The role of computerized ECG interpretation in clinical trials

The Lipid Research Clinics project consisted of the Prevalence Study and the Coronary Primary Prevention Study. The first study characterized the levels of blood lipids in the population studied, and related these to the presence of coronary artery disease as detected by medical history, symptom analysis, resting ECG and treadmill exercise ECG tests. The prevention trial tested whether a reduction of elevated serum cholesterol would reduce the incidence of coronary events (Table I). Both projects collected resting 12-lead ECGs and treadmill exercise test data on the participants. Instrumentation and methodology were standardized. Rest and exercise electrocardiographic data were recorded on analog magnetic tape. Tapes were sent to the ECG Center where they were digitized, analyzed and also reproduced for visual analysis according to the Minnesota Code of Blackburn. There were two reasons for using both visual coding and computer analyses in these projects. It was thought that computer methodology would enhance the accuracy of the electrocardiographic data, but on the other hand its accuracy had not been adequately documented and its output statements were not in the language of the Minnesota Code, which is traditionally accepted by epidemiologists. The other and more pressing reason for using the Minnesota Code is that the computer programming had not been completed at the start of the project, and visual coding was required in order to provide the ECG data necessary for the randomization of the Prevention Trial participants.