Search

This article describes the preparation of a clinical study on a connected medical device dedicated to facilitate the prevention of lymphedema for patients treated for breast cancer after axillary lymph node dissection (ALND), and to alert them of any manifestation or resurgence in order to prevent its aggravation. A simulator of the entire physical process called digital twin has been developed for this purpose, in a hardware-in-the-loop framework. Statistical calibration of its input parameters, by stochastic inversion and using sensitivity studies, led to establish one or more measurement protocols allowing to capture the signal on a mobile device (phone or tablet) and to detect signal breaks that are physically significant. The measured signal makes it possible to report quickly on the worsening of the patient’s condition and to warn the therapists within a very reasonable period of time. The general methodology of this work seems to us to be easily reproducible in the preparation of clinical studies of other types for connected devices, which aim to develop measurement protocols limiting the often significant cost of such studies. One of the immediate prospects of this work is the initiation of a clinical study on different patients who have been treated by surgery for breast cancer, after improving the robustness of the design of the prototype to take into account the uncertainties affecting the measurements.

Uncertain information on input parameters of computer models is usually modeled by considering these parameters as random, and described by marginal distributions and a dependence structure of these variables. In numerous real-world applications, while information is mainly provided by marginal distributions, typically from samples, little is really known on the dependence structure itself. Faced with this problem of incomplete or missing information, risk studies that make use of these computer models are often conducted by considering independence of input variables, at the risk of including irrelevant situations. This approach is especially used when reliability functions are considered as black-box models. Such analyses remain weakened in absence of in-depth model exploration, at the possible price of a strong risk misestimation. Considering the frequent case where the reliability output is a quantile, this article provides a methodology to improve risk assessment, by exploring a set of pessimistic dependencies using a copula-based strategy. In dimension greater than two, a greedy algorithm is provided to build input regular vine copulas reaching a minimum quantile to which a reliability admissible limit value can be compared, by selecting pairwise components of sensitive influence on the result. The strategy is tested over toy models and a real industrial case-study. The results highlight that current approaches can provide non-conservative results.

BACKGROUND Asthma is a heterogeneous disease in which the interaction between genetic and environmental factors plays a major role. The significance of blood eosinophil is unclear. The aim of the study was to determine the significance of blood eosinophil count in moderate-to-severe asthmatic children of preschool age and school age. METHODS This was a prospective cross-sectional study performed from 2011 to 2015 including children from the severe asthma molecular phenotype (SAMP) cohort at Trousseau Hospital (Paris, France). We included children with severe and moderate asthma, or severe and moderate recurrent wheeze, aged from 1 to 15 years at the time of exploration. RESULTS We analyzed data from 402 children: 248 of preschool age and 154 of school age. Blood eosinophil count third quartile thresholds were 322 and 600 cells/μL for the preschool- and school-age groups, respectively. In multivariate analysis, a blood eosinophil count over this threshold was associated with elevated total IgE (OR = 5.33, P

Relational query optimisers rely on cost models to choose between different query execution plans. Selectivity estimates are known to be a crucial input to the cost model. In practice, standard selectivity estimation procedures are prone to large errors. This is mostly because they rely on the so-called attribute value independence and join uniformity assumptions. Therefore, multidimensional methods have been proposed to capture dependencies between two or more attributes both within and across relations. However, these methods require a large computational cost which makes them unusable in practice. We propose a method based on Bayesian networks that is able to capture cross-relation attribute value dependencies with little overhead. Our proposal is based on the assumption that dependencies between attributes are preserved when joins are involved. Furthermore, we introduce a parameter for trading between estimation accuracy and computational cost. We validate our work by comparing it with other relevant methods on a large workload derived from the JOB and TPC-DS benchmarks. Our results show that our method is an order of magnitude more efficient than existing methods, whilst maintaining a high level of accuracy.

