Optimized Risk Score to Predict Mortality in Patients with Cardiogenic Shock in the ICU

Mortality prediction in critically ill patients with cardiogenic shock can guide triage and selection of potentially high-risk treatment options. In this work, we developed a simple risk score to predict in-hospital mortality for adults admitted to the cardiac ICU with SCAI Shock Stage C or greater cardiogenic shock. Our model was built and validated using a pair of real-world datasets that represent the largest patient cohorts for this task. We compared our model to alternative models built using conventional penalized logistic regression and to alternative models that were developed for this task – highlighting improvements in performance in diagnosis and risk estimation.