A Component Analysis of Some Health Indicators

Abstract

Through the procedure of principal component analysis, the authors investigated the utilities of five health indicatores; infant death rate, corrected death rate, correted PMI (proportional mortality indicator), Uematsu's health indicator (normal deviate for the mortality) and correc ted life lost in the regions with various scales of population. The results obtained were as follows : 1) In general, infant death rate was representative health indicator in small regions such as city, town or village. On the other hand, Uematsu's indicator was very proficient in the regions with the middle or large populations. 2) As for the structure of those indicators mentioned above, corrected PMI and corrected life lost had very resemble aspects each other, and infant death rate, corrected death rate and Uematsu's indicator showed the highly similar profiles in each region. 3) The value of linear combination in the primary principal component was valid for evaluation of health levels in the middle or large regions.