Hospital clustering in the treatment of acute myocardial infarction patients via a Bayesian nonparametric approach

We fit a Bayesian generalized mixed-effects model for binary data. For patient i = 1, . . . ,n j in each hospital j = 1, . . . ,J, let Yi j be a Bernoulli random variable with mean pi j, which represents the in-hospital probability that the patient survived after STEMI. The pi j’s are modelled through a logit regression with covariates xi j := (1,xi j1,xi j2,xi j3) which represent age, Symptom Onset to Ballon time in the log scale (the time since the symptoms onset up to the surgical practice of angioplasty) and killip (denoting the severity of the infarction), respectively, of the patients,