EUVIMEDTutustu EUVIMEDiin
Randomisierte kontrollierte Studie

Cluster‐Level Analyses to Estimate a Risk Difference in a Cluster Randomized Trial With Confounding Individual‐Level Covariates: A Simulation Study

Jules Antoine Pereira Macedo iD, Bruno Giraudeau iD, ESCIENT collaborators

Statistics in Medicine · 2025

DOI: 10.1002/sim.70341 ↗

At a glance

Randomisierte kontrollierte Studie in Statistics in Medicine, 2025. Written by Jules Antoine Pereira Macedo and 2 co-authors. Institution: Université de Tours, Nantes Université, INSERM, SPHERE U1246, Tours, France. Major MEDLINE topics: Randomized Controlled Trials as Topic. 31 references. Funded by: Agence Nationale de la Recherche.

TiivistelmäSubjectsBibliografiset tiedotCiteRelated studies

Abstract (original language)

Cluster randomized trials (CRTs) may be analyzed using cluster‐level analyses. For binary outcomes, proportions are estimated for each cluster, and a risk difference can be estimated. The confidence interval is estimated using a Student distribution. However, in doing so, individual‐level characteristics are not adjusted for even though CRTs are known to be prone to recruitment/identification bias, possibly implying individual‐level confounders. With a simulation study, we compared cluster‐level analyses to estimate a risk difference for a two‐arm parallel CRT with individual‐level confounders and cluster‐level covariates. We considered the unadjusted (UN) method, two two‐stage procedure (TSP) methods considering a binomial or a Gaussian distribution, G‐computation (GC), and targeted maximum likelihood estimation (TMLE) methods. As expected, the UN method was biased. TSP methods were also biased for scenarios with a treatment effect when the number of clusters per arm was small. GC and TMLE methods were unbiased. For these latter methods, adjustment on only individual‐level covariates led to better performance measures (type I error rate, coverage rate and relative error of the standard error) than adjustment on both individual‐ and cluster‐level covariates. TSP, GC and TMLE had very similar results except in scenarios with a small number of clusters: Biased results for TSP methods and convergence problems for GC methods. In this case, TMLE should be preferred.

Subjects

MeSH terms (MEDLINE)

Randomized Controlled Trials as TopicCluster AnalysisHumansComputer SimulationLikelihood FunctionsBiasConfounding Factors, EpidemiologicModels, StatisticalNormal Distribution

Author keywords

binary outcomecluster randomized trialcluster‐level analysesg‐computationrisk differencetargeted maximum likelihood estimation

Authors and institutions

  • Université de Tours, Nantes Université, INSERM, SPHERE U1246, Tours, France.
  • INSERM, CIC1415, CHRU de Tours, Tours, France.

ORCID profiles: Jules Antoine Pereira Macedo, Bruno Giraudeau

Funding

  • Agence Nationale de la Recherche

Cite

APA

Pereira Macedo, J. A., Giraudeau, B., & ESCIENT collaborators (2025). Cluster‐Level Analyses to Estimate a Risk Difference in a Cluster Randomized Trial With Confounding Individual‐Level Covariates: A Simulation Study. Statistics in Medicine. https://doi.org/10.1002/sim.70341

Vancouver

Pereira Macedo JA, Giraudeau B, ESCIENT collaborators. Cluster‐Level Analyses to Estimate a Risk Difference in a Cluster Randomized Trial With Confounding Individual‐Level Covariates: A Simulation Study. Statistics in Medicine. 2025. doi:10.1002/sim.70341

BibTeX
@article{PereiraMacedo2025,
  author = {Pereira Macedo, Jules Antoine and Giraudeau, Bruno and ESCIENT collaborators, ESCIENT collaborators},
  title = {Cluster‐Level Analyses to Estimate a Risk Difference in a Cluster Randomized Trial With Confounding Individual‐Level Covariates: A Simulation Study},
  journal = {Statistics in Medicine},
  year = {2025},
  doi = {10.1002/sim.70341}
}
Start for free →