Background: Patients admitted to the observation unit with chest pain have a high utilization of cardiac diagnostic testing (e.g. stress testing). Current practice does not incorporate the consistent use of a risk-stratification tool. HEART score is a validated tool that accurately identified low-risk chest pain patients in the Emergency Room in past studies. The object of this study was to determine if the HEART score could identify patients on a medical observation unit who would not benefit from further cardiac testing.
Methods: This was a retrospective, observational cohort study at a single academic medical center observation unit. Over a 12-month period, study data were collected and managed using REDCap. Data collection was done via manual chart review of patients with a chief complaint of chest pain. The HEART score was calculated using the history and physical, ED notes, EKG, and laboratory values. The endpoint was major adverse cardiac event (MACE), which was defined as acute myocardial infarction (AMI), percutaneous coronary intervention (PCI), surgical revascularization, or death within six weeks of the index visit. The type and results of cardiac diagnostic testing were also abstracted. Specificity, sensitivity, and frequency of cardiac testing were used for analysis.
Results: There were 282 patients identified within the time period, 151 (54%) female and 131 (46%) male. The groups were dichotomized into low risk (HEART score 0-3) and elevated risk (HEART score >3). There were 115 (41%) patients in the low-risk group and 167 (59%) patients in the elevated-risk group. The MACE rate for the low-risk group was 2 (1.7%) and 17 (10.1%) for the elevated-risk group, (p=.005). The two MACE events in the low-risk group were PCI and 17 MACE events in the elevated-risk group included 6 AMI, 11 PCI, 1 CABG, and 1 death. The death in the elevated-risk group was non-cardiac due to sepsis related to peritonitis/appendicitis. The HEART score sensitivity was 89% (95%CI: 66-98%) and specificity 43% (95%CI: 37-49%). Stress testing was performed 70/115 (60.9%) for the low-risk group and 102/167 (61%) for the elevated-risk group (p=.89). The low-risk group had 58/68 (85.3%) normal stress test results and 10/68 (14.7%) abnormal, 2 were not completed. The elevated-risk group had 67/102 (65.7%) normal stress test results and 35/102 (34.3%) abnormal results. There was a significant difference in the stress test results by group, (p=.005).
Conclusions: The HEART score identified a patient population at low risk for adverse cardiac events who received diagnostic testing at a similar rate as the elevated risk group. Incorporation of the HEART score into the admission protocol will provide decision support for clinicians and help improve decision-making with regards to cardiac testing for chest pain. This could reduce unnecessary diagnostic testing and decrease wasteful spending.