Convergence and discontinuities: Trends in mortality by cause in developed countries, 1950--2000

This work examines mortality trends by class of cause in developed countries. It begins by using principal component analysis to represent mortality trends between 1950 and 2000 for classes of causes and identifies two general groups of causes, one consisting of infectious, digestive, and other diseases for which mortality consistently decreased in the period, and the other including circulatory diseases and malignant neoplasms for which mortality first increased in some countries before generally declining. Countries overall and in most regions display similar mortality trajectories, despite differences in economic and health characteristics.;It then shows that the age pattern of all-cause mortality decline has varied with economic and health development. Mortality rates in high-income countries fell proportionally at middle and older ages during the period, whereas rates at some ages increased in Eastern European countries. Mortality rates proportionally decreased faster at middle ages compared to older ages in countries that were less economically developed at the beginning of the period, a trend due in large part to dramatic reductions in mortality at middle ages from causes such as infectious diseases.;The work then analyzes the causes of observed slowdowns in mortality declines for females at older ages in some developed countries including the U.S. It finds, using a modified version of the indirect method proposed by Peto et al., that smoking exposure accounted for approximately half of the difference between e65 values for females in the U.S. and some other developed countries. A longitudinal mixed effects model finds that dietary factors, specifically animal fat consumption, also had a significant effect on national values for life expectancy for females at older ages. Social and economic factors, such as per capita income and income equality, are found to have little effect on these values.;Finally, the work evaluates claims concerning the relationship between infant and early childhood mortality and mortality at older ages using an age-period-early cohort mortality model. It finds a substantive effect of infant mortality on mortality at later ages that is concentrated at middle ages and is most pronounced for causes with an infectious origin.