In this study, we develop Three Markov models which are continuous time-homogeneous Model, time piecewise constant intensities Markov model and semi-Markov model with Weibull distribution as the waiting time distribution to evaluate malaria serology evolution. We consider two-state model describing antibody reactivity defined by immunologists. We discuss in detail the application of these models to identify relationships between malaria control program and serological measurements of malaria transmission

Immunoglobulin-E mediated-cow's milk allergy (IgE-CMA) is the one of the most frequent IgE mediated-food allergy in children in industrialized countries and may cause life- threatening anaphylaxis. The natural history of CM IgE mediated food allergy is favorable for most children. However, the clinical outcome of CMA seems worse than previously described, with less chance of recovery after the age of 31. This article is protected by copyright. All rights reserved.

International audience; Allergic asthma is an inflammatory disease of the airways, resulting from an inappropriate immune response to inhaled allergens, which involves multiple inflammatory cells and mediators that contribute to airway hyper-responsiveness. Omalizumab is a monoclonal antibody, specifically directed to the C-Epsilon3 domain of immunoglobulin E (IgE), indicated as an additional therapy in severe allergic asthma [1]. Impaired function of regulatory T cells (Treg) is incriminated in the aberrant asthma inflammatory response [2].

Introduction L’omalizumab est une biotherapie indiquee chez l’enfant presentant un asthme allergique severe. Il est donc important de definir plus clairement la population cible qui peut beneficier au mieux de cette biotherapie. Objectifs Le but de cette etude est de definir les facteurs predictifs de reponse complete a l’omalizumab a 4 mois de traitement. Methodes Nous avons inclus retrospectivement des enfants presentant un asthme severe ou refractaire traite par omalizumab. La reponse complete a l’omalizumab etait evaluee a 4 mois lors d’une reunion multidisciplinaire (en tenant compte de l’ACT, des exacerbations severes, des EFR) et classee en reponse partielle, complete ou non reponse. Resultats Pour les 45 inclus dans l’etude, 22 patients (49 %) avaient une reponse complete, 20 (44 %) une reponse partielle, et 3 (7 %) etaient non repondeurs. Dans l’analyse par regression logistique la presence d’un eczema et un VEMS eleve etaient des facteurs predictifs de reponse complete a l’omalizumab a 4 mois (en comparaison aux non-repondeurs ou aux repondeurs partiels) avec un OR de 1,50 (IC 95 % [1,09 ; 1,9], p = 0,005), et 1,02 (IC 95 % [1,01 ; 1,20], p = 0,001), respectivement. Nous avons identifie 3 clusters independants de reponse a l’omalizumab. Cluster 1, 16 patients avec un asthme eosinophilique (eosinophilie > 300/mm3, n = 15 (94 %), p = 0,001), VEMS moyen 65,4 ± 10,2, (p = 0,006) avec une faible reponse a l’omalizumab (n = 1 (6 %), p Conclusion De multiples morbidites allergiques (dont l’eczema) avec des sensibilisations allergiques multiples et une fonction pulmonaire normale s’associent a une reponse precoce a l’omalizumab.

Background Evaluation of local Plasmodium falciparum malaria transmission has been investigated previously using the reversible catalytic model based on prevalence of antibody responses to single antigen to estimate seroconversion rates. High correlations were observed between seroconversion rates and entomological inoculation rates (EIR). However, in this model, the effects of malaria control interventions and clinical episodes on serological measurements were not assessed. This study monitors the use of antibody responses to P. falciparum crude extracts for assessing malaria transmission, compares seroconversion rates estimated from longitudinal data to those derived from cross-sectional surveys and investigates the effects of malaria control interventions on these measures in an area of declining malaria transmission. In addition, the validity of this model was evaluated by comparison with the alternative model. Methods Five cross-sectional surveys were carried out at the end of the wet season in Dielmo, a malaria-endemic Senegalese rural area in 2000, 2002, 2008, 2010 and 2012. Antibodies against schizonts crude extract of a local P. falciparum strain adapted to culture (Pf 07/03) were measured by ELISA. Age-specific seroprevalence model was used both for cross-sectional surveys and longitudinal data (combined data of all surveys). Results A total of 1504 plasma samples obtained through several years follow-up of 350 subjects was used in this study. Seroconversion rates based on P. falciparum schizonts crude extract were estimated for each cross-sectional survey and were found strongly correlated with EIR. High variability between SCRs from cross-sectional and longitudinal surveys was observed. In longitudinal studies, the alternative catalytic reversible model adjusted better with serological data than the catalytic model. Clinical malaria attacks and malaria control interventions were found to have significant effect on seroconversion. Discussion The results of the study suggested that crude extract was a good serological tool that could be used to assess the level of malaria exposure in areas where malaria transmission is declining. However, additional parameters such as clinical malaria and malaria control interventions must be taken into account for determining serological measurements for more accuracy in transmission assessment. Electronic supplementary material The online version of this article (doi:10.1186/s12936-017-2052-0) contains supplementary material, which is available to authorized users.

