92 JOURNAL OF STUDIES ON ALCOHOL / SUPPLEMENT NO. 15, 2005 92 Section 4: Assessment in Combined Pharmacotherapy and Behavior Therapy Trials of Alcoholism Treatment: Issues and Domains* DENNIS DONOVAN, PH.D. Alcohol and Drug Abuse Institute, University of Washington, 1107 NE 45th Street, Suite 120, Seattle, Washington 98105–4631 A SSESSMENT REPRESENTS a critical element in the development of clinical research protocols and in clini- cal treatment planning (Connors et al., 1994; Donovan, 1999, 2003a,b). It seems that it would be relatively simple in the arena of alcoholism to determine an outcome: drinking or not, abstinent or not. However, this “either/or” perspective belies the complexities involved in the assessment process. It is no longer appropriate to end one’s assessment of an individual with only a dichotomous “yes/no” determination with respect to a diagnosis or an outcome. The field has grown well beyond this binary categorization. There is a need for a much more detailed and differentiated view of drinking behavior and its antecedents and consequences. It is also important to be able to track the changes in drinking behavior and alcohol intake across time and to predict their course from information gained at the outset of a study or entrance into treatment and from process measures collected during the course of treatment. Different types of treat- ment—behavioral, pharmacological or their combination— may be expected to exert their effects on different aspects of drinking behavior, thus requiring multiple measures of the targeted behavior. The set of articles in this section reflect the increased sophistication and breadth of coverage involved in the as- sessment process in clinical research of combined behav- ioral and pharmacological interventions. In the first article, Gastfriend et al. present the rationale for the baseline as- sessment in COMBINE. They provide a description of the process one should follow to determine the appropriate do- mains for assessment and, once having done that, to choose among available measures those that will best measure these domains with both reliability and validity. Given the focus on medications in COMBINE, there is increased attention to physical examinations, medical conditions and biologi- cal markers of liver function. This extends well beyond the assessment of such variables typically found in studies of behavioral interventions but is important to monitor safety parameters with the two medications being used in COM- BINE. Many of the measures included at baseline will pro- vide a marker of participants’ function or status at intake into the study, against which change in function and behavior can be determined. Finally, many of the measures have been included to evaluate their role as mediating or moderating variables. A goal for the assessment was that it be compre- hensive and capable of serving these multiple needs yet not exceedingly long. We feel that this goal has been met. The second article, by Anton and Randall, focuses more specifically on the assessment of drinking behavior and al- cohol consumption. The question is: What are the best mea- sures of drinking to be included in COMBINE? The answer to this question has a direct bearing on the primary out- come measure(s) that will be used to determine the effi- cacy or effectiveness of an intervention. Consideration must be given to the purported mechanism of action and goals of the treatments as well as type of outcomes that would have clinical meaning. Thus the answer to this question will vary from one study to the next based on the specific parameters of drinking to be affected by the particular in- terventions being evaluated. This is exemplified by the fact that the primary outcome measures used in COMBINE dif- fer from those used in Project MATCH (Babor et al., 1994). The decision about the measures of drinking to include in COMBINE was based on the hypothesized effects of acamprosate and naltrexone on drinking. Abstinence-pro- moting medications or therapies require abstinence-based outcome variables (e.g., percentage of days abstinent, time to first drink, cumulative days abstinent). Relapse preven- tion therapies and/or medications that reduce alcohol’s re- inforcing properties require drinking variables that measure moderation or reduction (e.g., drinks per day, number of heavy drinking days, time to first and successive heavy drinking days). There continues to be concern about the validity of self- report of drinking among alcohol-dependent individuals. Two methods have been used to corroborate such reports. *This work was supported by National Institute on Alcohol Abuse and Alcoholism Cooperative Agreement U10AA11799.