Automated Heart Rate and QRS Axis Classification on Digital Electrocardiograms to Determine Associated Disease

A computational tool was developed in Python to automatically analyze digitized electrocardiograms (ECGs) from the PhysioNet database. Open-source libraries such as NeuroKit, SciPy, and BioScipy were employed to detect fiducial points corresponding to the P wave and QRS complex. Based on these measurements, temporal intervals within the ECG were extracted to estimate heart rate, cardiac rhythm, and electrical axis, thus allowing the identification of potential underlying cardiac abnormalities. The proposed script demonstrated the ability to accurately detect conditions including bradycardia (mild, moderate, or severe), tachycardia, rhythm classification (regular or irregular), and electrical axis orientation (normal, leftward deviation, rightward deviation, or indeterminate). In this preliminary testing phase, a total of 100 ECGs were analyzed. From these, 60% were classified as bradycardia, 10% as tachycardia, 30% as normal frequency; on the other hands, electrical axis was classified as leftward (20%), indeterminate (20%) deviations and 60% as normal, while no rightward deviations were observed. Examples of specific cardiac conditions identified by the algorithm include Left Ventricular Hypertrophy (LVH), Sinus Node Dissociation (SND) and electrical instability (EI). These results were validated by a cardiologist from Hospital General Desiderio Rosado Carvajal in Comalcalco, Tabasco, Mexico, who performed the analyses through conventional diagnostic methods and, in all cases the predictions were exactly equals. The high accuracy observed supports the potential of this computational approach as a technological aid in clinical settings.