EUVIMEDOpdag EUVIMED
Open access · CC BY 4.0

Causal Machine Learning Methods and Use of Cross‐Fitting in Settings With High‐Dimensional Confounding

Susan Ellul iD, Stijn Vansteelandt iD, John B Carlin, Margarita Moreno-Betancur

Statistics in Medicine · 2025

DOI: 10.1002/sim.70272 ↗

At a glance

Journal article in Statistics in Medicine, 2025. Written by Susan Ellul and 3 co-authors. Institution: Murdoch Children's Research Institute, Parkville, Victoria, Australia. Major MEDLINE topics: Machine Learning, Causality. Cited 6 times so far (Crossref); 69 references. The full text is freely available (open access, CC BY 4.0). Funded by: Australian National Health and Medical Research Council (NHMRC) Investigator Grant Emerging Leadership Level 2, Australian Government Research Training Program Scholarship, Victorian Government's Operational Infrastructure Support Program.

AbstractSubjectsBibliografiske oplysningerCiteRelated studies

Abstract (original language)

Observational epidemiological studies commonly seek to estimate the causal effect of an exposure on an outcome. Adjustment for potential confounding bias in modern studies is challenging due to the presence of high‐dimensional confounding, which occurs when there are many confounders relative to sample size or complex relationships between continuous confounders and exposure and outcome. Doubly robust methods such as Augmented Inverse Probability Weighting (AIPW) and Targeted Maximum Likelihood Estimation (TMLE) have the potential to address these challenges, using data‐adaptive approaches and cross‐fitting, but despite recent advances, limited evaluation and guidance are available on their implementation in realistic settings where high‐dimensional confounding is present. Motivated by an early‐life cohort study, we conducted an extensive simulation study to compare the relative performance of AIPW and TMLE using data‐adaptive approaches for estimating the average causal effect (ACE). We evaluated the benefits of using cross‐fitting with a varying number of folds, as well as the impact of using a reduced versus full (larger, more diverse) library in the Super Learner ensemble learning approach used for implementation. We found that AIPW and TMLE performed similarly in most cases for estimating the ACE, but TMLE was more stable. Cross‐fitting improved the performance of both methods, but was more important for variance estimation and coverage than for point estimates, with the number of folds a less important consideration. Using a full Super Learner library was important to reduce bias and variance in complex scenarios typical of modern health research studies.

Subjects

MeSH terms (MEDLINE)

Machine LearningCausalityHumansLikelihood FunctionsComputer SimulationConfounding Factors, EpidemiologicBiasCohort StudiesModels, StatisticalObservational Studies as Topic

Author keywords

augmented inverse probability weightingcausal inferencecross‐fittingdoubly robusthigh‐dimensional confoundingtargeted maximum likelihood estimation

Authors and institutions

  • Murdoch Children's Research Institute, Parkville, Victoria, Australia.
  • Department of Paediatrics, University of Melbourne, Melbourne, Victoria, Australia.
  • Department of Mathematics, Computer Science and Statistics, Ghent University, Ghent, Belgium.

ORCID profiles: Susan Ellul, Stijn Vansteelandt

Funding

  • Australian National Health and Medical Research Council (NHMRC) Investigator Grant Emerging Leadership Level 2
  • Australian Government Research Training Program Scholarship
  • Victorian Government's Operational Infrastructure Support Program
  • National Health and Medical Research Council

Cite

APA

Ellul, S., Vansteelandt, S., Carlin, J. B., & Moreno-Betancur, M. (2025). Causal Machine Learning Methods and Use of Cross‐Fitting in Settings With High‐Dimensional Confounding. Statistics in Medicine. https://doi.org/10.1002/sim.70272

Vancouver

Ellul S, Vansteelandt S, Carlin JB, Moreno-Betancur M. Causal Machine Learning Methods and Use of Cross‐Fitting in Settings With High‐Dimensional Confounding. Statistics in Medicine. 2025. doi:10.1002/sim.70272

BibTeX
@article{Ellul2025,
  author = {Ellul, Susan and Vansteelandt, Stijn and Carlin, John B and Moreno-Betancur, Margarita},
  title = {Causal Machine Learning Methods and Use of Cross‐Fitting in Settings With High‐Dimensional Confounding},
  journal = {Statistics in Medicine},
  year = {2025},
  doi = {10.1002/sim.70272}
}
Start for free →