Improving Mortality Projections By Incorporating Cohort Smoking Histories

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): This research is aimed at improving mortality projections in the United States by incorporating information on a cohort's history of cigarette smoking. In terms of attributable risk, cigarette smoking accounts for more deaths in the United States than any other risk factor. Previous research has demonstrated powerful effects of a cohort's smoking history on its subsequent mortality. We intend to incorporate these risks into the Lee-Carter projection model, which has not previously been developed in a way that accommodates both period and cohort influences. Even modest improvements in mortality projections will have major implications for the fiscal future of the United States; movement from the high to the low mortality projection series of the Social Security Administration changes the fiscal balance of the system by $4.1 trillion, the largest impact of variation in any underlying projection parameter. An important byproduct of the analysis will be an index of period mortality conditions that is purged of the disturbances introduced by disparate cohort smoking patterns. Much of the smoking behavior that is relevant during the projection period 2005-2050 has already been observed. For those cohorts for whom little or no smoking behavior has been observed, we will employ an age/period/cohort model of smoking behavior that is estimated on 25 National Health Interview Surveys that collected comparable smoking data from 1965 to 2003.