The estimate of coronary incidence following different case finding procedures

In the pool of two rural population groups made up of 1695 coronary-free men aged 40-59 years at entry and followed-up for 25 years, repeated field examination at 5, 10, 20 and 25 years after enrollment and mortality checking, allowed us to identify 256 first major coronary events (coronary deaths and definite myocardial infarction)--the basic procedure. An extra effort in data collection, involving checking hospital admissions and discharges, inquiries to general practitioners and some postal questionnaires, allowed us to identify 45 extra cases of non-fatal myocardial infarction--the special procedure. The reported incidence of a first major coronary event in 25 years increased in this way by 17.8%, from 15.1% to 17.7%. The 45 cases identified by the special procedure were younger and had more favourable levels of risk factors than the other 256. However the coefficients of six risk factors estimated by the Cox model (age, mean blood pressure, serum cholesterol, cigarette consumption, forced expiratory volume and arcus senilis) were rather similar, and all were significant in the solution that included the cases identified by the basic procedure, compared to the solution that included cases identified by both procedures.