Background Childhood recurrent wheezing and consequently asthma corresponds to various phenotypes. Our aim was to link genetic variants of asthma candidate genes to the phenotypes of early onset wheezing. Study design We included very young consecutive children presenting with recurrent wheezing who had been evaluated for the severity of wheezing, associated atopic comorbidities, and tested for biomarkers of atopy and inflammation. All were genotyped for 16 single nucleotide polymorphisms (SNPs) linked with asthma or atopy. An unsupervised hierarchical bottom-up method was used for clustering the phenotypes and a multinomial logistic regression was performed for each individual SNP. Results We replicated the three phenotypes previously described Trousseau Asthma Program in 317 children aged 21.5 ± 7.9 months: cluster 1 (nonatopic uncontrolled severe wheeze), n = 207, a severe viral-induced wheeze, cluster 2 (atopic multiple trigger wheeze), n = 61, with multiple allergic comorbidities, and cluster 3 (episodic viral wheeze), n = 49, a mild viral-induced wheeze. The TT-genotype of the IL-4 rs2070874 polymorphism was significantly associated with the nonatopic uncontrolled severe wheeze compared to the episodic viral wheeze (OR 7.9; CI95% [2.5-25.3]; P = 0.001). Conclusion Association between the TT-genotype of IL-4 rs2070874 polymorphism and a severe phenotype of viral-induced wheeze further underlines the role IL-4 plays in the inflammation pathway leading to viral respiratory infections.

28 pages; This paper is about variable selection with random forests algorithm in presence of correlated predictors. In high-dimensional regression or classification frameworks, variable selection is a difficult task, that becomes even more challenging in the presence of highly correlated predictors. Firstly we provide a theoretical study of the permutation importance measure for an additive regression model. This allows us to describe how the correlation between predictors impacts the permutation importance. Our results motivate the use of the Recursive Feature Elimination (RFE) algorithm for variable selection in this context. This algorithm recursively eliminates the variables using permutation importance measure as a ranking criterion. Next various simulation experiments illustrate the efficiency of the RFE algorithm for selecting a small number of variables together with a good prediction error. Finally, this selection algorithm is tested on a real life dataset from aviation safety where the flight data recorders are analysed for the prediction of a dangerous event.

Introduction L’asthme allergique est une maladie inflammatoire des bronches liee a une reponse immune inappropriee secondaire a l’inhalation de pneumallergenes. Un deficit en lymphocytes T regulateurs (Tregs) pourrait etre responsable de cette reponse immune aberrante. L’omalizumab est un anticorps monoclonal dirige contre le domaine C-Epsilon3 des IgE, utilise en traitement additionnel dans l’asthme severe allergique. L’objectif de notre etude est d’evaluer le taux de Treg circulants (evalue par marquage FoxP3) comme marqueur d’efficacite de l’omalizumab en relation avec le controle de l’asthme. Methodes Nous avons realise une etude longitudinale, prospective, monocentrique chez des enfants atteints d’asthme severe allergique, non controle sous traitement maximal, traites par omalizumab. Nous avons evalue avant traitement puis apres 16 semaines : le controle de l’asthme, les EFR, le taux sanguin de polynucleaires eosinophiles et de Treg evalues par FoxP3. Resultats Vingt et un patients ont ete inclus. Le taux de patients avec un asthme controle ou partiellement controle augmentait a 16 semaines de 0 % a 42,9 % et 52,4 % respectivement. Le pourcentage de Treg augmentait significativement de 4,6 ± 1,6 % a 5,7 ± 1,9 % a 16 semaines (p = 0,02) et le taux de polynucleaires eosinophiles diminuait significativement de 460 ± 316 a 320 ± 200/mm3 (p = 0,02). L’evolution du controle de « non controle » a « controle » a 16 semaines etait expliquee par la variable Treg (p Conclusion Dans notre population, le traitement par omalizumab ameliore le controle de l’asthme. L’amelioration du controle est associee a une augmentation significative des Treg. Le taux de Treg parait associe au controle de l’asthme et pourrait representer un biomarqueur d’efficacite de l’omalizumab.

