Greater than 50% of patients who present to the ED with chest pain are admitted but <10% are ultimately diagnosed with ACS. This pervasive over-triage costs $10-13 billion annually. The HEART Pathway uses a validated clinical decision aid and serial troponin measures to provide real-time decision support to providers. In prior studies, the HEART Pathway decreased hospitalizations, stress testing, and hospital length of stay, without increasing adverse events, demonstrating efficacy at a single site. However, it is unclear what its effectiveness would be in a multicenter study. In preparation for possible widespread implementation of the HEART Pathway, we sought to determine health care utilization and ACS rates for ED patients with acute chest pain across the 4 Carolinas Collaborative health systems (Wake Forest Baptist Health, Duke University, University of North Carolina, and Medical University of South Carolina). We hypothesized that sites that had not implemented the HEART pathway would have higher hospitalization and stress testing rates compared to the one site that had done so. This multi-center observational study uses the Carolina Collaborative infrastructure to pool data from a common electronic health record platform across 4 academic university-affiliated health care systems. We included adult patients (≥21 years old) presenting to the ED from 1/1/2015-12/31/2015 with a chief complaint of “chest pain” or “heart problem,” for whom the provider ordered troponins. We excluded those with a diagnosis code of ST-segment elevation myocardial infarction (STEMI). We anticipate a total of 20,000 patients being eligible based on prior work. The Carolinas Collaborative has an existing ontology of variables including demographics, diagnoses, medications, and selected laboratory values. We assessed rates of hospitalization, stress testing, angiography, recurrent ED visits, readmissions, death, myocardial infarction, revascularization, and length of stay to determine the potential impact of implementing the HEART Pathway at included sites. Further variables were extracted and definitions harmonized across sites. Rates of hospitalization and stress testing from the Heart Pathway site, Wake Forest Baptist Health (WFBH) will be compared to other sites in the Collaborative. We will calculate p values with alpha =0.05 for these two comparisons with and without propensity score adjustment for comorbid conditions. We will report basic demographics, including age, sex, and race. We will report unadjusted rates of hospitalization and stress testing for non-STEMI patients with chest pain at the Heart Pathway site versus non-HEART pathway sites. After controlling for patient severity and comorbidity, we will determine whether patients at non-HEART pathway sites are more/less likely (odds ratios) to be admitted or to have stress testing. In this comprehensive multicenter EHR data analysis, we anticipate that rates of hospitalization and stress testing among eligible patients will vary across sites. We will be able to determine whether implementing the HEART Pathway is associated with more or less hospitalization and stress testing.