255 



were targeted at the elderly population. However, we found that the elderly population is not 

 often the focus of medical or epidemiologic research. In lieu of elderly-specific relative risks, 

 we used relative risks for the general adult population. 



Population Attributable Risk 



Relative risks in epidemiologic studies are calculated by dividing the incidence of 

 disease in the exposed group by the incidence in the disease in the non-exposed group.' Some 

 researchers go one step further and calculate the risk, not just for the individual, but to the 

 larger community, by measuring the Population Attributable Risk (PAR). The PAR is the 

 proportion of cases for a given disease that may be attributable to an etiologic factor (e.g. 

 cigarene smoking) and is calculated using the following formula:' 



bir-l)*l 



where b = prevalence of substance use in the population 

 r = relative risk for a given disease 



For each disease or medical condition where a substance abuser's relative risk was 

 reported in the epidemiologic literature, we calculated a Medicare-specific population 

 attributable risk (PAR) using relative risks reported in the studies and the specific prevalence 

 estimates for tobacco, alcohol, and drugs in the Medicare population. Appendix 11 provides 

 a listing of those conditions, the related substances and their PARs. The PAR was computed 

 using the above formula or, where different relative risks were reported by current and former 

 users or by level of consumption, the PAR formula was revised to reflect this specificity. 



13 



