Abstract

JARAMILLO ELORZA, Mario César  and  LOPERA GOMEZ, Carlos Mario. Study of the effect of failure imputation in estimating the survival curve under interval censoring. Ciencia en Desarrollo [online]. 2017, vol.8, n.1, pp.21-28. ISSN 0121-7488.

Most survival analyzes are based on exact failure times and right censored observations, using methods widely known as the Kaplan-Meier (KM). When the data are interval censored is necessary to use the Turnbull's method to estimate the survival function, but in practice is often used the imputation of failure times in this kind of censorship through the midpoint of the interval, the right end of the interval or generating a random point within the interval using the uniform distribution. This paper studies through simulation the effect of three types of imputation on the estimates of the survival curve compared to the method developed by Turnbull. Different simulation scenarios based on the sample size and the time between visits were analyzed. In all scenarios simulation functions estimated using data imputation differ significantly from the true survival function S(t).

Keywords : Survival Analysis; Interval Censoring; Data Imputation.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )