Reproducibility of analysis of the electrocardiogram in epidemiology using the 'Minnesota code'.

Use of the electrocardiogram is necessary to advance knowledge of the epidemiology of ischaemic heart disease. Analysis of the electrocardiogram in epidemiological practice, however, may be difficult since the clinical method of interpretation cannot be used because of its deficiencies in standardization and reproducibility (Blackburn, 1968; Burgess, Fejfar, and Kagan, 1963; Davies, 1958; Higgins, Kannel, and Dawber, 1965; Rose, 1965; Rose and Blackburn, 1968; Scandinavian Committee on ECG Classification, 1967). Therefore, the epidemiologist must avoid any interpretation of the electrocardiogram. He has to give a description of it which should be as reproducible as possible and yet relevant to the diseases under study. Such a description has to be based on a set of measurements and not on the qualitative evaluation of certain patterns. The method of classification known as the Minnesota Code (Blackburn, Keys, Simonson, Rautaharju, and Punsar, 1960; Blackburn, 1968) answers this requirement. Since 1966 the Groupe d'Etude sur l'Epidemiologie de l'Atheroscl6rose (GREA) has used a particular protocol for this code; details have been published elsewhere* (Elgrishi, Ducimetiere, Richard, and Gelin, 1969). The present study examines the reproducibility of an analysis carried out in this way.