The influence of editing on the performance of a computer program for serial comparison of electrocardiograms.

In this study, three tests of computer ECG analysis function were conducted: 1) interpretation of serial tracings using the single record analysis program, i.e. without serial comparisons; 2) interpretation of the same tracings using a computer-comparison program; and 3) interpretation of these tracings using not only a computer-comparison program but also an editing program which allows amending of each serial record and insertion of any corrections into the computer data base before the next comparison. There was an over-all reduction of 84% in the number of statement changes needed to produce a correct final report by using the combined comparison and editing program. When this combined program was used, 81% of all computer interpretations required no alteration, --a considerable saving of time in analyzing serial electrocardiographic tracings. Serial comparison was required in in 27--45% of all ECG's taken in four medical centers.