Methods to Estimate the Association between Risk Factors for Mortality and Slowdowns in Reductions in Mortality for Females at Older Ages in Some Developed Countries

Brian Rostron, University of California, Berkeley

Slowdowns in gains in life expectancy for females at older ages have been observed for recent years in some developed countries such as Denmark, the Netherlands, and the U.S.A. I use and expand the indirect method developed by Peto et al. to measure the extent to which previous smoking behavior, as indicated by lung cancer mortality rates, is responsible for these trends, calculating incidence rates and relative risks for causes of death by specific sex- and age-groups. Results suggest that some but not all of the differences between e65 values for females in these and other developed countries are attributable to smoking. Other factors that may account for some portion of the remaining differences in life expectancy include differences between countries in social, economic, and health systems as well as dietary habits such as consumption of animal products and saturated fats.