Forecasting Health Expectancy for Old Adults in the United States Using Cohort Smoking and Obesity History

Bochen Cao, University of Pennsylvania

Forecasting health expectancy is critical in studying population health, to which smoking and obesity are the two most destructive factors. Smoking is found to be responsible for shortening life expectancy both with and without disability, while obesity is associated with higher risk of disabling but non fatal conditions. The prevalence of obesity is expected to increase and smoking is expected to continue declining in the next decades. This study forecasts health expectancy for the U.S. population above age 55, using a multistate life table approach and a modified Lee-Carter model to forecast future transition rates. Cohort history of smoking and obesity are included to capture their impacts on mortality and morbidity, and to avoid implausible long-term forecasts that may be produced by the Lee-Carter model. The results suggest that smoking leads to loss in life expectancy and obesity leads to absolute gain in life expectancy with disability for both sexes.