Beyond the alarm: proactive predictions for cardiac arrest incidents in hospitals using Interpretable machine learning models

Code Blue is an emergency alert signal activated whenever a hospitalized patient experiences cardiac or respiratory arrest. This research aims to build an interpretable machine-learning model for predicting Code Blue events before they can happen. This study uses the electronic medical record data of the patient for the last 24 hours and the doctor’s clinical notes. We use the technique of Natural Language Processing (NLP) to extract features from the doctor’s clinical notes. The extracted features are combined with other electronic medical data, such as vital information, pre-existing diseases, and organ dysfunction, to build the final prediction model. We use Explainable Artificial Intelligence (XAI) methods to interpret the model. The enhanced ability to explain the model output will help healthcare professionals better understand the early warning signals of Code Blue. The prompt alert from an early warning signal ensures that patients receive immediate care from medical practitioners, thereby substantially increasing the potential for saving lives. Based on these findings, hospital managers can reframe the policy regarding the announcement of Code Blue. Hospitals can use a decision support tool to prioritize patients and manage the operations of the code blue team.