Comparison of PR, QRS, and QT interval measurements by seven ECG interpretation programs

Area of Science:

• Cardiology
• Biomedical Engineering
• Medical Instrumentation

Background:

• Automated electrocardiogram (ECG) interval measurements are presumed more accurate than manual methods.
• However, performance variations among different ECG analysis programs remain under-evaluated.

Purpose of the Study:

• To compare the accuracy and consistency of PR, QRS, and QT interval measurements across multiple commercial electrocardiograph programs.
• To identify systematic biases and the extent of clinically significant errors generated by these automated ECG analysis tools.

Main Methods:

• Digital ECGs were replayed through seven commercial electrocardiographs to generate analog ECGs.
• Measurements for PR interval, QRS duration, and QT interval were extracted and compared against a median of program outputs and population mean values.

Main Results:

• Small, statistically significant systematic biases were observed between different ECG measurement programs.
• While most pairwise comparisons showed similar accuracy for routine ECGs, substantial differences emerged in the frequency and magnitude of clinically significant errors, particularly in challenging cases.
• Variations in error magnitudes (e.g., at the 98th percentile) could differ by a factor of two, with inconsistencies in overestimation and underestimation patterns.

Conclusions:

• Users of digital ECGs must acknowledge the presence of systematic differences between analysis programs.
• Awareness of potential large, clinically relevant errors in difficult ECG tracings is crucial for accurate interpretation.