rightcensoring

Right-censoring occurs in time-to-event data when the event of interest has not occurred for an individual by the end of observation or the individual is lost to follow-up. In such cases the observed time is the minimum of the true event time and the censoring time, and the exact event time is unknown but exceeds the observed time.

Common causes include administrative end of study, withdrawal, or loss to follow-up. Censoring is called right-censoring when the event would occur later if followed longer. It is contrasted with left censoring and interval censoring, where the event is known to lie below a threshold or within an interval.

Assumptions: analyses typically assume non-informative (independent) censoring, meaning the probability of being censored is independent of the event time given covariates. Violations can bias results. Methods: the Kaplan-Meier estimator computes the survival function using censored observations; the Nelson-Aalen estimator computes the cumulative hazard; Cox proportional hazards models use partial likelihood to incorporate censored data.

In practice, right-censoring reduces information, and high censoring rates inflate variance. Data summaries report the number of events and the number censored. Software implementations are standard in survival analysis packages.