Modelos de riscos proporcionais e aditivos para o tratamento de covariÃveis dependentes do tempo

AUTOR(ES)

Tarciana Liberal Pereira

DATA DE PUBLICAÇÃO

2004

RESUMO

Frequently in survival analysis when covariates are incorporated in the analysis its registered values are those measured in the origin of the time or the beginning of the study. However in many studies that involve data of survival covariates exist that are monitored during the study and its values move in this period. These covariveis whose values if modify with the time are known as Covariates Dependent of the Time and have much utility in the analysis of survival data therefore can be used in such a way to accomodate measured that vary with the time during the study as also to model the effect of individuals that move of group during a treatment. Analyses that consider these covariates can supply resulted more necessary and the inclusion of them cannot cause serious vices of estimation. A sufficiently flexible and extensively used model in analysis of survival for incorporating the effect of fixed covariates is the model of proportional risks of Cox that can be generalized to incorporate the effect of dependent covariates of the time. Although not to be very used in the practical one, alternative models to the one of Cox have been suggested to the long one of the years. Aalen considered a model of additive or linear risk that supplies a useful alternative the model of Cox. This model has shown frequently practical advantages especially when the covariates have effect varying in the time therefore separately allow the control of the influence of the variation in the time of each covariate. In this work these two models are presented and are shown to the use in the presence of dependents covariates of the time. Two real data bases are used to compare the adjustments of these two models. In the first application these models are used to evaluate factors that can be related with the duration of the maternal breastfeeding. In the second application it is verified if the infection for the HIV is factor of risk for the development of the sinusite

ASSUNTO(S)

covariÃveis dependentes estatistica modelo de cox

Documentos Relacionados

• Testes de hipoteses com alternativas restritas em modelos de riscos proporcionais
• Modelos de tratamento para o transtorno do pânico
• Estabilidade dinâmica para sistemas quânticos dependentes do tempo.
• Motivações de dependentes químicos para o tratamento: percepção de familiares
• Formulação LTSn para problemas de transporte sem simetria azimutal e problemas dependentes do tempo