Improving Risk Stratification of Emergency Department Patients With Acute Heart Failure

Improving Risk Stratification of Emergency Department Patients With Acute Heart Failure: Building and Testing a Machine-learning Platform for Personalized, Accurate, Real-time Risk Prediction

The primary goal is to build and test a previously developed and validated risk model and clinical decision support tool embedded within the electronic health record to improve risk stratification of emergency department (ED) patients with acute heart failure (AHF).

The study team will build an electronic health record-embedded clinical decision support tool using a recently developed risk prediction model that curates patient-specific data in real-time, accurately estimates short-term patient risk, and presents tailored clinical recommendations. This will be a regional implementation study in which the tool is turned on at 21 emergency departments (ED) across Kaiser Permanente Northern California (KPNC). The study team will validate risk predictions and study key clinical outcomes as part of this trial.

Inclusion Criteria:

• Adult (≥18y) KPNC members treated in the ED for presumed acute heart failure

Exclusion Criteria:

• Children (<18y).
• Patients who left against medical advice or eloped prior to ED physician evaluation.