International audience; BackgroundLittle is known about inflammatory pathways of severe recurrent wheeze in preschool children and severe asthma in children.ObjectivesThe aim of the Severe Asthma Molecular Phenotype cohort was to characterize phenotypes of severe recurrent wheeze and severe asthma during childhood in terms of triggers (allergic or not), involved cells (eosinophil or neutrophil), and corticoid responsiveness.MethodsChildren with moderate-to-severe asthma and preschool children with moderate-to-severe recurrent wheeze were enrolled prospectively. They underwent standardized clinical and blood workup, and bronchoalveolar lavage (BAL) evaluation. Cluster analysis was applied to 350 children with 34 variables.ResultsThree clusters were identified: cluster 1, Neutrophilic steroid-refractory recurrent wheeze phenotype, with 138 children uncontrolled despite high-dose inhaled corticosteroids (ICS) (92%, P < .001), with more history of pneumonia (31%, P < .001), more gastroesophageal reflux disease (37%, P < .001), and the highest blood neutrophil count (mean 4.524 cells/mm3, P = .05); cluster 2, Severe recurrent wheeze with sensitization to a single aeroallergen (12%, P = .002), with 104 children controlled with high-dose ICS (63%, P < .001); cluster 3, Eosinophilic steroid-refractory asthma phenotype, with 108 children uncontrolled despite high-dose ICS (76%, P < .001) with more allergic rhinitis, atopic dermatitis, and food allergies (82%, 40%, 31%, P < .001, respectively). They also had a higher blood eosinophil count and a higher percentage of BAL eosinophil (506/mm3, 2.6%, P < .001 respectively).ConclusionsInflammation pathway of asthma and recurrent wheeze are related to eosinophil cells in older children and neutrophil cells in younger children. These results could improve personalized treatments.

Les modeles multi-etats permettent d’etudier l’evolution dans le temps d’un phenomene observe a differents moments. Le processus associe est souvent Markovien afin de resumer le passe. Ces modeles sont appliques avec succes dans differents domaines. Les modeles semi-Markoviens sont une alternative interessante puisque l’evolution du processus depend du temps passe dans un etat. En effet, la distribution des temps de sejour est exponentielle dans les modeles Markoviens alors que c’est un parametre du modele semi-Markovien. Une revue des principales methodes d’estimation dans ces modeles est presentee. Les modeles Markoviens homogenes avec des intensites constantes et les modeles Markoviens non homogenes avec des intensites dependantes du temps sont presentes. Une methode d’estimation parametrique dans les modeles semi-Markoviens est ensuite proposee. La methode du maximum de vraisemblance permet d’estimer les parametres des distributions des temps de sejours. On discute ensuite l’utilisation des modeles multi-etats dans les analyses cout-efficacite. En effet, un modele de cout peut aisement etre ajuste pour evaluer les couts et l’efficacite d’une intervention. L’application de ces modeles au cas de l’asthme permet d’illustrer le potentiel de ces methodes.

Introduction La securite et l’efficacite de l’immunotherapie allergenique sous-cutanee (ITSC) dans l’asthme reste debattue. Nous avons evalue la securite et l’efficacite de l’ITSC sur l’asthme et les comorbidites allergiques chez des enfants atteints d’asthme allergique modere a severe. Methodes Nous avons collecte de facon retrospective des donnees sur les effets indesirables, le nombre annuel d’exacerbations severes, la dose de corticosteroides inhales (CSI) et l’activite des comorbidites allergiques, dans une cohorte d’enfants beneficiant d’une ITSC acceleree envers Dermatophagoides pteronyssinus. Resultats Soixante-douze enfants, ayant en majorite un asthme severe (56 %), ont ete inclus Les reactions systemiques severes representaient 0,3 % des 3455 injections realisees. L’utilisation d’adrenaline a ete necessaire pour un patient. Le taux d’exacerbations de l’asthme diminuait de facon significative apres 1 an d’ITSC (p Discussion Les donnees sur la securite de l’ITSC dans cette cohorte de patients atteints d’asthme persistant modere a severe presentent un interet pour la prise en charge de ce type de malades. Une des principales limites de cette etude est sa methodologie retrospective, sans groupe temoin. Conclusion Cette etude montre un taux faible de RSS au vu du nombre total d’injections realisees et une efficacite significative de l’ITSC sur l’asthme, particulierement en cas de phenotype exacerbateur.

