Testes de significancia repetidos em analise de sobrevivencia

AUTOR(ES)
DATA DE PUBLICAÇÃO

2002

RESUMO

The main objective of this work is to elaborate a detailed exposition of the theory of repeated significance tests, with particular emphasis on non parametric tests used in Survival Analysis. A second objective is to study and use in applications the somehow new Weighted Kaplan Meier statistic that was created to compare survival in two populations. This statistics is an integrated weighted difference of Kaplan Meier estimates of survival curves in two groups. In the first chapter, a revision of basic definitions and tests used in Survival Analysis is made. After, the principles of the theory of repeated significance tests, developed around 1969 for comparison of means of a response variables Y with Normal distribution in two populations are presented (Armitage et al (1969), Pocock (1977), O Brien &Fleming (1979), Lan &DeMets (1983)). Then, at Chapter 3, the results for non immediate or censored response variables with an asymptotic normal distribution are presented. Finally, a resampling procedure for making a repeated significance test using the Weighted Kaplan Meier (Pepe &Fleming, 1989). and log rank statistics at each interim analysis is proposed. A study by simulation of this procedure is presented

ASSUNTO(S)

testes de hipoteses estatisticas bootstrap (estatistica)

ACESSO AO ARTIGO

Documentos Relacionados