Fig. 2From: Estimating a population cumulative incidence under calendar time trendsCumulative incidence in a hypothethical cohort divided into two strata with a time trend given by a proportional hazards model. In the most prevalent stratum, an administrative censoring is imposed at time t=1 but no censoring is happening in the least prevalent stratum. Based on a sample of 50,000 individuals in each stratum, the pooled Kaplan–Meier estimator and the Kaplan–Meier estimator based on a proportional hazards model is computedBack to article page