International audience; BackgroundAtopic dermatitis (AD) is known to predate asthma and other atopic disorders described under the term “atopic march”. However, this classic sequence is not always present and only a few studies have addressed children at risk of developing asthma. The objective of this study is to define early-onset AD phenotypes leading to asthma.MethodsWe performed a cluster analysis with 9 variables of 214 infants with early-onset AD prospectively enrolled in the ORCA cohort and followed each year on the occurrence of asthma until the age of 6.ResultsWe identified 3 clusters - cluster 1 (n = 94) with low to no sensitization to food (27.7%) or aeroallergens (10.6%) and moderate AD severity (SCORAD 25.29 +/- 14.6) called “AD with low sensitization”; - cluster 2 (n = 84) characterized by a higher AD severity (SCORAD 32.66+/-16.6) and frequent sensitization to food (98.9%) or aeroallergens (26.2%), most likely multiple (96.4% for food allergens), called “AD with multiple sensitizations” - cluster 3 (n = 36) with parental history, moderate AD severity (SCORAD 24.46+/-15.7), moderate rate of sensitization to food allergens (38.9%) (exclusively single) with no sensitization to aeroallergens, called “AD with familial history of asthma”. Percentages of children suffering from asthma at the age of 6 were higher in clusters 2 and 3 (36.1% and 33.3% respectively versus 14.9% in cluster 1, p

The selection of grouped variables using the random forest algorithm is considered. First a new importance measure adapted for groups of variables is proposed. Theoretical insights into this criterion are given for additive regression models. Second, an original method for selecting functional variables based on the grouped variable importance measure is developed. Using a wavelet basis, it is proposed to regroup all of the wavelet coefficients for a given functional variable and use a wrapper selection algorithm with these groups. Various other groupings which take advantage of the frequency and time localization of the wavelet basis are proposed. An extensive simulation study is performed to illustrate the use of the grouped importance measure in this context. The method is applied to a real life problem coming from aviation safety.

In studies of disease states and their relation to evolution, data on the state are usually obtained at in frequent time points during follow-up. Moreover in many applications, there are measured covariates on each individual under study and interest centres on the relationship between these covariates and the disease evolution. We developed a continuous-time Markov model with use of time-dependent covariates and a Markov model with piecewise constant intensities to model asthma evolution. Methods to estimate the effect of covariates on transition intensities, to test the assumption of time homogeneity and to assess goodness-of-fit are proposed. We apply these methods to asthma control. We consider a three-state model and we discuss in detail the analysis of asthma control evolution. Copyright © 2003 John Wiley & Sons, Ltd.

International audience

Background Rules for predicting the course of asthma in wheezy infants have low specificity. Objective To determine if the novel phenotypes—mild early viral wheeze (EVW), atopic multiple-trigger wheeze (MTW), and nonatopic uncontrolled wheeze (NAUW)—have different courses during the preschool period. Methods Part of the prospectively followed Trousseau Asthma Program cohort was phenotyped using cluster analysis with 12 parameters (sex, asthma severity and control with inhaled corticosteroid [ICS], parental asthma, allergic rhinitis, eczema, food allergy, EVW or MTW, and allergen exposure trigger). Wheezing trajectories were assessed by crossing the original phenotypes with the phenotypes obtained at 5 years. Results Four clusters were identified at 5 years of age: asymptomatic children (n = 47) with no wheezing (98%), children with mild EVW (n = 40, 87% with EVW, 50% with EVW controlled with low-dose ICS), those with atopic MTW (n = 30, 100% with MTW, only 17% with MTW controlled with low-dose ICS, more significant for pollen asthmatic trigger), and those with atopic severe UW (n = 33, 63% with UW uncontrolled despite high doses of ICS, more significant for allergic rhinitis and dust as asthmatic trigger). Those with mild EVW became asymptomatic or remained with mild EVW. Those with atopic MTW remained with atopic MTW and those with NAUW developed severe UW in most cases. Conclusion These results show that remission is most frequently observed in mild EVW and that no remission is observed in atopic MTW.

