EUVIMEDEUVIMED entdecken
LeitlinieOpen Access · CC BY 4.0

Guidelines and Best Practices for the Use of Targeted Maximum Likelihood and Machine Learning When Estimating Causal Effects of Exposures on Time‐To‐Event Outcomes

Denis Talbot iD, Awa Diop, Miceline Mésidor, Yohann Chiu, Caroline Sirois, Andrew J Spieker iD, Antoine Pariente, Pernelle Noize, Marc Simard, Miguel Angel Luque Fernandez iD, Michael Schomaker, Kenji Fujita und 2 weitere

Statistics in Medicine · 2025

DOI: 10.1002/sim.70034 ↗

Kurz gefasst

Leitlinie in der Zeitschrift Statistics in Medicine, 2025. Verfasst von Denis Talbot und 13 weiteren Autor:innen. Beteiligte Einrichtung: Département de Médecine Sociale et Préventive, Université Laval, Québec, Canada. Hauptthemen laut MEDLINE: Machine Learning, Causality. Die Arbeit wurde bisher 7-mal zitiert (Crossref) und nennt 71 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: International Society for Pharmacoepidemiology, Fonds de Recherche du Québec - Santé, CIHR.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Targeted maximum likelihood estimation (TMLE) is an increasingly popular framework for the estimation of causal effects. It requires modeling both the exposure and outcome but is doubly robust in the sense that it is valid if at least one of these models is correctly specified. In addition, TMLE allows for flexible modeling of both the exposure and outcome with machine learning methods. This provides better control for measured confounders since the model specification automatically adapts to the data, instead of needing to be specified by the analyst a priori . Despite these methodological advantages, TMLE remains less popular than alternatives in part because of its less accessible theory and implementation. While some tutorials have been proposed, none address the case of a time‐to‐event outcome. This tutorial provides a detailed step‐by‐step explanation of the implementation of TMLE for estimating the effect of a point binary or multilevel exposure on a time‐to‐event outcome, modeled as counterfactual survival curves and causal hazard ratios. The tutorial also provides guidelines on how best to use TMLE in practice, including aspects related to study design, choice of covariates, controlling biases and use of machine learning. R‐code is provided to illustrate each step using simulated data ( https://github.com/detal9/SurvTMLE ). To facilitate implementation, a general R function implementing TMLE with options to use machine learning is also provided. The method is illustrated in a real‐data analysis concerning the effectiveness of statins for the prevention of a first cardiovascular disease among older adults in Québec, Canada, between 2013 and 2018.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Machine LearningCausalityHumansLikelihood FunctionsModels, StatisticalProportional Hazards Models

Stichwörter der Autor:innen

causal inferencedouble robustnessmachine learningobservational studiessurvival analysistargeted maximum likelihood estimation

Autor:innen und Institutionen

  • Département de Médecine Sociale et Préventive, Université Laval, Québec, Canada.
  • Axe Santé des Populations et Pratiques Optimales en Santé, Centre de Recherche du CHU de Québec - Université Laval, Québec, Canada.
  • Faculté de Pharmacie, Université Laval, Québec, Canada.
  • Bureau d'information et d'études en santé des populations, Institut national de santé publique du Québec, Québec, Canada.
  • VITAM-Centre de Recherche en Santé Durable, Centre Intégré de Santé et de Services Sociaux de la Capitale Nationale, Québec, Canada.
  • Department of Biostatistics, Vanderbilt University Medical Center, Nashville, Tennessee, USA.

ORCID-Profile: Denis Talbot, Andrew J Spieker, Miguel Angel Luque Fernandez, Mireille E Schnitzer

Förderung

  • International Society for Pharmacoepidemiology
  • Fonds de Recherche du Québec - Santé
  • CIHR
  • Canada Research Chairs
  • Canadian Institutes of Health Research

Zitieren

APA

Talbot, D., Diop, A., Mésidor, M., Chiu, Y., Sirois, C., Spieker, A. J., Pariente, A., Noize, P., Simard, M., Luque Fernandez, M. A., Schomaker, M., Fujita, K., Gnjidic, D., & Schnitzer, M. E. (2025). Guidelines and Best Practices for the Use of Targeted Maximum Likelihood and Machine Learning When Estimating Causal Effects of Exposures on Time‐To‐Event Outcomes. Statistics in Medicine. https://doi.org/10.1002/sim.70034

Vancouver

Talbot D, Diop A, Mésidor M, Chiu Y, Sirois C, Spieker AJ, et al. Guidelines and Best Practices for the Use of Targeted Maximum Likelihood and Machine Learning When Estimating Causal Effects of Exposures on Time‐To‐Event Outcomes. Statistics in Medicine. 2025. doi:10.1002/sim.70034

BibTeX
@article{Talbot2025,
  author = {Talbot, Denis and Diop, Awa and Mésidor, Miceline and Chiu, Yohann and Sirois, Caroline and Spieker, Andrew J and Pariente, Antoine and Noize, Pernelle and Simard, Marc and Luque Fernandez, Miguel Angel and Schomaker, Michael and Fujita, Kenji and Gnjidic, Danijela and Schnitzer, Mireille E},
  title = {Guidelines and Best Practices for the Use of Targeted Maximum Likelihood and Machine Learning When Estimating Causal Effects of Exposures on Time‐To‐Event Outcomes},
  journal = {Statistics in Medicine},
  year = {2025},
  doi = {10.1002/sim.70034}
}
Kostenlos starten →