MDMatthieu Doutreligne

Matthieu Doutreligne

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Matthieu Doutreligne has a total of 37 co-authors. Here are the top co-authors by number of shared publications:

  • Claire Morgand

    Affiliation Agence Régionale de Santé Ile-de-France, Saint-Denis, France

    3 shared publications
  • Gaël Varoquaux

    Affiliation Soda Team, Inria Saclay, Palaiseau, France

    3 shared publications
  • Judith Abecassis

    Affiliation Soda Team, Inria Saclay, Palaiseau, France

    3 shared publications
  • Tristan Struja

    Affiliations Laboratory for Computational Physiology, Massachusetts Institute of Technology, Cambridge, Massachusetts, United States of America, Medical University Clinic, Division of Endocrinology, Diabetes & Metabolism, Kantonsspital Aarau, Aarau, Switzerland

    3 shared publications
  • Antoine Lamer

    University of Lille, Regional Federation for Mental Health and Psychiatry Research Hauts-de-France, ULR 2694 METRICS , Lille

    2 shared publications
  • Benjamin Popoff

    Department of Digital Health, Department of Anesthesiology, Critical Care and Perioperative Medicine, CHU Rouen, Rouen Cedex, France

    2 shared publications

Latest Publications

  • Representations and inference from time-varying routine care data

    by Matthieu Doutreligne

    Représentations et inférence à partir de données de santé temporelles collectées en routine: Les bases de données de vie réelle sont de plus en plus accessibles, exhaustives, avec des détails temporels précis. Contrairement aux données utilisées dans la recherche clinique traditionnelle, elles capturent l'organisation routinière des soins. Ces données de soins quotidiens ouvrent la porte à de nouvelles questions de recherche, notamment en ce qui concerne la qualité des soins, l'efficacité des interventions après leur mise sur le marché, l'hétérogénéité de leurs bénéfices dans les populations mal desservies ou le développement de traitements personnalisés. D'un autre côté, la complexité et la nature à grande échelle de ces bases de données posent un certain nombre de défis pour leur utilisation efficace. Pour remédier à ces problèmes, les économètres et les épidémiologistes ont récemment proposé l'utilisation de modèles flexibles combinant l'inférence causale et l'apprentissage automatique en grande dimension. Dans un premier temps, nous illustrons par trois exemples la tension actuelle entre ces nouvelles sources de données, l'apprentissage automatique et des problématiques modernes de santé publique. Ces exemples motivent notre principale question de recherche : Comment des modèles flexibles peuvent-ils aider à fournir un traitement approprié à chaque patient afin d'améliorer sa santé ? Afin de mieux comprendre les infrastructures modernes de collecte et d'analyse des dossiers patients informatisés (DPI), nous faisons la synthèse d'entretiens semi-structurés menés dans le cadre d'une étude de cas nationale portant sur les entrepôts de données cliniques des 32 hôpitaux régionaux et universitaires français. Reconnaissant la difficulté d'accéder à des échantillons de grande taille et à la puissance de calcul pour développer des modèles prédictifs généralisables, nous étudions un gradient de complexité dans les représentations et les algorithmes prédictifs sur DPI. En se tournant vers le cadre causal, nous détaillons ensuite les éléments clés nécessaires pour estimer de manière robuste l'effet du traitement à partir de données de DPI variant dans le temps. Nous documentons l'impact de différents choix méthodologiques pour l'étude de l'effet de l'albumine sur la mortalité dans des cas de septicémie avec la base de données MIMIC-IV (Medical Information Mart for Intensive Care). Dans des problèmes à grande dimension, la sélection d'hyperparamètres pour les modèles causaux est cruciale pour éviter le sous-apprentissage ou le sur-apprentissage. Pour une simulation et trois ensembles de données semi-simulées, nous montrons que le risque R doublement robuste surpasse d'autres risques causaux existants.

  • Open Access

    Step-by-step causal analysis of EHRs to ground decision-making

    by Matthieu Doutreligne, Tristan Struja, Judith Abecassis, Claire Morgand, Leo Anthony Celi, Gaël Varoquaux

    Causal inference enables machine learning methods to estimate treatment effects of medical interventions from electronic health records (EHRs). The prevalence of such observational data and the difficulty for randomized controlled trials (RCT) to cover all population/treatment relationships make these methods increasingly attractive for studying causal effects. However, researchers should be wary of many pitfalls. We propose and illustrate a framework for causal inference estimating the effect of albumin on mortality in sepsis using an Intensive Care database (MIMIC-IV) and comparing various sensitivity analyses to results from RCTs as gold-standard. The first step is study design, using the target trial concept and the PICOT framework: Population (patients with sepsis), Intervention (combination of crystalloids and albumin for fluid resuscitation), Control (crystalloids only), Outcome (28-day mortality), Time (intervention start within 24h of admission). We show that too large treatment-initiation times induce immortal time bias. The second step is selection of the confounding variables based on expert knowledge. Increasingly adding confounders enables to recover the RCT results from observational data. As the third step, we assess the influence of multiple models with varying assumptions, showing that a doubly robust estimator (AIPW) with random forests proved to be the most reliable estimator. Results show that these steps are all important for valid causal estimates. A valid causal model can then be used to individualize decision making: subgroup analyses showed that treatment efficacy of albumin was better for patients >60 years old, males, and patients with septic shock. Without causal thinking, machine learning is not enough for optimal clinical decision on an individual patient level. Our step-by-step analytic framework helps avoiding many pitfalls of applying machine learning to EHR data, building models that avoid shortcuts and extract the best decision-making evidence.


