The method of adjustment of death rate as presently employed is inadequate for comparing death rates for disease with low frequency and different age distributions. A new death rate index, including a kind of inverse sine transformation, was applied to standardize death rates for chronic disease such as Aplastic Anemia, Chronic Glomerulonephritis, Idiopathic Cardiomyopathy, Paralysis Agitans, Pulmonary Fibrosis, and Systemic Lupus Erythematosus, and compared to standardizations presently employed, such as the direct and indirect methods. The problem of comparing death rates was discussed with relation to this method of standardization of death rates for dieseases with differing sex and age distributions. The applicability of this index to the detection of clustering of dieseases was also discussed.
The death rate from peptic ulcer decreased gradually during 25 years after the Second World War in Japan (Vital Statistics, Ministry of Health & Welfare). The difference in the death rate during the period was analysed every five years (5 year difference), between 46 prefectures (inter-prectural difference), by means of analysis of variance (two-way layout). The decrease in the death rate (5 year difference) was significant in all age groups of 15-74 year old males, and 15-79 year old females (p <0.05). The decrease was most significant for males aged 40-69 years. The inter-prefectural difference was most significant for males aged 40-69 years as well. The difference was greater for males than for females in the 15-64 year age groups. It was assumed that steady and extensive development of the screening system for gastric cancer and peptic ulcer for working people was the main factor leading to the age and sex characteristics of the decrease and inter-prefectural difference in the death rate.
Levels of screening tests were studied by one hour values of GTT on diabetes melitus. The results were as follows:1) Mean blood sugar values of both males and females were increased by aging, and the value for females was slightly higher than for males in all age groups.2) The frequency distributions of the values of GTT showed normal distribution types for both sex and age.3) When the sujbects were screened uniformly by the value of 220 mg/dl of blood sugar, 4.2% of the 5961 males were selected for detailed examination. However, when considering the age factor, subjects were screened by M + 2 SD values, 2.5% were selected for detailed examination.4) We tried similar tests on workers in a certain company. In this test, the value of 170 mg/dl blood sugar was used, and 34.5% of 339 males were selected for detailed examination. But when the subjects were screened by M + 1 SD values of blood sugar, and when those values were corrected for age class, 26.5% were selected for detailed examination. The fre quency of abnormality detected with the M + 1 SD method was 8% less than with the uni form method.5) We checked the efficacy of this method by evaluating laboratory tests. The method that corrected blood sugar value by M + 1 SD, within age classes, was slightly inferior to the uniform method in sensitivity score, but was superior in specificity score. From these results, we determined the M + 1 SD method to be suitable as a screening method for diabetes melitus.