Cardiogenic Shock Detection Using Electronic Medical Records: A Review and Blueprint for Clinical Implementation and Future Research

Despite advances in drug and device technology, health care delivery, and research infrastructure, cardiogenic shock (CS) continues to have nearly 50% in‐hospital mortality. In patients with CS, both the initial severity of Society for Cardiovascular Angiography and Intervention CS and its subsequent trajectory predicts the clinical outcomes. Accordingly, delayed initial recognition and failure to escalate or deescalate treatment can significantly affect the outcomes of CS. Traditional assessment methods, with the exception of blood pressure measurement, require a high index of suspicion and frequent reassessment by the clinical team. Electronic medical record‐based detection has been successfully implemented in acute and critical care patients with septic shock and acute kidney injury. In CS, electronic medical record‐based studies have largely focused on using models to predict outcomes in patients with CS, with limited data on electronic medical record‐based tools to assist with either predicting CS or providing real time alerts when escalation or de‐escalation might be indicated. Early detection of CS may be associated with detection of earlier Society for Cardiovascular Angiography and Intervention stages of CS and potentially prevent deterioration to higher stages. In this review, we seek to highlight a blueprint for electronic medical record‐based detection of CS that focuses on reproducibility, convenience, clinical decision support, and research aspects.