  • Open Access

    How to select predictive models for decision-making or causal inference

    by Matthieu Doutreligne, Gaël Varoquaux

    Background: We investigate which procedure selects the most trustworthy predictive model to explain the effect of an intervention and support decision-making. Methods: We study a large variety of model selection procedures in practical settings: finite samples settings and without a theoretical assumption of well-specified models. Beyond standard cross-validation or internal validation procedures, we also study elaborate causal risks. These build proxies of the causal error using “nuisance” reweighting to compute it on the observed data. We evaluate whether empirically estimated nuisances, which are necessarily noisy, add noise to model selection and compare different metrics for causal model selection in an extensive empirical study based on a simulation and 3 health care datasets based on real covariates. Results: Among all metrics, the mean squared error, classically used to evaluate predictive modes, is worse. Reweighting it with a propensity score does not bring much improvement in most cases. On average, the $R\text{-risk}$, which uses as nuisances a model of mean outcome and propensity scores, leads to the best performances. Nuisance corrections are best estimated with flexible estimators such as a super learner. Conclusions: When predictive models are used to explain the effect of an intervention, they must be evaluated with different procedures than standard predictive settings, using the $R\text{-risk}$ from causal inference.


  • Barriers encountered with clinical data warehouses: Recommendations from a focus group

    by Antoine Lamer, Benjamin Popoff, Boris Delange, Matthieu Doutreligne, Emmanuel Chazard, Romaric Marcilly, Sonia Priou, Paul Quindroit


  • Open Access

    Step-by-step causal analysis of Electronic Health Records to ground decision making

    by Matthieu Doutreligne, Tristan Struja, Judith Abecassis, Claire Morgand, Gaël Varoquaux, Leo Anthony Celi

    Accurate predictions, as with machine learning, may not suffice to provide optimal healthcare for every patient. Indeed, prediction can be driven by shortcuts in the data, such as racial biases. Causal thinking is needed for data-driven decisions. Here, we give an introduction to the key elements, focusing on routinely-collected data, electronic health records (EHRs) and claims data. Using such data to assess the value of an intervention requires care: temporal dependencies and existing practices easily confound the causal effect. We present a step-by-step framework to help build valid decision making from real-life patient records by emulating a randomized trial before individualizing decisions, eg with machine learning. Our framework highlights the most important pitfalls and considerations in analysing EHRs or claims data to draw causal conclusions. We illustrate the various choices in studying the effect of albumin on sepsis mortality in the Medical Information Mart for Intensive Care database (MIMIC-IV). We study the impact of various choices at every step, from feature extraction to causal-estimator selection. In a tutorial spirit, the code and the data are openly available.


  • Open Access

    Adjusting for the progressive digitization of health records: working examples on a multi-hospital clinical data warehouse

    by Adam Remaki, Benoît Playe, Paul Bernard, Simon Vittoz, Matthieu Doutreligne, Gilles Chatelier, Etienne Audureau, Emmanuelle Kempf, Raphaël Porcher, Romain Bey

    Objectives To propose a new method to account for time-dependent data missingness caused by the increasing digitization of health records in the analysis of large-scale clinical data. Materials and Methods Following a data-driven approach we modeled the progressive adoption of a common electronic health record in 38 hospitals. To this end, we analyzed data collected between 2013 and 2022 and made available in the clinical data warehouse of the Greater Paris University Hospitals. Depending on the category of data, we worked either at the hospital, department or unit level. We evaluated the performance of this model with a retrospective cohort study. We measured the temporal variations of some quality and epidemiological indicators by successively applying two methods, either a naive analysis or a novel complete-source-only analysis that accounts for digitization-induced missingness. Results Unrealistic temporal variations of quality and epidemiological indicators were observed when a naive analysis was performed, but this effect was either greatly reduced or disappeared when the complete-source-only method was applied. Discussion We demonstrated that a data-driven approach can be used to account for missingness induced by the progressive digitization of health records. This work focused on hospitalization, emergency department and intensive care units records, along with diagnostic codes, discharge prescriptions and consultation reports. Other data categories may require specific modeling of their associated data sources. Conclusions Electronic health records are constantly evolving and new methods should be developed to debias studies that use these unstable data sources.


