The utility of self-report and biological measures of alcohol consumption in alcoholism treatment outcome studies

Despite the significant advances of the past 20 years, serious questions have been raised about the value of alcoholism treatment outcome research due to the purported invalidity of alcoholics' self-report data for which biological measures of alcohol consumption have been suggested as a solution. The literature does not support widespread skepticism of alcoholics' self-reports or the existence of a systematic underreporting bias except when alcoholics have a positive blood alcohol level. Available evidence shows good reliability and validity for reports of hospital/jail stays and frequency of drinking/abstinence; good reliability and modest agreement with collateral informants for measures of problem severity and alcohol dependence symptoms. For amount consumed, reliability is good for outpatient but not for inpatient or residential alcoholics and validity in the form of agreement with collateral reports is not good except in problem drinkers at followup. Unfortunately, currently available biological markers do not solve the problems of measuring alcohol consumption in outcome studies. Markers correlate only modestly with self-reported consumption, are affected by factors other than alcohol consumption, show large individual differences in response to variations in alcohol intake, have an overly long half-life in certain cases, provide a specific but not very sensitive indicator of poor outcome, and may present compliance problems especially among chronic alcoholics. Nonetheless, markers can be used as part of a convergent validity approach and as a measure of negative alcohol-related health consequences. Suggestions for current practices in the field and for future research directions are provided.