Electronic Health Record Integration of Predictive Analytics to Select High-Risk Stable Patients With Non–ST-Segment–Elevation Myocardial Infarction for Intensive Care Unit Admission
Aman D. Kansal, Cynthia L. Green, Eric David Peterson, L. Kristin Newby, Tracy Y. Wang, Mark Sendak, Suresh Balu, Manesh R. Patel, Alexander C. Fanaroff · Circulation Cardiovascular Quality and Outcomes · 2021
In the United States, ≈40% of patients with non-ST-segment-elevation myocardial infarction (NSTEMI) who initially present without cardiogenic shock or cardiac arrest are admitted to the intensive care unit (ICU). 1,2Importantly, ICU utilization for these initially stable NSTEMI patients varies substantially between hospitals, and severity of illness on presentation is similar for NSTEMI patients treated and not treated in the ICU, suggesting that ICU admission decisions are largely based on hospital policies and local clinician preferences. 1isk-based ICU utilization-admitting those at the highest risk of developing complications requiring ICU care to the ICU and admitting those at lower risk to a non-ICU setting-has the potential to better align resource use with patient needs.For patients with initially stable NSTEMI, the Acute Coronary Treatment and Intervention Outcomes Network (ACTION) ICU risk score uses data available at the time of hospital presentation to predict risk of clinical deterioration requiring ICU carein-hospital death, cardiac arrest, shock, respiratory failure, heart block requiring pacemaker placement, and stroke. 3Risk scores are infrequently used in clinical practice, 4 but directly embedding a risk score calculator with associated decision support into the electronic health record (EHR) could increase uptake.The goal of this innovation was to achieve risk-based ICU utilization for patients with NSTEMI presenting to the emergency department (ED) by embedding the ACTION ICU risk score into the EHR as a modified best practice advisory (BPA). LOCAL CHALLENGES IN IMPLEMENTATIONMajor challenges in implementation included lack of clinician familiarity with the BPA interface, placement of the risk calculator within ED clinicians' existing workflow, and finding consensus on the appropriate threshold for ICU admission.Although NSTEMI is a relatively common diagnosis, a large number of clinicians see patients in the ED at tertiary academic medical centers, including attending physicians, advanced practice providers, fellows, and residents from multiple specialties.Any one clinician, therefore, sees patients with NSTEMI infrequently and is likely to be unfamiliar with the interface of an NSTEMI risk prediction calculator.Successful deployment of predictive analytics at the point of care requires that the tool overcome clinicians' unfamiliarity by integrating into existing workflows.Since the risk calculator and decision support tool was intended for use by ED clinicians, we designed the tool in consultation with a group of ED clinicians, with a goal of understanding how a new tool could be best incorporated into their workflow.ED