Skip to main content

Advertisement

Figure 4 | BMC Medical Research Methodology

Figure 4

From: How to handle mortality when investigating length of hospital stay and time to clinical stability

Figure 4

Probability of hospital discharge from simulated data. Median values of each estimation method for the probability of hospital discharge, based on 1,000 simulations, as detailed in the 'Simulated Data' section of the Methods. The underlying cumulative incidence curves are shown in solid black lines. Median estimates of the restricted analysis estimator 1- Ŝ d i s R ( t ) (top left panel, orange curve) and the complement of the Kaplan-Meier estimator 1- Ŝ d i s W ( t ) (top right panel, green curve) each overestimate the probability of discharge. Median estimates from the cumulative incidence estimator C I ^ d i s ( t ) (purple, lower left panel) and the worst outcome estimator 1- Ŝ d i s W ( t ) (blue, lower right panel) coincided exactly with the true cumulative incidence function, as was expected. Shaded polygons give the 97.5% and 2.5% percentiles of the estimates from the simulations, for each time point.

Back to article page