Early Prediction of Cardiogenic Shock Using Machine Learning
Yale Chang1, Corneliu Antonescu2,3, Shreyas Ravindranath1
1Philips Research North America, Cambridge, MA, United States.
Insights
A new machine learning model accurately predicts cardiogenic shock (CS) development up to two hours in advance using electronic health record data. This tool aids early intervention for severe cardiac events, improving patient outcomes.
Area of Science:
- * Cardiology and Artificial Intelligence in Healthcare
- * Clinical Decision Support Systems
- * Predictive Analytics in Critical Care
Background:
- * Cardiogenic shock (CS) presents a significant clinical challenge with high in-hospital mortality rates.
- * Timely identification and management of CS are often hindered by varied patient presentations and settings (ED, WARD, CC).
- * Existing diagnostic and management pathways face challenges in early detection.
Purpose of the Study:
- * To develop and retrospectively evaluate a machine learning (ML) model for predicting cardiogenic shock.
- * To enable early identification of patients at risk for CS, facilitating timely interventions.
- * To assess the model's predictive performance across different clinical settings.
Main Methods:
- * Development of an XGBoost (XGB) based ML model utilizing de-identified electronic health record (EHR) data from 2010-2017.
- * Input variables included demographics, vital signs, lab values, orders, and pre-existing diagnoses.
- * Model designed to predict CS need 2 hours prior to intervention (inotropes, vasopressors, mechanical support).
Main Results:
- * The ML model achieved a high overall area under the curve (AUC) of 0.87.
- * Performance varied by setting: AUC of 0.81 in Critical Care (CC), 0.84 in Emergency Department (ED), and 0.97 in General Wards (WARD).
- * Model precision increased when refined for specific subpopulations like acute myocardial infarction (AMI) or congestive heart failure (CHF).
Conclusions:
- * A machine learning model utilizing EHR data can effectively predict the development of cardiogenic shock.
- * The developed XGBoost model demonstrates significant clinical utility for early CS detection.
- * Further refinement of the model can enhance its precision for specific patient groups, supporting proactive clinical management.
Abstract:
Cardiogenic shock (CS) is a severe condition with in-hospital mortality of up to 50%. Patients who develop CS may have previous cardiac history, but that may not always be the case, adding to the challenges in optimally identifying and managing these patients. Patients may present to a medical facility with CS or develop CS while in the emergency department (ED), in a general inpatient ward (WARD) or in the critical care unit (CC). While different clinical pathways for management exist once CS is recognized, there are challenges in identifying the patients in a timely manner, in all settings, in a timeframe that will allow proper management. We therefore developed and evaluated retrospectively a machine learning model based on the XGBoost (XGB) algorithm which runs automatically on patient data from the electronic health record (EHR). The algorithm was trained on 8 years of de-identified data (from 2010 to 2017) collected from a large regional healthcare system. The input variables include demographics, vital signs, laboratory values, some orders, and specific pre-existing diagnoses. The model was designed to make predictions 2 h prior to the need of first CS intervention (inotrope, vasopressor, or mechanical circulatory support). The algorithm achieves an overall area under curve (AUC) of 0.87 (0.81 in CC, 0.84 in ED, 0.97 in WARD), which is considered useful for clinical use. The algorithm can be refined based on specific elements defining patient subpopulations, for example presence of acute myocardial infarction (AMI) or congestive heart failure (CHF), further increasing its precision when a patient has these conditions. The top-contributing risk factors learned by the model are consistent with existing clinical findings. Our conclusion is that a useful machine learning model can be used to predict the development of CS. This manuscript describes the main steps of the development process and our results.
