Kaplan-Meier estimate

The Kaplan-Meier estimator is a nonparametric way to estimate a survival function: the probability of remaining event-free up to each point in time. Its strength is handling censoring, meaning participants who leave the study or are still event-free when observation stops contribute information for as long as they were followed. The resulting curve steps downward at each time an event occurs, and the height at any time approximates the proportion event-free. When reading a study, look at the number-at-risk table beneath the curve, because estimates late in follow-up often rest on few remaining participants and become unreliable; also check whether groups are compared with an appropriate test such as the log-rank test. Example: in a cancer trial, a Kaplan-Meier curve might show 70 percent of patients alive at two years, while a widening gap between treatment and control curves suggests a survival difference worth quantifying with a hazard ratio.