[Age-, period-, and birth-cohort-specific effects on the male proportion in Japanese newborns and projections for male proportion for 20 periods (2008-2027)]

[Gender Differences in Projected Life Expectancy in Japan (2023-2047) Determined by Bayesian Age-Period-Cohort Analysis]

OBJECTIVES

In this study, we aimed to (1) determine the effects of age, period, and cohort on mortality rate trends between 1958 and 2012 in Japan and (2) assess gender differences in projected life expectancy (LE) for the 2023-2047 period.

METHODS

A time trend study was conducted using age-period-cohort (APC) analysis. A Bayesian APC model was fitted to describe mortality rate trends for the 1958-2012 period and to project mortality rates for 2023-2047. LE was predicted by Chiang's method using projected mortality rates.

RESULTS

Age, period, and cohort effects showed similar patterns between males and females. As time passes, gender differences in projected LE were larger among individuals over 65 years than among those under 65 years. Time series change rates of the extension of projected LE after excluding specific causes of death showed the following: smaller extension of projected LE in males in terms of mortality risk from malignant neoplasms, heart diseases, pneumonia, and accidents (under 65 years) and in females in terms of mortality risk from heart diseases, cerebrovascular diseases, and suicide (over 65 years).

CONCLUSIONS

Gender differences in projected LE are expected to be smaller before middle age and to be larger among seniors. These projected gender differences stem in part from the lower mortality risk among men than among women from malignant neoplasms, heart diseases, pneumonia, and accidents (under 65 years), and among women compared to men from heart disease, cerebrovascular disease, and suicide (over 65 years).