QRS Detection Algorithm for Telehealth Electrocardiogram Recordings.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Cardiology

Background:

• Electrocardiogram (ECG) analysis is crucial for diagnosing cardiac conditions.
• Existing QRS detection algorithms are primarily optimized for clean clinical data, limiting their effectiveness in telehealth settings.
• Telehealth ECG recordings often suffer from lower quality and artifacts, necessitating specialized algorithms.

Purpose of the Study:

• To develop and evaluate a novel QRS detection algorithm, termed UNSW, specifically designed for analyzing both clinical and telehealth ECG recordings.
• To compare the performance of the UNSW algorithm against established QRS detection methods, Pan-Tompkins (PT) and Gutiérrez-Rivas (GR).

Main Methods:

• The UNSW algorithm extracts features from ECG amplitude and derivatives, applying frequency-based filtering and an adaptive threshold.
• Performance evaluation involved testing on the MIT-BIH Arrhythmia database, a clinical ECG database with added noise (MIT-BIH noise stress test), and telehealth ECG records.
• QRS detection metrics including sensitivity (Se) and positive predictivity (+P) were calculated and compared across algorithms.

Main Results:

• For clean clinical ECG, UNSW demonstrated comparable performance (>99% Se and +P) to PT and GR.
• On noisy clinical ECG data, UNSW achieved >99% Se and a superior +P (98%) compared to PT and GR.
• For telehealth ECG, UNSW significantly outperformed PT and GR, achieving 98% Se and 95% +P.

Conclusions:

• The UNSW algorithm is the first QRS detection method specifically designed and validated for telehealth ECG data, showing superior performance.
• Its effectiveness on both clinical and lower-quality telehealth ECG makes it a valuable tool for managing increasing telehealth ECG analysis workloads.