Comparer des méthodes

	Modèle conjoint pour données longitudinales et données de temps d'événement	Modèle de survie pour événements récurrents
Domaine	Analyse de survie	Analyse de survie
Famille	Survival analysis	Survival analysis
Année d'origine	2004	1981
Auteur d'origine	Tsiatis, A.A. & Davidian, M.; Rizopoulos, D.	Andersen & Gill (AG, 1982); Prentice, Williams & Peterson (PWP, 1981); Wei, Lin & Weissfeld (WLW, 1989)
Type	Semiparametric regression model	Semi-parametric hazard model for repeated events
Source fondatrice	Rizopoulos, D. (2012). Joint Models for Longitudinal and Time-to-Event Data. CRC Press.	Cook, R.J. & Lawless, J.F. (2007). The Statistical Analysis of Recurrent Events. Springer.
Alias	joint model, shared random effects model, longitudinal-survival joint model, Joint Model (Boylamsal + Sağkalım Birleşik Model)	Tekrarlayan Olay Modeli (Recurrent Events), Andersen-Gill model, AG model, Wei-Lin-Weissfeld model
Apparentées	5	4
Résumé	The joint model for longitudinal and time-to-event data, formalised by Tsiatis and Davidian in 2004 and extended comprehensively by Rizopoulos in 2012, simultaneously estimates a mixed-effects model for repeatedly measured biomarkers and a survival model for the time to an event, linking the two processes through shared random effects. It resolves two major problems that simpler approaches cannot handle: informative dropout from longitudinal studies and the endogeneity of time-varying biomarkers used as covariates in a Cox model.	A recurrent event model is a survival analysis extension, formalised through the landmark contributions of Prentice, Williams and Peterson (1981), Andersen and Gill (1982), and Wei, Lin and Weissfeld (1989), that models time-to-event data when the same event — such as a hospital readmission, disease relapse, or equipment failure — can occur multiple times in the same individual. The three principal frameworks are the Andersen-Gill (AG) model, the Prentice-Williams-Peterson (PWP) stratified model, and the Wei-Lin-Weissfeld (WLW) marginal model, each making different assumptions about within-subject dependence.