Objective IgE-mediated allergic asthma phenotype appears to be heterogeneous. We set out to define distinct allergic phenotypes by unsupervised cluster analysis. Study design A total of 18 variables were analyzed: sex and age, eczema and food allergy, asthma duration, asthma severity and control, severe exacerbations, total IgE level, allergic sensitization, fractional exhaled nitric oxide, and functional parameters. Clusters obtained were cross-tabulated with environmental parameters. Results Four clusters were identified in 125 children (average age 8.9 years): (1) 57 children constituted the “House dust mite Sensitization and Mild Asthma” cluster, 98% of these were monosensitized and had mild asthma (74%); (2) 12 children had “Pollen Sensitization with Severe Exacerbations,” 92 % with severe exacerbations and pollen sensitization; (3) 20 children had “Multiple Allergies and Severe Asthma,” with 95% having moderate to severe asthma, and a significantly decreased forced expiratory flow rate at 25%-75% of forced vital capacity, 100% had eczema and higher values of IgE (1123 kU/L) and fractional exhaled nitric oxide (67 ppb) (this cluster was associated with molds at home [P = .004]); and (4) 36 children had “Multiple Allergic Sensitizations and Mild Asthma,” 97% of these with multiple sensitizations and 100% mild asthma. Conclusions The identification of 2 novel severe allergic asthma phenotypes “Pollen Sensitization with Severe Exacerbations”and “Multiple Allergies and Severe Asthma” could lead to specific targeted treatment.

Background Recurrent wheezing during infancy is a heterogeneous disorder that has been associated with early-onset asthma. Objective To identify phenotypes of severe recurrent wheezing and therapeutic approaches. Methods We performed cluster analysis with 20 variables of 551 children with active asthma, younger than 36 months old, and enrolled in the Trousseau Asthma Program. Results We identified 3 independent clusters of children with wheezing. Cluster 1, mild episodic viral wheeze (n = 327), consisted of children with wheezing related only to colds (71%), mild disease (76%), and mainly normal chest x-ray results. Cluster 2, nonatopic uncontrolled wheeze (n = 157), was characterized by moderate to severe disease (91%), uncontrolled wheezing despite high doses of inhaled corticosteroids (55%), parents with asthma, and increased levels of ferritine. Cluster 3, atopic multiple-trigger wheeze (n = 67), included more children with multiple-trigger wheeze (68%) than did clusters 1 or 2; eczema (75%); a positive result from the Phadiatop Infant test (90%); increased levels of IgE, IgA, and IgG; and abnormal results from chest x-rays. In separate analysis, 1 parameter for boys (increased total level of IgE) and 2 parameters for girls (wheezing severity and increased total level of IgE) properly classified 90% of boys and 83% of girls in the appropriate cluster. Significant associations were found between overcrowding, molds and cockroaches at home, and atopic multiple-trigger wheeze and between day-care attendance and nonatopic uncontrolled wheeze in other parts. Conclusion We identified different phenotypes of recurrent wheezing in young children by using cluster analysis with usual variables. These phenotypes require confirmation in longer, follow-up studies.

In this paper, we provide a nonparametric estimator of the distribution of bivariate censored lifetimes, in a model where the two censoring variables differ only through an additional observed variable. This situation is motivated by a particular application to insurance, where the supplementary variable corresponds to the age difference between two individuals. Asymptotic results for our estimator are provided. The new tools that we develop are used to perform goodness-of-fit tests for survival copula models. The practical performance is illustrated through simulations and a real data analysis.

