Semi- and non-parametric competing risks analysis of right censored data and failure cause missing completely at random

Laurent Bordes (1) , Jean-Yves Dauxois (2) , Pierre Joly (3)

1 LMAP - Laboratoire de Mathématiques et de leurs Applications [Pau]
2 LMB - Laboratoire de Mathématiques de Besançon (UMR 6623)
3 ISPED - Institut de Santé Publique, d'Epidémiologie et de Développement

We consider a nonparametric and a semiparametric (in presence of covariates) additive hazards rate competing risks model with censoring and failure cause possibly missing completely at random. Estimators of the unknown parameters are proposed in order to satisfy some optimality criteria. Large ample results are given for all the estimators. Our nonparametric method is applied to a real data set and the behavior of the semiparametric estimators are analyzed through a Monte Carlo study.

Mots clés: Additive hazards; competing risks; counting processes; missing failure cause; reliability; survival analysis.; survival analysis

Domaines: Statistiques [math.ST]

Laurent Bordes, Jean-Yves Dauxois, Pierre Joly. Semi- and non-parametric competing risks analysis of right censored data and failure cause missing completely at random. 2007. ⟨hal-00193032⟩