Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)

Graduate Group


First Advisor

Michel Guillot

Second Advisor

Samuel Preston


Assessing the Impacts of Smoking and Obesity on Mortality and Morbidity

in the United States

Bochen Cao

Michel Guillot

Smoking and obesity are two leading risk factors that account for the current US lags in advances in health and longevity compared to other wealthy nations. This dissertation consists of three independent studies of the impacts of smoking and obesity on population health outcomes among older adults in the United States. The first study estimates the effects of the recent smoking decline on future all-cause mortality, based on the association observed between cohort smoking pattern and cohort death rates from lung cancer. We find that change in smoking is expected to have a large effect on U.S. mortality. However, compared to men, women are expected to have smaller increase in future life expectancy, because of their lagged decline in smoking. The second study extends the first one and estimates the joint effects of smoking and obesity on both mortality and disability. A multistate lifetable approach is applied to estimate the transition rates between different health states, which are in turn projected up to 2040 using a modified Lee-Carter model that incorporates cohort histories of smoking and obesity. The results indicate men and women both are expected to experience compression of disability, with increasing proportions of their future gain in life expectancy likely to be disability free. Nevertheless, due to gender difference in smoking history and in response to obesity, men will likely to have an advantage over women in health improvement in the next three decades. The third study investigates the direct effects of both obesity and weight change on mortality. A dynamic causal model is applied to adjust for reverse causality that is attributable to illness-associated and smoking-associated weight loss in a time-dependent fashion, a problem that prior studies often fail to adequately handle. This study demonstrates that both the confounding by illness and by smoking lead to overestimates of the effects of being underweight and of weight loss, but underestimates the effect of being obese. Moreover, not only being underweight or severe obese, but also sharp weight fluctuations are associated with excess mortality risk.