In this paper we study a class of M-estimators in a regression model under bivariate random censoring and provide a set of sufficient conditions that ensure asymptotic n 1 / 2 - convergence . The cornerstone of our approach is a new estimator of the joint distribution function of the censored lifetimes. A copula approach is used to modelize the dependence structure between the bivariate censoring times. The resulting estimators present the advantage of being easily computable. A simulation study enlighten the finite sample behavior of this technique.

Multi-state stochastic models are useful tools for studying complex dynamics such as chronic diseases. Semi-Markov models explicitly define distributions of waiting times, giving an extension of continuous time and homogeneous Markov models based implicitly on exponential distributions. This paper develops a parametric model adapted to complex medical processes. (i) We introduced a hazard function of waiting times with a U or inverse U shape. (ii) These distributions were specifically selected for each transition. (iii) The vector of covariates was also selected for each transition. We applied this method to the evolution of HIV infected patients. We used a sample of 1244 patients followed up at the hospital in Nice, France.

Control and severity of asthma are two different but complementary concepts. The severity of asthma could influence the control over time. The aim of this study was to demonstrate this relationship. A total 365 patients with persistent asthma (severity) were enrolled and followed-up prospectively. Data were analysed using a continuous time homogeneous Markov model of the natural history of asthma. Control of asthma was defined according to three health states which were qualified: optimal, suboptimal and unacceptable control (states 1, 2 and 3). Transition forces (denoted λij from state i to state j) and transition probabilities between control states were assessed and the results stratified by asthma severity were compared. Models were validated by comparing expected and observed numbers of patients in the different states. Transition probabilities stabilised between 100–250 days and more rapidly in patients with mild-to-moderate asthma. Patients with mild-to-moderate asthma in suboptimal or unacceptable control had a high probability of transition directly to optimal control. Patients with severe asthma had a tendency to remain in unacceptable control. A Markov model is a useful tool to model the control of asthma over time. Severity modified clearly the health states. It could be used to compare the performance of different approaches to asthma management. This study was supported by an unrestricted educational grant from GlaxoSmithKline, France.

International audience; Background: In the preschool period, allergic rhinitis (AR) is infrequent and thus under-diagnosed. However, recent works have highlighted the occurrence of AR in toddlers although the causes of AR in this young population remain unknown. The objective of this study was to identify determinants of AR in young children with asthma. Methods: We carried out a case-control study of 227 children with active asthma and enrolled in the Trousseau Asthma Program. AR and other allergic diseases (asthma, food allergy and eczema) were diagnosed by medical doctors using standardized questionnaires. Parental history of AR and asthma, biological markers of atopy (total IgE, blood eosinophilia, allergic sensitization towards food and aeroallergens) and environmental parameters were also collected. Results: Forty one of the children (18.1%) had AR. By univariate logistic regression analysis, AR was mainly associated with peanut sensitization (OR = 6.75; p = 0.002); food allergy (OR = 4.31; p = 0.026); mold exposure (OR = 3.81 p,0.01) and parental history of AR (OR = 1.42; p = 0.046). Due to the strong link between food allergy and peanut sensitization three models of multivariate logistic regression were performed and confirmed that AR is associated with peanut sensitization but also food allergy and mold exposure. A random forest analysis was also performed to explain AR. The results reinforced the logistic analysis that peanut sensitization and mold exposure were the principal determinants of AR. Conclusions & Clinical Relevance: These results stress the importance of investigating AR in young children with asthma to potentially diagnose a particularly severe allergic asthmatic phenotype. Moreover, these data evoke the hypothesis that peanut could be an aeroallergen.

International audience

The relationship between asthma and obesity appears to be quite complex. The aim of this study was to assess the effect of excess weight on asthma control evolution in a cohort of asthmatics. A prospective database was set up, which enrolled adult asthmatics with persistent (mild, moderate or severe) asthma. The control of asthma was defined as a binary variable, acceptable or unacceptable. In order to evaluate the effect of body mass index (BMI; or =25), data were analysed using a continuous time homogeneous Markov model in which the forces ruling the transition between the two health states were estimated. The following confounding covariates were also evaluated in the model: severity of asthma, current treatment with oral corticosteroids (OCS) and history of OCS over the year preceding inclusion. About 406 asthmatics were included who made a total of 1639 consultations; the median length of follow up was 182 days. Using a univariate model, overweight patients had a lower risk of transiting from the unacceptable to the acceptable health state (RR = 0.45; P < 0.01). The effect of weight remained significant (RR = 0.53; P < 0.01) in the multivariate model including the other covariates. Moreover, transition probabilities stabilized more rapidly for patients with BMI < 25 (200 vs 300 days). In this study, we thus demonstrated that there is an association between excess weight and transition from unacceptable to acceptable control. Because control of asthma clearly drives asthma management, this finding has consequences for defining original new strategies for managing asthma in overweight patients.

