229 



and costs. However, mortality SAFs, which measure smokers' risk of dying of a disease, arc 

 different than morbidity SAFs, or smokers' risk of contracting a disease. Thus, mortality SAFs 

 cannot be used reliably for estimating morbidity or hospital costs. 



Recognizing the shortcomings of using mortality SAFs in estimating health care costs, Rice 

 (1986) developed a different methodology for identifying smokers' attributable risk of using 

 health services using NHIS data. For people who had neoplastic, circulatory, and rcspiratory 

 diseases. Rice analyzed the use of hospital days and physician visits by smokers compared to 

 non-smokers by age and sex. From these ratios, Rice was able to calculate morbidity attributable 

 risks which she then applied to hospital and outpatient expenditures for these diseases to estimate 

 annual smoking-related health care costs. While not as disease-specific as the mortality-based 

 studies. Rice's methodology set a standard for estimating aimual health care costs associated with 

 smoking. 



In addition to these point-in-time estimates, others have studied the lifetime costs of smoking. 

 For example, Manning concludes that the cumulative impact of excess medical care required by 

 smokers at all ages far outweighs shorter life expectancy. Hodgson using survey data from the 

 National Medical Expenditures Survey (NMES) and the National Health Interview Survey 

 (NHIS), breaks down the differences in smokers and non-smokers expenditures by payer, 

 revealing that over the long term, payers that cover the younger age groups (i.e. private 

 insurers and Medicaid) bear a greater burden of smokers' costs than does, for example. 

 Medicare. These studies have current relevance in countering the arguments that measures 

 designed to reduce smoking (e.g., increased cigarette tax) will, in fart, increase health care 



costs. 



-58- 



