A retrospective external validation study of the HEART score among patients presenting to the emergency department with chest pain

Matthew Jay Streitz, Joshua James Oliver, Jessica Marie Hyams, Richard Michael Wood, Yevgeniy Mikhaylovich Maksimenko, Brit Long, Robert Michael Barnwell, Michael David April

Emergency physicians must be able to effectively prognosticate outcomes for patients presenting to the Emergency Department (ED) with chest pain. The HEART score offers a prognostication tool, but external validation studies are limited. We conducted an external retrospective validation study of the HEART score among ED patients presenting to our ED with chest pain from 1 January 2014 to 9 June 2014. We utilized chart review methodology to abstract data from each patient's electronic medical record. We collected data relevant to each of the five elements of the HEART score: history, electrocardiogram (ECG) interpretation, patient age, patient risk factors, and troponin levels. We calculated the diagnostic accuracy of the HEART score (0-10) for predicting the primary outcome of major adverse cardiac events (MACE) over 6 weeks following the ED visit (coronary revascularization, myocardial infarction, or mortality). We randomly selected 10% of patient charts from which a second investigator abstracted all data to assess inter-rater reliability for all study variables. Of 625 charts reviewed, we abstracted data on 417 (66.7%) consecutive patients meeting study inclusion criteria. Thirty-one (7.4%) of these patients experienced 6-week MACE. We observed no instances of MACE within 6 weeks among subjects with a HEART score of 3 or less. The area under the receiver operator curve (AUROC) is 0.885 (95% confidence interval 0.838-0.931). Patients with a HEART score ≤3 are at low risk for 6-week MACE. Hence, these patients may be candidates for outpatient follow-up instead of inpatient admission for cardiac risk stratification.