Given the heterogeneous nature of substance abuse, it is notable that several predictors of response are independent of the primary drug of abuse or the treatment setting [208]. Although the strength of the relationship of predictor to outcome varies, the following factors have been identified consistently: severity of dependence or withdrawal; psychiatric comorbidity; substance-related problems; motivation (abstinence commitment); length of treatment; negative affective states; cognitive factors; personality traits and disorders; coping skills; multiple substance abuse; contingency contracting or coercion; genetic factors; sleep architecture; urges and craving; self-efficacy; and economic and social factors. Although it is well known that severity of dependence (including polysubstance abuse), serious psychiatric comorbidity, and social problems are associated with poor treatment response, only recently has research examined the efficacy of intervention strategies that specifically address these problems. Adequate treatment of psychiatric comorbidity and improvement in social, economic, and family functioning lead to better treatment outcomes. The development of specific techniques to enhance self-efficacy, motivation, coping skills, and functioning in the community are concrete examples of how the identification of factors associated with positive outcomes has led to the development of new treatments. Despite significant accomplishments, the field is left with many unanswered questions. Although several biologic markers, such as neuroendocrine response and sleep architecture, show promise as outcome predictors, it is not known whether these are critical factors in the initiation of substance use or its progression to dependence. Determining whether biologic markers are epiphenomena reflecting the amount and duration of substance abuse or are fundamental to the pathophysiology of dependence is a matter of urgent concern. With some exceptions, identification of biologic predictors has not led to innovative therapies. One of these exceptions is the development of naltrexone for the treatment of alcoholism, which was based in a solid theoretical rationale and followed by hypothesis-driven experiments. Similar opportunities should emerge from current basic science and clinical research. The application of pharmacogenetic techniques to the field of addiction also holds great promise. As future studies are undertaken, researchers and clinicians must be mindful that differences in outcome predictors across drugs of abuse and treatments may emerge as subgroups of individuals with addictive disorders and new therapies are identified. There is already evidence that early onset alcoholism is associated with poor response under some circumstances, yet may be a predictor of response to targeted pharmacotherapy with ondansetron [64, 112]. As the ability to subtype disorders based on meaningful biologic differences grows, it is anticipated that several relevant outcome predictors that are specific for pharmacotherapy will emerge.