Smoothing techniques for improving the repeatability of the automated diagnosis of inferior myocardial infarction.

This study describes the implementation of novel techniques that have been designed with the aim of improving the repeatability of the diagnostic section of the Glasgow electrocardiographic (ECG) analysis program. Specific reference is made to the agreement in consecutive computer-assisted diagnoses of inferior myocardial infarction (IMI). Inherent repeat variation was estimated in ECG parameters of interest and used in conjunction with smoothing methods to produce a continuous Q-wave index ranging from 0 (no IMI) to 1 (IMI). A decision as to the presence or absence of IMI was then made on the basis of this smooth index. The sensitivity and specificity of the new approach remain unchanged from the conventional procedure when analyzing single ECGs. However, consistency in interpretation of day-to-day and minute-to-minute ECG interpretations was enhanced. Specific reference is made to the agreement between consecutive pairs of computer-assisted diagnoses of ECGs from the same patient with which one or both interpretations was that of IMI.