食料供給量と死亡率の関連に関する研究

抄録

We studied the relationship between food supplies and mortality rates of several diseases. Data was collected from two Japanese Publications, 1951-1976, Food Balance Sheet and Vital Statistics. In the first stage, the variables, amount of food supplied annually, age-adjusted death rates, are grouped into five categories according to their types of regression lines: (1) monotonous increment, (2) monotonous decrement, (3) concave, (4) convex, (5) other. Statistics calculated by means of principal component analysis represented characteristics of complicated phenomena. After that, canonical correlation analysis was applied to typical foods and mortality rates which were selected because of their history of changing patterns. The first and second canonical correlation coefficients were extremely high. It was suggested that food supply and mortality rates had a significant relationship during 1951-1976 . Furthermore, it was suggested that rice consumption changed concurrently with hypertension, cerebrovascular disease and stomach cancer; and consumption of meat, oil, fat, dairy products, eggs and sugar increased concurrently with the increasing rate of cancers of the colon, rectum, breast and pancreas. The latter diseases have already been linked to Western or European food patterns. The results in this study strongly support previous results which have been reported by other studies.