Improving CPR Predictive Model in ED: The Role of Initial Data and KTAS

Area of Science:

• Emergency Medicine
• Clinical Informatics
• Cardiology

Background:

• Cardiac arrest (CPR) is a critical event in emergency departments.
• Predictive models can aid in early intervention and prevention.
• Initial patient data is crucial for timely risk assessment.

Purpose of the Study:

• To develop and validate a predictive model for CPR using initial emergency department admission data.
• To compare the predictive performance of models with and without the Korean Triage and Acuity Scale (KTAS).
• To assess the utility of integrating vital signs and KTAS for CPR risk stratification.

Main Methods:

• Retrospective analysis of electronic medical record (EMR) data from Severance Hospital (2018-2022).
• Development of two predictive models: one with initial vital signs and patient information, and another including KTAS.
• Exclusion criteria included patients under 18, missing vital signs or KTAS, and those with Do-Not-Resuscitate (DNR) orders.

Main Results:

• The model incorporating KTAS demonstrated significantly superior predictive performance compared to the model using only vital signs and patient information.
• The integration of vital signs and KTAS proved effective in predicting the likelihood of CPR.
• The study identified key predictors for CPR within the initial emergency department assessment.

Conclusions:

• Combining initial vital signs with KTAS provides a robust approach for predicting CPR in the emergency department.
• This predictive model can support clinical decision-making to potentially prevent cardiac arrest.
• Utilizing KTAS in predictive modeling enhances the accuracy of identifying high-risk patients for early intervention.