STBRIER: Stata module to compute Brier score for censored time-to-event (survival) data

stbrier computes the Brier score for risk prediction models in survival analysis based on right censored data by weighting individuals by their inverse probability of being uncensored [Graf et al. 1999; Gerds & Schumacher 2006; Binder et al. 2009] (see brier for non-censored data). In settings without competing risks stbrier can compute the Brier score of a parametric (streg) or semi-parametric Cox regression model (stcox), and in the presence of competing risks, the Brier score of a Fine-Gray regression model (stcrreg). To serve as a benchmark (null model), stbrier can be computed without covariates [Gerds et al. 2008]. Additionally, the user can use a simple wrapper program to compute the integrated Brier score [Graf et al.1999] across all values of btime (see the example below).