sous certaines hypothèses, la loi jointe des deux durées résiduelles converge, après une normalisation adéquate. Cela peut être utilisé pour analyser le risque parmi les survivants aux âges élevés. Parallèlement, la distribution d’hétérogénéité parmi les survivants converge vers une distribution semi-paramétrique. / This thesis comprises three essays on identification and estimation problems in bivariate survival models with individual and common frailties.The first essay proposes a model to capture the mortality dependence of the two spouses in a couple. It allows to disentangle two types of dependencies : the broken heart syndrome and the dependence induced by common risk factors. An analysis of their respective effects on joint insurance premia is also proposed.The second essay shows that, under reasonable model specifications that take into account the longevity effect, we can identify the joint distribution of the long-term care and mortality risks from the observation of cohort mortality data only. A numerical application to the French population data is proposed.The third essay conducts an analysis of the tail of the joint distribution for general bivariate survival models with proportional frailty. We show that under appropriate assumptions, the distribution of the joint residual lifetimes converges to a limit distribution, upon normalization. This can be used to analyze the mortality and long-term care risks at advanced ages. In parallel, the heterogeneity distribution among survivors converges also to a semi-parametric limit distribution. Properties of the limit distributions, their identifiability from the data, as well as their implications are discussed.

Syftet med uppföljningsstudien var att undersöka deltagares upplevelser av verksamma/hindrande komponenter i behandlingen Hästunderstödd Kognitiv Beteendeterapi (HU-KBT) 12-18 månader efter avslutad behandling. Syftet var också att undersöka vad deltagarna upplever att de uppnått genom att delta i behandlingen HU-KBT. Semistrukturerade intervjuer genomfördes med fem deltagare. Intervjumaterialet analyserades med hjälp av kvalitativ tematisk analys. Ett flertal komponenter i behandlingen beskrevs som hjälpsamma; gruppen, terapeuten, hästen, miljön och olika specifika psykoterapeutiska komponenter. Hindrande behandlingskomponenter som framkom var negativa aspekter av att behandlas i grupp samt för lite tid. Effekter av behandlingen som beskrevs av deltagarna var förbättrad ångesthanteringsförmåga, förhöjd energinivå, förbättrad gränssättningsförmåga, minskat behov av att ha kontroll och ta ansvar för andra, ökad självinsikt, en ökad medvetenhet om och förståelse för känslor och hur dessa kan hanteras samt en förbättrad arbetsförmåga. Övrigt som framkom i intervjuerna var att förväntningarna var blandade, att det kunde finnas övriga livsomständigheter som kan ha påverkat utfallet av behandlingen, att effekten av behandlingen ofta var fördröjd, att behandlingen upplevdes som effektiv, att det förekom olika sätt som användes för att förhindra återfall samt att det fanns önskemål om att få ta del av ytterligare behandling och att behandlingen ska rikta sig till fler grupper av människor och även kunna erbjudas i förebyggande syfte. Resultaten ser lovande ut men det behövs ytterligare forskning för att vidare undersöka behandlingens effektivitet och verksamma komponenter. / The purpose of the follow-up study was to examine the participants' experiences of active/inhibitory components in the treatment Equine Assisted Cognitive Behavioral Therapy (HU-KBT) 12-18 months after treatment completion. The purpose was also to investigate the participant´s self-perceptions of achievement by participating in the HU-KBT treatment. Semi-structured interviews were conducted with five participants. The interview material was analyzed using qualitative thematic analysis. Several components of the treatment were described as helpful; the group, the therapist, the horse, the environment and several specific psychotherapeutic components. Obstructive treatment components that emerged were negative aspects of being treated in group and shortage of time. Effects of the treatment described by the participants were improved anxiety management, increased energy levels, improved ability to set boundaries, reductions in need to control others or to take responsibility for them, increased self-awareness, increased awareness and understanding of emotions and how they can be managed and an improved ability to work. Other things that emerged in the interviews were expectations were mixed; other life circumstances may have affected the outcome of the treatment; the effect of the treatment was often delayed; the treatment was perceived as effective; various ways were used to prevent relapse; there was a desire to gain access to further treatment and that the treatment should be aimed at more groups of people and also be offered for preventive purposes. The results look promising, but further research is needed to further investigate the efficacy and the operating components of the treatment.

