254 



APPENDIX I: METHODS 



The methodology for this study relied primarily on the existing epidemiologic 

 literature, as well as on consultations with physicians knowledgeable about substance abuse and 

 related disorders. 



Epidemiologic Literature Search 



We conducted a Medline search of epidemiologic or etiologic studies that identified 

 substance abuse (tobacco, alcohol, or drugs) as a major risk factor for acquiring a given 

 disease/condition. In this search, we selected individual studies, reviews, or meta-analyses, that 

 quantified either a relative risk or an attributable risk, and that generally met the criterion 

 established by the Surgeon General for establishing causality.' Specifically, we favored studies 

 tliat 1) reported stratified relative risks by levels of consumption or by age and sex, 2) 

 demonstrated a dose-response relationship, 3) found diminishment of risk upon cessation of 

 use, and 4) had findings that were generally consistent with other studies. In addition, with 

 a very few exceptions, we used studies that measured the abusers' increased risk of acquiring 

 a disease relative to a non-substance abuser (morbidity studies), as opposed to their increased 

 risk of dying of the disease (mortality studies)."** 



Studies that were reviews or meta-analyses of other studies took priority since they 

 combined the results of multiple studies and often reported a composite attributable or relative 

 risk. If these were not available, we selected large prospective or case-control studies and 

 calculated an average relative risk from these studies. When possible, we selected studies that 



A complete bibliography is available from CASA. 



12 



