Extension to multiple covariates in the estimation of instantaneous risk function of the Aalen additive model / Extensão para várias covariáveis do método de estimação da função de risco instantâneo do modelo aditivo de Aalen

AUTOR(ES)
DATA DE PUBLICAÇÃO

2009

RESUMO

The Aalen additive model assesses the risk of occurrence of an event over time. The model is non parametric and estimates only accumulated risks. So be has no information regarding the risk of occurrence of the event at a given time. Thus, this study developed a estimation of the instantaneous risk function considering several covariates, using accumulated risk functions obtained from parametric distributions to smooth cumulative risk by stratifying the data parcel. To check goodness of fit of the parametric curves cumulative risks to nonparametric cumulative risks, it was implemented a test through compute-intensive methods. It was concluded that in presence of two dichotomous covariates, the cumulative risk of model Aalen can be estimated stratifying the data set and that this methodology can be used in a data set with any number of covariates. It also was concluded that the test proposed of adequacy controls type I error for different average proportions censorship (p = 0.30, 0.20, 0.10) and different sample sizes (n = 30, 50, 60, 90), being more conservative for small average proportions of censorship.

ASSUNTO(S)

distribuições paramétricas aalen model estatistica parametric distributions bootstrap modelo de aalen suavização smoothing bootstrap

Documentos Relacionados