An investigation of an excess incidence of heart disease

In a large complex chemical company, it is difficult to distinguish clusters of disease of occupational origin from those due to non-occupational factors and those that have arisen by chance. The mechanical application of statistical procedures to health data to detect occupational hazards will produce many false positives. More refined techniques were used to demonstrate that the incidence of coronary disease at a small site was too high to be attributed to chance alone. An intensive search for a causative factor proved abortive.