The interactions considered Staurosporine ic50 were prespecified following presentation of an initial model with no interaction terms and requesting input from the NCAA Steering Group, representatives from hospitals participating in NCAA, and an Expert Group of clinicians, statisticians and health services researchers formed to advise on risk modelling (see ‘Acknowledgements’ section). The interactions considered were: • age with sex; Interaction terms were added to the full model and retained if significant at P < 0.01. For the interaction of location of arrest with presenting rhythm, in order to reduce the potentially large number of interaction terms, combining interaction terms for similar groups

of categories of both presenting rhythm (e.g. all shockable arrests,

all non-shockable arrests) and location of arrest (e.g. ED and EAU, EAU and ward, CCU and cardiac catheter laboratory) Trichostatin A in vitro was considered. Comparisons of models (for testing linearity, combining categories, stepwise reduction and adding interactions) were performed with likelihood ratio tests. The resulting models were validated for discrimination, calibration and accuracy in: (1) the development dataset; (2) the full validation dataset; and (3) the validation data from hospitals that commenced participation in NCAA from April 2012 onwards and were therefore not included in the development dataset (providing true external validation in a smaller sample of hospitals). To reduce overfitting, model estimates were shrunk using the uniform (heuristic) shrinkage method of Van Houwelingen and Le Cessie.15 Discrimination was assessed by the c index. Calibration was assessed graphically and tested using the Hosmer–Lemeshow test for perfect calibration in ten equal sized groups by predicted probability of survival. As the Hosmer–Lemeshow test does not provide a measure of the degree of miscalibration and is very sensitive to sample size,16 calibration was also assessed using Cox’s calibration regression, which assesses the degree of linear miscalibration by fitting a logistic regression of observed survival on the predicted log odds of survival

from the risk model.17 Accuracy was assessed by Brier’s score Protein tyrosine phosphatase and Shapiro’s R, and the associated approximate R-squared statistics (termed the ‘sum-of-squares’ R-squared and the ‘entropy-based’ R-squared, respectively)18, which are obtained by scaling each measure relative to the value achieved from a null model. Measures of model performance were calculated using the marginal predicted probabilities from the risk model, i.e. without taking into account hospital level effects, representing the predicted probability of survival for a patient with the given characteristics in an ‘average’ hospital. The final risk models were refitted to all data (development and validation datasets combined) to maximise precision and generalisability, with shrinkage applied to reported coefficients.