Abstract 45: Reexamination of a Clinical Decision Rule to Discontinue Resuscitation During In-Hospital Cardiac Arrest
Background: Several clinical decision rules have been developed to help practitioners avoid futile resuscitative efforts after in-hospital cardiac arrests (IHCA). One parsimonious decision rule, developed and validated by van Walraven and colleagues (Arch Int Med, 1999; JAMA 2001) was highly predictive of survival to discharge after IHCA in a small sample. However, it’s large-scale applicability remains unknown.
Hypothesis: The model will continue to perform well in the context of a diverse population and among contemporary hospitals.
Methods: We examined the model’s sensitivity and specificity for predicting survival to discharge after IHCA in 75,272 patients from 372 US hospitals within the Get With The Guidelines-Resuscitation registry. The model defines a chance of survival if the arrest met any one of the following three conditions: 1) was witnessed or monitored, 2) initial rhythm was ventricular tachycardia or ventricular fibrillation, or 3) was less than 10 minutes in duration. We compared observed and predicted survival to hospital discharge with a 2x2 contingency table. A secondary analysis assessed favorable neurological status among survivors using cerebral performance category scores of ≤ 2.
Results: The van Walraven model correctly identified 12,003 (16%) patients who survived to hospital discharge with a sensitivity of 98.2% (95% CI, 97.9%-98.4%) and a specificity of 8.1% (7.9%-8.3%). Of the 5,322 patients who were predicted to have no chance of survival to discharge, only 222 (1.8%) actually survived, representing a negative predictive value (NPV) of 95.8% (95.3%-96.3%). Combining favorable neurological status with survival to discharge increased the sensitivity and NPV to 98.4% (98.2%-98.7%) and 97.0% (96.6%-97.5%), respectively. Overall, the application of this tool to a broader population would only miss 1.3% of all survivors with a favorable neurological outcome.
Conclusion: This simple clinical decision rule, using three intra-arrest variables, had over 95% predictive accuracy in a large national sample for identifying IHCA patients who will not survive to discharge, and this negative predictive accuracy increased to 98.7% when it examined survival with favorable neurologic status.
Author Disclosures: D.N. Bennett: None. R.A. Hayward: None. S.M. Bradley: None. B.K. Nallamothu: None. S.E. Moser: None. P.S. Chan: None. Z.D. Goldberger: None.
- © 2014 by American Heart Association, Inc.