Use of epidemiological data to predict the occurrence of arteriosclerotic diseases in urban residents

Abstract

From 1983 to 1986, 100,352 urban residents were screened for cardiovascular disease. Glucose and protein levels in urine, total cholesterol, triglycerides, high density lipoprotein cholesterol, blood pressure, and aortic pulse wave velocity were measured. Retinal examinations and electrocardiography were also done. A follow-up survey was conducted two years after screening. A total of 34,895 subjects had no disease and 301 had arteriosclerotic diseases. The occurrence of cardio- and cerebrovascular diseases could not have been predicted from any one abnormal result on a screening test, but might have been predicted from multiple abnormal results, with each factor having its own weight and a unique association with other factors. The risk of disease increased multiplicatively higher rather than additively with the number of abnormal test results. We used Hayashi's quantification method III to determine which combination of abnormal test results was associated with a higher risk of arterio sclerotic diseases, and thus developed a comprehensive indicator of disease risk by grouping several abnormal test results that were affected by different factors. This evaluation system can be used to predict the onset of arteriosclerotic diseases to some degree and can contribute to preventive medicine.