BACKGROUND: A prearrest prediction tool can aid clinicians in consolidating objective findings with clinical judgement and in balance with the values of the patient be a part of the decision process for do-not-attempt-resuscitation (DNAR) orders. A previous prearrest prediction tool for in-hospital cardiac arrest (IHCA) have not performed satisfactory in external validation in a Swedish cohort. Therefore our aim was to develop a prediction model for the Swedish setting.

METHODS: Model development was based on previous external validation of The Good Outcome Following Attempted Resuscitation (GO-FAR) score, with 717 adult IHCAs. It included redefinition and reduction of predictors, and addition of chronic comorbidity, to create a full model of 9 predictors. Outcome was favourable neurological survival defined as Cerebral Performance Category score 1-2  at discharge. The likelihood of favourable neurological survival was categorised into very low (<1%), low (1-3%) and above low (>3%).

RESULTS: We called the model the Prediction of outcome for In-Hospital Cardiac Arrest (PIHCA) score. The AUROC was 0.808 (95% CI 0.807-0.810) and calibration was satisfactory. With a cutoff of 3% likelihood of favourable neurological survival sensitivity was 99.4% and specificity 8.4%. Although specificity was limited, predictive value for classification into ≤3% likelihood of favorable neurological survival was high (97.4%) and false classification into ≤3% likelihood of favourable neurological survival was low (0.6%).

CONCLUSION: The PIHCA score has the potential to be used as an objective tool in prearrest prediction of outcome after IHCA, as part of the decision process for a DNAR order.