MODELING ALL-CAUSE MORTALITY - PROJECTIONS OF THE IMPACT OF SMOKING CESSATION BASED ON THE NHEFS

Objectives. A model that relates clinical risk factors to subsequent m ortality was used to simulate the impact of smoking cessation. Methods . Survivor functions derived from multivariate hazard regressions fitt ed to data from the first National Health and Nutrition Examination Su rvey (NHANES I) Epidemiologic Followup Study, a longitudinal survey of a representative sample of US adults, were used to project deaths fro m all causes. Results. Validation tests showed that the hazard regress ions agreed with the risk relationships reported by others, that proje cted deaths for baseline risk factors closely matched observed mortali ty, and that the projections attributed deaths to the appropriate leve ls of important risk factors. Projections of the impact of smoking ces sation showed that the number of cumulative deaths would be 15% lower after 5 years and 11% lower after 20 years. Conclusions. The model pro duced realistic projections of the effects of risk factor modification on subsequent mortality in adults. Comparison of the projections for smoking cessation with estimates of the risk attributable to smoking p ublished by the Centers for Disease Control and Prevention suggests th at cessation could capture most of the benefit possible from eliminati ng smoking.