Abstract Sun1108: Development of a Novel Categorization Framework for In-Hospital Cardiac Arrest

Introduction: In-hospital cardiac arrest (IHCA) affects roughly 300,000 patients each year in the United States. Compared to out-of-hospital cardiac arrest (OHCA), IHCA remains understudied. OHCA categorization is well established, allowing for efficient and granular quality improvement and clinical research to improve OHCA outcomes. Similar approaches have not been applied to phenotype IHCA despite the abundance of available data from Electronic Medical Records (EMR). We hypothesized that IHCA data collected in the EMR can be used to categorize patients into clinically relevant phenotypes and consequently improve understanding of prognostic risk factors. Methods: We performed a retrospective cohort study of patients admitted to the Hospital of the University of Pennsylvania who suffered IHCA. Trained study team members extracted detailed information from the EMR, including the patient’s pre-admission status, hospital course, and the events in the 24 hours leading up to IHCA. The study team reviewed the pre-arrest trajectory of each case to identify phenotypes of pre-arrest deterioration (phenotypes are not mutually exclusive). Clinical criteria for each trajectory were also developed using expert input of study team members and available scientific literature. These phenotypes were iteratively developed, refining or adding phenotype definitions as new patterns emerged. A multidisciplinary panel of cardiac arrest experts reviewed each case until ten cases were classified without the need to modify or add a clinical phenotype. Additional IHCA cases were then categorized using these novel phenotypes. Results: Our cardiac arrest expert panel identified definitions and criteria for nine novel pre-arrest categories using the initial 100 cases (Figure 1). After review of 40 additional IHCA cases, they assigned all patients into at least one pre-arrest category. The most common pre-arrest trajectories were unstable tachyarrhythmia (41/202, 20%), unstable bradycardia (36/202, 18%), and respiratory failure (39/202, 19%). Conclusion: Our findings demonstrate that IHCA patients can be categorized into clinically relevant phenotypes using EMR-derived criteria. These phenotypes provide insight into pre-arrest trajectory and may help identify prognostic risk factors. By leveraging this categorization, we plan to investigate how these phenotypes are associated with patient outcome and whether they can inform resuscitation strategies or enhance quality improvement efforts.