Pitfalls in Time-to-Event Analysis of Registry Data: A Tutorial based on Simulated and Real Cases - Archive ouverte HAL Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2023

Pitfalls in Time-to-Event Analysis of Registry Data: A Tutorial based on Simulated and Real Cases

Résumé

Based on both the use of real-world data and a simulation approach, this tutorial article looks at how overlooking right censoring, left truncation, competing events and recurrent event methods in analyses of time-to-event data can lead to suboptimal or biased estimations. The simulation approach enables a comparison with the truth and thus concrete estimations of bias for methods that (i) do not consider right censoring (leading to underestimation of survival), (ii) do not consider left truncation (leading to overestimation of survival), (iii) treat competing risks as right-censoring (leading to overestimation of survival), and (iv) consider only the first of a series of recurrent events for a given individual.
Fichier principal
Vignette du fichier
CEREDIH_EPISTAT_Manuscript_Survival_20230928_corrected_OB_V4.pdf (1.86 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04358291 , version 1 (21-12-2023)

Identifiants

  • HAL Id : hal-04358291 , version 1

Citer

Mickaël Alligon, Nizar Mahlaoui, Olivier Bouaziz. Pitfalls in Time-to-Event Analysis of Registry Data: A Tutorial based on Simulated and Real Cases. 2023. ⟨hal-04358291⟩
8 Consultations
9 Téléchargements

Partager

Gmail Facebook X LinkedIn More