International audience

Databases, and in particular relational databases, are a common paradigm for storing and querying data. Data is stored in relations which are linked with each other. Users may interrogate the database by issuing queries via a declarative query language. One of the challenges for a relational database management system (RDBMS) is to evaluate each query as quickly possible. However, as it turns out, there are many ways in which a query can be executed. The RDBMS delegates the choice of the query execution plan to a query optimiser. In turn, the query optimiser tasks a cost model with the responsibility of estimating the cost of each execution plan considered by the optimiser. The cost estimation therefore has a large influence on the quality of the execution plans, which in turns affects the running time that the user perceives. The most important input in the cost model is the selectivity, which is the amount of data that is expected to flow in and out of each operator in a plan. From a statistical point of view, this is a case of density estimation, where the goal is to determine the amount of data that verifies certain conditions. It is common practice to use density estimation models that make strong simplifying assumptions, in large part for efficiency reasons. For instance, it is usually assumed that attributes are independent with each other. Alas, in practice, these assumptions are more often that not unverified, which leads to large estimation errors. The goal of this PhD is thus to propose models that offer a better compromise between selectivity estimation accuracy and computational cost. First of all, we propose a methodology based on Bayesian networks, where we constrain each network to have a tree topology, therefore allowing a linear computational complexity. Secondly, we train online machine learning models to correct an existing selectivity estimation model, whilst adapting to concept drift. Our experimental results, which are based on the TPC-DS and JOB benchmarks, are very encouraging with respect to both of our lines of work., Les bases de données relationnelles sont couramment utilisées pour stocker et interroger des données. Les données y sont stockées dans des relations qui sont liées entre elles. Un utilisateur interroge ces données via des requêtes exprimées dans un langage déclaratif. Un des défis pour le Système de Gestion de Bases de Données Relationnelles (SGBDR) est d’évaluer les requêtes le plus rapidement possible. Pour cette évaluation, il existe un large choix de plans d’exécution associé à une requête donnée. Le SGBDR délègue le choix du plan d’exécution le plus efficace à un optimiseur. Dans ce processus, le modèle de coûts, quand à lui, a la responsabilité d’estimer le temps de calcul de chaque plan d’exécution considéré par l’optimiseur. Par conséquent, le modèle de coûts influe sur la qualité des plans d’exécution produit par l’optimiseur, et donc sur le temps de réponse perçu par l’utilisateur.Le paramètre le plus influant dans le modèle de coûts est la sélectivité, qui permet d’estimer la quantité de données qui transite entre chaque opérateur d’un plan d’exécution. D’un point de vue statistique, ceci correspond à de l’estimation de densité, qui consiste à déterminer la quantité de données qui vérifie un ensemble de conditions. Il est d’usage courant d’utiliser des modèles avec des hypothèses simplificatrices pour des raisons de performances. Par exemple, il est souvent supposés qu’il n’existe pas de dépendances de valeurs entre différents attributs de relations. Dans la majeure partie des applications cette hypothèse se révèle non vérifiée et conduit à de grandes erreurs d’estimation. Le but de cette thèse est donc de proposer des modèles plus réalistes qui offrent un meilleur compromis entre la précision des estimations et la complexité calculatoire requise. Premièrement, nous proposons une méthode s’appuyant sur les réseaux Bayésiens dont la représentation sous forme d’arbre nous permet d’améliorer la précision des estimations avec une complexité linéaire. Par la suite, nous explorons l’usage de modèles d’apprentissage en ligne pour corriger un modèle d’estimation existant tout en s’adaptant à l’évolution des données. Nos résultats expérimentaux basés sur les bechmaks JOB et TPC-DS montrent des résultats très encourageants.