Fig. 3.

Event survival curves and hazard ratios for acute coronary events (ACS, CABG, or PCI) within a 3-year observation period in patients with no prior ASCVD (acute coronary syndrome, peripheral arterial disease, stroke, PCI, CABG, or carotid artery intervention) and low (A), borderline (B), intermediate (C), or high (D) ASCVD risk as determined by the pooled cohort equations (PCE). These results suggest an ability of ECG-AI to risk stratify patients independently of the PCE and over shorter time periods. Risk stratification is also observed in patients under the age of 40 (E). Patients were censored upon loss to follow-up if there was no EHR evidence of survival beyond the observation period or death during the observation period. For Hazard Ratios, the number in parentheses represents the 95% confidence interval. All patients were required to have evidence of primary care at Mayo Clinic. “ECG-AI score” is defined as the number of positive results on the three disease-specific models.