A prediction survival model for out-of-hospital cardiopulmonary resuscitations

Abstract 

Aims

The aims of this study were to double check old (Resuscitation Predictor Scoring [RPS], Advanced Cardiac Life Support, and Early Prediction Score [EPS]) and form new (Serbian Quality of Life immediately [SR-QOLi], Serbian Quality of Life short-term [SR-QOLs], and Serbian Quality of Life long-term [SR-QOLl]) scores for survival prediction in out-of-hospital cardiopulmonary resuscitation (OHCPR) in Serbia.

Methods

A prospective, 2-year, multicentric study was designed. By the means of the Utstein style, OHCPR performed and its outcome were followed. In every patient, immediate (i) (Return of Spontaneus Circulation [ROSC] >20 min), short-term (s) (to hospital discharge), and long-term [1] (1 year upon) survival after the OHCPR, under the application of RPS, ASCLS, and EPS models, was evaluated. We assessed the association between survival rate and individual predictors of OHCPR using RPS, ASCLS, and EPS: cardiopulmonary resuscitation (CPR) started (>4 or <4 minutes after out-of-hospital cardiac arrest), swallowing activity (present or not), the primary arrest mode (cardiac or respiratory), and initial pupillar photoreaction (present or absent). By the successive-logistic and linear-regression analysis method, the additional model of the type SR-QOL (SR-QOLi, SR-QOLs, and SR-QOLl) was created.

Results

We found that bystander CPR, witnessed arrest, shockable rhythms, CPR within 4 minutes, pupillar photoreaction, and primary cardiac arrest mode were associated with improved survival. Cumulative survival upon OHCPR was 12.7% for immediate, 11.3% before patient's discharge, and 10% after 12 months. Applied on our sample, standard scores displayed satisfactory (RPS) and good (Advanced Cardiac Life Support and EPS) degree of survival prediction in OHCPR. In receiver operator characteristic (ROC) analysis, SR-QOLi (ROC = 0.833) and SR-QOLs (ROC = 0.882) were defined as a good models and SR-QOLl (ROC = 0.913) was defined as an excellent model for prediction of outpatient CPR outcomes.

Conclusion

In the course of the research, SR-QOL models were created for prediction of the immediate (SR-QOLi), short-term (SR-QOLs), and long-term (SR-QOLl) survival after the OHCPR, better predictions in our environment.

Keywords: Cardiopulmonary resuscitations, Out-of-hospital, Survival, Predictions, Serbian Quality of Life, Models