Different interpretations for the sub-survival functions in a competing risks framework

• Autores: Norberto Porta, G. Gómez, María Luz Calle Rosingana
• Localización: XXX Congreso Nacional de Estadística e Investigación Operativa y de las IV Jornadas de Estadística Pública: actas, 2007, ISBN 978-84-690-7249-3
• Idioma: inglés
• Texto completo no disponible (Saber más ...)
• Resumen

  • Competing risks data arises when the failure of an individual may be classified into one of k (k>1) mutually exclusive causes of failure. The data motivating this paper come from a cohort of newly diagnosed bladder cancer cases. After first-line therapy, these patients are susceptible to distinct events: i) recurrence, ii) progression, and iii) death from bladder cancer. The time to the first event is clinically important to characterize distinct courses of the disease.

    A common error is to describe each failure type by using the Kaplan-Meier (KM) estimate. For instance, in the bladder cancer study, one could analyze recurrence by treating other events as censored observations and applying the standard KM estimate. However, in this case, such estimate does not represent a proper survival function, not even a sub-survival function, unless the different events are independent. The goal of this paper is to point out the role of fully interpretable functions, such as sub-distributions and sub-survivals, and to characterize the relationship among them.