Time to event analysis (survival)

time to event analysis In a prospective or retrospective cohort studies, subjects with or without exposure are followed longitudinally over a fixed period of time for a given discrete event or outcome. But if the event of interest does not occur during study period, data of such study subjects are censored. Censoring, however, does not mean deletion of such data and their data till the time of censoring (i.e. till such subjects are under observation) are analyzed. Censoring can be left sided (onset of exposure not well defined) or right sided (subjects lost to follow up, withdrawn prematurely from study or who did not have event till end of study period).

Most common univariate survival analysis technique is non-parametric Kaplan Meier survival analysis whereas popular semi-parametric multivariate survival analysis involves Cox proportional hazard analysis.