<p>    Adaptive MCMC is advantageous over traditional MCMC due to its ability to automatically adjust its proposal distributions during the sampling process, providing improved sampling efficiency and faster convergence to the target distribution, especially in complex or high-dimensional problems. However, designing and validating the adaptive scheme cautiously is crucial to ensure algorithm validity and prevent the introduction of biases. This dissertation focuses on the use of Adaptive MCMC for deep learning, specifically addressing the mode collapse issue in Generative Adversarial Networks (GANs) and implementing Fiducial inference, and its application to Causal inference in individual treatment effect problems.</p> <p><br></p> <p>    First, GAN was recently introduced in the literature as a novel machine learning method for training generative models. However, GAN is very difficult to train due to the issue of mode collapse, i.e., lack of diversity among generated data. We figure out the reason why GAN suffers from this issue and lay out a new theoretical framework for GAN based on randomized decision rules such that the mode collapse issue can be overcome essentially. Under the new theoretical framework, the discriminator converges to a fixed point while the generator converges to a distribution at the Nash equilibrium.</p> <p><br></p> <p>    Second, Fiducial inference was generally considered as R.A. Fisher's a big blunder, but the goal he initially set, <em>making inference for the uncertainty of model parameters on the basis of observations</em>, has been continually pursued by many statisticians. By leveraging on advanced statistical computing techniques such as stochastic approximation Markov chain Monte Carlo, we develop a new statistical inference method, the so-called extended Fiducial inference, which achieves the initial goal of fiducial inference. </p> <p><br></p> <p>    Lastly, estimating ITE is important for decision making in various fields, particularly in health research where precision medicine is being investigated. Conditional average treatment effect (CATE) is often used for such purpose, but uncertainty quantification and explaining the variability of predicted ITE is still needed for fair decision making. We discuss using extended Fiducial inference to construct prediction intervals for ITE, and introduces a double neural net algorithm for efficient prediction and estimation of nonlinear ITE.</p>

Les modèles uplift (levier en français) traitent de l'inférence de cause à effet pour un facteur spécifique, comme une intervention de marketing. En pratique, ces modèles sont construits sur des données individuelles issues d'expériences randomisées. Un groupe traitement comprend des individus qui font l'objet d'une action; un groupe témoin sert de comparaison. La modélisation uplift est utilisée pour ordonner les individus par rapport à la valeur d'un effet causal, par exemple, positif, neutre ou négatif. Dans un premier temps, nous proposons une nouvelle façon d'effectuer la sélection de modèles pour la régression uplift. Notre méthodologie est basée sur la maximisation du coefficient Qini. Étant donné que la sélection du modèle correspond à la sélection des variables, la tâche est difficile si elle est effectuée de manière directe lorsque le nombre de variables à prendre en compte est grand. Pour rechercher de manière réaliste un bon modèle, nous avons conçu une méthode de recherche basée sur une exploration efficace de l'espace des coefficients de régression combinée à une pénalisation de type lasso de la log-vraisemblance. Il n'y a pas d'expression analytique explicite pour la surface Qini, donc la dévoiler n'est pas facile. Notre idée est de découvrir progressivement la surface Qini comparable à l'optimisation sans dérivée. Le but est de trouver un maximum local raisonnable du Qini en explorant la surface près des valeurs optimales des coefficients pénalisés. Nous partageons ouvertement nos codes à travers la librairie R tools4uplift. Bien qu'il existe des méthodes de calcul disponibles pour la modélisation uplift, la plupart d'entre elles excluent les modèles de régression statistique. Notre librairie entend combler cette lacune. Cette librairie comprend des outils pour: i) la discrétisation, ii) la visualisation, iii) la sélection de variables, iv) l'estimation des paramètres et v) la validation du modèle. Cette librairie permet aux praticiens d'utiliser nos méthodes avec aise et de se référer aux articles méthodologiques afin de lire les détails. L'uplift est un cas particulier d'inférence causale. L'inférence causale essaie de répondre à des questions telle que « Quel serait le résultat si nous donnions à ce patient un traitement A au lieu du traitement B? ». La réponse à cette question est ensuite utilisée comme prédiction pour un nouveau patient. Dans la deuxième partie de la thèse, c’est sur la prédiction que nous avons davantage insisté. La plupart des approches existantes sont des adaptations de forêts aléatoires pour le cas de l'uplift. Plusieurs critères de segmentation ont été proposés dans la littérature, tous reposant sur la maximisation de l'hétérogénéité. Cependant, dans la pratique, ces approches sont sujettes au sur-ajustement. Nous apportons une nouvelle vision pour améliorer la prédiction de l'uplift. Nous proposons une nouvelle fonction de perte définie en tirant parti d'un lien avec l'interprétation bayésienne du risque relatif. Notre solution est développée pour une architecture de réseau de neurones jumeaux spécifique permettant d'optimiser conjointement les probabilités marginales de succès pour les individus traités et non-traités. Nous montrons que ce modèle est une généralisation du modèle d'interaction logistique de l'uplift. Nous modifions également l'algorithme de descente de gradient stochastique pour permettre des solutions parcimonieuses structurées. Cela aide dans une large mesure à ajuster nos modèles uplift. Nous partageons ouvertement nos codes Python pour les praticiens désireux d'utiliser nos algorithmes. Nous avons eu la rare opportunité de collaborer avec l'industrie afin d'avoir accès à des données provenant de campagnes de marketing à grande échelle favorables à l'application de nos méthodes. Nous montrons empiriquement que nos méthodes sont compétitives avec l'état de l'art sur les données réelles ainsi qu'à travers plusieurs scénarios de simulations. / Uplift models deal with cause-and-effect inference for a specific factor, such as a marketing intervention. In practice, these models are built on individual data from randomized experiments. A targeted group contains individuals who are subject to an action; a control group serves for comparison. Uplift modeling is used to order the individuals with respect to the value of a causal effect, e.g., positive, neutral, or negative. First, we propose a new way to perform model selection in uplift regression models. Our methodology is based on the maximization of the Qini coefficient. Because model selection corresponds to variable selection, the task is haunting and intractable if done in a straightforward manner when the number of variables to consider is large. To realistically search for a good model, we conceived a searching method based on an efficient exploration of the regression coefficients space combined with a lasso penalization of the log-likelihood. There is no explicit analytical expression for the Qini surface, so unveiling it is not easy. Our idea is to gradually uncover the Qini surface in a manner inspired by surface response designs. The goal is to find a reasonable local maximum of the Qini by exploring the surface near optimal values of the penalized coefficients. We openly share our codes through the R Package tools4uplift. Though there are some computational methods available for uplift modeling, most of them exclude statistical regression models. Our package intends to fill this gap. This package comprises tools for: i) quantization, ii) visualization, iii) variable selection, iv) parameters estimation and v) model validation. This library allows practitioners to use our methods with ease and to refer to methodological papers in order to read the details. Uplift is a particular case of causal inference. Causal inference tries to answer questions such as ``What would be the result if we gave this patient treatment A instead of treatment B?" . The answer to this question is then used as a prediction for a new patient. In the second part of the thesis, it is on the prediction that we have placed more emphasis. Most existing approaches are adaptations of random forests for the uplift case. Several split criteria have been proposed in the literature, all relying on maximizing heterogeneity. However, in practice, these approaches are prone to overfitting. In this work, we bring a new vision to uplift modeling. We propose a new loss function defined by leveraging a connection with the Bayesian interpretation of the relative risk. Our solution is developed for a specific twin neural network architecture allowing to jointly optimize the marginal probabilities of success for treated and control individuals. We show that this model is a generalization of the uplift logistic interaction model. We modify the stochastic gradient descent algorithm to allow for structured sparse solutions. This helps fitting our uplift models to a great extent. We openly share our Python codes for practitioners wishing to use our algorithms. We had the rare opportunity to collaborate with industry to get access to data from large-scale marketing campaigns favorable to the application of our methods. We show empirically that our methods are competitive with the state of the art on real data and through several simulation setting scenarios.