An Utstein-based model score to predict survival to hospital admission: The UB-ROSC score

Abstract

Aims: To develop and validate a multi-parametric practical score to predict the probability of survival to hospital admission of an out-of-hospital cardiac arrest (OHCA) victim by using Utstein Style-based variables.

Methods: All consecutive OHCA cases occurring from 2015 to 2017 in two regions, Pavia Province (Italy) and Canton Ticino (Switzerland) were included. We used random effect logistic regression to model survival to hospital admission after an OHCA. We computed the model area under the ROC curve (AUC ROC) for discrimination and we performed both internal and external validation by considering all OHCAs occurring in the aforementioned regions in 2018. The Utstein-Based ROSC (UB-ROSC) score was derived by using the coefficients estimated in the regression model. The score value was obtained adding the pertinent score components calculated for each variable. The score was then plotted against the probability of survival to hospital admission.

Results: 1962 OHCAs were included (62% male, mean age 73 ± 16 years). Age, aetiology, location, witnessed OHCA, bystander CPR, EMS arrival time and shockable rhythm were independently associated with survival to hospital admission. The model showed excellent discrimination (AUC 0.83, 95%CI 0.81-0.85) for predicting survival to hospital admission, also at internal cross-validation (AUC 0.82, 95%CI 0.80-0.84). The model maintained good discrimination after external validation by using the 2018 OHCA cohort (AUC 0.77, 95%CI 0.74-0.80).

Conclusions: UB-ROSC score is a novel score that predicts the probability of survival to hospital admission of an OHCA victim. UB-ROSC shall help in setting realistic expectations about sustained ROSC achievement during resuscitation manoeuvres.

Keywords: Out of hospital cardiac arrest; Prediction; Return of spontaneous circulation; Score; Survival.