    • 31 July 2023
    • Published by HAL in HAL
    Open Access

    Causal thinking for decision making on Electronic Health Records: why and how

    by Matthieu Doutreligne, Tristan Struja, Judith Abecassis, Claire Morgand, Anthony Leo, Gaël Celi, Leo Celi, Gaël Varoquaux

    Accurate predictions, as with machine learning, may not suffice to provide optimal healthcare for every patient. Indeed, prediction can be driven by shortcuts in the data, such as racial biases. Causal thinking is needed for data-driven decisions. Here, we give an introduction to the key elements, focusing on routinely-collected data, electronic health records (EHRs) and claims data. Using such data to assess the value of an intervention requires care: temporal dependencies and existing practices easily confound the causal effect. We present a step-by-step framework to help build valid decision making from real-life patient records by emulating a randomized trial before individualizing decisions, eg with machine learning. Our framework highlights the most important pitfalls and considerations in analysing EHRs or claims data to draw causal conclusions. We illustrate the various choices in studying the effect of albumin on sepsis mortality in the Medical Information Mart for Intensive Care database (MIMIC-IV). We study the impact of various choices at every step, from feature extraction to causal-estimator selection. In a tutorial spirit, the code and the data are openly available.

  • Open Access

    Good practices for clinical data warehouse implementation: A case study in France

    by Matthieu Doutreligne, Adeline Degremont, Pierre-Alain Jachiet, Antoine Lamer, Xavier Tannier

    Real-world data (RWD) bears great promises to improve the quality of care. However, specific infrastructures and methodologies are required to derive robust knowledge and brings innovations to the patient. Drawing upon the national case study of the 32 French regional and university hospitals governance, we highlight key aspects of modern clinical data warehouses (CDWs): governance, transparency, types of data, data reuse, technical tools, documentation, and data quality control processes. Semi-structured interviews as well as a review of reported studies on French CDWs were conducted in a semi-structured manner from March to November 2022. Out of 32 regional and university hospitals in France, 14 have a CDW in production, 5 are experimenting, 5 have a prospective CDW project, 8 did not have any CDW project at the time of writing. The implementation of CDW in France dates from 2011 and accelerated in the late 2020. From this case study, we draw some general guidelines for CDWs. The actual orientation of CDWs towards research requires efforts in governance stabilization, standardization of data schema, and development in data quality and data documentation. Particular attention must be paid to the sustainability of the warehouse teams and to the multilevel governance. The transparency of the studies and the tools of transformation of the data must improve to allow successful multicentric data reuses as well as innovations in routine care.


  • Open Access

    Standardized Description of the Feature Extraction Process to Transform Raw Data Into Meaningful Information for Enhancing Data Reuse: Consensus Study

    by Antoine Lamer, Mathilde Fruchart, Nicolas Paris, Benjamin Popoff, Anaïs Payen, Thibaut Balcaen, William Gacquer, Guillaume Bouzillé, Marc Cuggia, Matthieu Doutreligne, et al.

    Background: Despite the many opportunities data reuse offers, its implementation presents many difficulties, and raw data cannot be reused directly. Information is not always directly available in the source database and needs to be computed afterwards with raw data for defining an algorithm. Objective: The main purpose of this article is to present a standardized description of the steps and transformations required during the feature extraction process when conducting retrospective observational studies. A secondary objective is to identify how the features could be stored in the schema of a data warehouse. Methods: This study involved the following 3 main steps: (1) the collection of relevant study cases related to feature extraction and based on the automatic and secondary use of data; (2) the standardized description of raw data, steps, and transformations, which were common to the study cases; and (3) the identification of an appropriate table to store the features in the Observation Medical Outcomes Partnership (OMOP) common data model (CDM). Results: We interviewed 10 researchers from 3 French university hospitals and a national institution, who were involved in 8 retrospective and observational studies. Based on these studies, 2 states (track and feature) and 2 transformations (track definition and track aggregation) emerged. “Track” is a time-dependent signal or period of interest, defined by a statistical unit, a value, and 2 milestones (a start event and an end event). “Feature” is time-independent high-level information with dimensionality identical to the statistical unit of the study, defined by a label and a value. The time dimension has become implicit in the value or name of the variable. We propose the 2 tables “TRACK” and “FEATURE” to store variables obtained in feature extraction and extend the OMOP CDM. Conclusions: We propose a standardized description of the feature extraction process. The process combined the 2 steps of track definition and track aggregation. By dividing the feature extraction into these 2 steps, difficulty was managed during track definition. The standardization of tracks requires great expertise with regard to the data, but allows the application of an infinite number of complex transformations. On the contrary, track aggregation is a very simple operation with a finite number of possibilities. A complete description of these steps could enhance the reproducibility of retrospective studies.