Projecting the Effect of Changes in Smoking and Obesity on Future Life Expectancy in the United States Samuel H. Preston University of Pennsylvania Retirement Research Consortium Conference Washington, D.C. August 2, 2012 In Collaboration with Andrew Stokes and Bochen Cao, University of Pennsylvania and Neil Mehta, Emory University
Figure 1. Trends in Smoking and Obesity in the United States Sources: Cigarette consumption data per adult per year are extracted from Tobacco Situation and Outlook Report Yearbook. U.S. Department of Agriculture, October Obesity data based on measured body mass index in NHANES from 1960 to 2010.
Contrasting the Dynamics of Smoking and Obesity in the United States
Table 1A. Transition Probabilities across Categories of Body Mass Index, US Adults Ages 25-84, BMI in 2008 BMI in 1998NormalOverObese 1Obese 2 Normal Overweight Obese Class Obese Class
Table 1B. Differences in Transition Probabilities (Standard Errors) between and , US Adults Ages BMI in later year BMI in early yearNormalOverObese 1Obese 2 Normal * (0.019) 0.044* (0.016) 0.011* (0.004) 0.002* (0.001) Overweight (0.011) * (0.017) (0.015) 0.022* (0.010) Obese Class (0.004) * (0.023) (0.019) 0.100* (0.035) Obese Class (0.001) * (0.013) * (0.035) 0.117* (0.047) * p<0.05
Table 1C. Differences in Transition Probabilities (Standard Errors) between and , US Adults Ages BMI in later year BMI in early yearNormalOverObese 1Obese 2 Normal (0.020) (0.017) * (0.005) (0.001) Overweight (0.013) (0.022) (0.021) (0.014) Obese Class (0.004) (0.024) (0.022) (0.031) Obese Class (0.001) (0.014) (0.039) (0.053) * p<0.05
Figure 2A. Actual and Projected Trends in Body Mass Index (Males)
Figure 2B. Actual and Projected Trends in Body Mass Index (Females)
Data and Methods for the Analysis of the Mortality Risks of Obesity Data: Baseline data are obtained by pooling NHANES 3 ( ) and NHANES continuous waves Mortality data are obtained from the National Death Index. Sample size: 21,554 respondents and 2,976 deaths Method: Discrete hazards model, controlling for sex, race/ethnicity, educational attainment and smoking
Baseline Model Parameters for Regression Equation Predicting Mortality as a Function of Current Obesity and Obesity at Age 25 CovariatesCoefficientsSEtP>t Female Age Age 25 Overweight (BMI 25-30) Obese (BMI>30) Baseline Obese class 1 (BMI 30-35) Obese class 2 (BMI>35) Obese class 1 x Age Obese class 2 x Age Constant
Figure 3A. Effect of Projected Trends in Body Mass Index on Age-Specific Death Rates (Males)
Figure 3B. Effect of Projected Trends in Body Mass Index on Age-Specific Death Rates (Females)
Table 5B. Changes in Life Expectancy at Age 40 Resulting from Changes in Obesity YearMalesFemales
Figure 4. Trends in Smoking and Lung Cancer Mortality in the United States Sources: Cigarette consumption data: International Smoking Statistics (2011); Lung cancer mortality data: National Center for Health Statistics.
Figure 5. U.S. Male Lung Cancer Mortality by Cohorts
Figure 6. U.S. Male Lung Cancer Mortality by Period
Equation for Projecting the Mortality Effects of Smoking
Figure 7. Mean Number of Years Spent as a Cigarette Smoker before Age 40 by Cohort
Figure 9A: Effect of Projected Trends in Smoking on Age- Specific Death Rates (Males)
Figure 9B: Effect of Projected Trends in Smoking on Age- Specific Death Rates (Females)
Table 5A. Changes in Life Expectancy at Age 40 Resulting from Changes in Smoking YearMalesFemales
Figure 10A. Effect of Projected Trends in Smoking and Body Mass Index on Age-Specific Death Rates (Males)
Figure 10B. Effect of Projected Trends in Smoking and Body Mass Index on Age-Specific Death Rates (Females)
Table 5C. Changes in Life Expectancy at Age 40 Resulting from Changes in Smoking and Obesity Changes in Smoking Alone Changes in Obesity Alone Changes in Smoking and Obesity YearMalesFemalesMalesFemalesMalesFemales
Projected Changes in Life Expectancy at Age 40, MalesFemales SSA projected increase* Gain from reduced smoking Penalty from higher obesity *Source: Felicitie C.Bell and Michael L. Miller Life Tables for the United States Social Security Area Social Security Administration Actuarial Study No Washington, D.C.
Table 6. Sensitivity of Results to Changes in Procedures Change in ProcedureEffect on Life Expectancy at age 40 Relative to Main Projection MaleFemaleMaleFemaleMaleFemale Use of mortality rates with no control except age & sex Use of mortality rates without inclusion of BMI at age Use of uncorrected data on obesity at age Use of alternative series translating lung cancer into all-cause mortality