The Use of Artificial Intelligence in ECG Interpretation in the Outpatient Setting: A Scoping Review

Cardiovascular disease remains the leading cause of death across all demographics globally. The 12-lead ECG is a key diagnostic tool for early detection; however, its interpretation is complex and prone to error, particularly in outpatient settings. AI, especially deep learning models such as convolutional neural networks (CNNs), has shown potential in improving ECG interpretation, but its effectiveness outside hospital environments remains underexplored. This scoping review summarizes the use of deep learning algorithms for ECG interpretation in outpatient settings, focusing on diagnostic accuracy and clinical utility. A systematic, comprehensive literature search was conducted using EMBASE, Ovid MEDLINE, and Web of Science. Included studies examined AI-based ECG interpretation in outpatient care. Studies set in emergency departments, case reports, reviews, and purely theoretical AI models were excluded. Data on study design, AI methods used, and clinical outcomes were extracted. Study quality was assessed using the Joanna Briggs Institute Critical Appraisal tools. Findings from this review demonstrate that AI-assisted ECG interpretation, particularly using deep learning, improves diagnostic accuracy and enables earlier detection of cardiac conditions. These tools show the greatest potential in resource-limited outpatient settings, helping reduce misdiagnosis rates and easing the burden on specialists. However, limitations persist, including small and homogenous sample sizes, and a tendency for some algorithms to overdiagnose. Issues such as data standardization, lack of diversity in training datasets, and limited external validation by cardiologists hinder real-world application. Larger, more representative datasets and robust clinical testing are essential to enhance the generalizability and reliability of these tools. AI integration into outpatient ECG interpretation holds significant promise for improving cardiovascular diagnostics, particularly in underserved areas. Future research should prioritize real-world validation across diverse populations and emphasize cardiologist oversight to ensure clinical safety and effectiveness.