J Med Internet Res. 2026 Feb 11;28:e76087. doi: 10.2196/76087.
ABSTRACT
BACKGROUND: Self-reported, computerized history taking (CHT) may enable efficient collection of medical histories for acute chest pain management.
OBJECTIVE: The primary aim is to determine the diagnostic performance of 4 CHT-derived chest pain risk scores for ruling out 30-day major adverse cardiac events (MACEs) or acute coronary syndrome (ACS). The secondary aim is to assess their impact on patient disposition in the emergency department (ED).
METHODS: This is a prospective cohort study conducted at a tertiary hospital ED in Stockholm, Sweden. Clinically stable adults (≥18 years) with chest pain and an electrocardiogram (ECG) not indicating an acute disease requiring immediate care provided medical histories via a tablet-based CHT program (Clinical Expert Operating System [CLEOS]). CHT data and ECG interpretations and troponin values were used to calculate the History, ECG, Age, Risk Factors, and Troponin (HEART) score, Danderyd HEART (D-HEART) score, Emergency Department Assessment of Chest Pain Score combined with an Accelerated Diagnostic Protocol (EDACS-ADP), and Troponin-only Manchester Acute Coronary Syndrome (T-MACS). The primary outcome was 30-day ACS; the secondary outcome was 30-day MACE (ACS, revascularization, or cardiovascular death).
RESULTS: Among 1000 participants (age: mean 55 years, SD 17 years; 456/1000, 45.60%, women), risk scores could be calculated in 838 (83.80%). Within 30 days, 65 (6.50%) participants experienced ACS, and 72 (7.20%) had a MACE. Negative predictive values were 0.99 (95% CI 0.97-1.00) for both outcomes. Sensitivity for MACE was 0.91 (95% CI 0.81-0.97) for HEART, 0.94 (95% CI 0.86-0.98) for D-HEART, 0.94 (95% CI 0.86-0.98) for EDACS-ADP, and 0.97 (95% CI 0.90-1.00) for T-MACS, with similar results for ACS. As many as 89 of the 528 (16.9%) patients admitted could be reclassified from "nonlow risk" to "low risk." Among reclassified patients, 30-day MACE or ACS occurred in 0-4 cases; miss rates were below 1% for D-HEART (4/416, 0.96%) and T-MACS (2/286, 0.7%), but exceeded 1% for HEART (6/406, 1.5%) and EDACS-ADP (4/346, 1.2%).
CONCLUSIONS: Automated, self-reported CHT provided sufficient data to calculate 4 chest pain risk scores in 838 of 1000 (83.80%) patients with acute chest pain, with score calculation dependent on physician-interpreted ECGs. These CHT-derived risk scores demonstrated good diagnostic performance for ruling out 30-day MACE and ACS. Performance was broadly comparable with prior studies using physician-acquired scores, although suggested safety thresholds were primarily met by D-HEART and T-MACS. The improved safety of D-HEART compared with HEART is likely attributable to the incorporation of serial 0/1-hour troponin testing. Use of CHT-derived risk scores may reclassify a substantial fraction of admitted patients as "low risk," potentially supporting discharge decisions in selected patients, while admission may still be required for non-ACS reasons. However, any gains in discharge rates should be weighed against the possibility of missed events among reclassified patients. Multicenter studies are needed to confirm generalizability, operational feasibility, and safety.
TRIAL REGISTRATION: ClinicalTrials.gov NCT03439449; https://clinicaltrials.gov/ct2/show/NCT03439449.
INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): RR2-10.1136/bmjopen-2019-031871.
PMID:41671573 | DOI:10.2196/76087