Patient safety in hospitals - a Bayesian analysis of unobservable hospital and specialty level risk factors.

摘要: This paper demonstrates how Bayesian hierarchical modelling can be used to evaluate the performance of hospitals. We estimate a three-level random intercept probit model attribute unexplained variation in hospital-acquired complications hospital effects, hospital-specialty effects and remaining variations, controlling for observable patient complexities. The combined information provided by posterior means densities latent specialty assess need scope improvements safety at different organizational levels. Posterior are not conventionally presented assessment but provides valuable additional policy makers on what poorly performing hospitals specialties may prioritized action. use surgical administrative data 2005/2006 16 35 public Victoria, Australia. compare safety. variances also compared identify clinical areas with greatest improvement. show that same rank markedly differently specialties.