Outcome variables and their assessment in alcohol treatment studies: 1968-1998

Background: This article provides a historical overview of the assessment of outcome variables in alcohol treatment studies that were first published between 1968 and 1998. The review focuses on changes over time in (1) the number of outcome variables and the number of different types of outcome variables assessed, (2) the likelihood of assessing specific types of outcome variables, (3) the methods used to assess outcome variables, and (4) the status of outcome assessment in more recent studies first published between 1990 and 1998. Methods: Reports of 357 alcohol treatment trials with two or more treatment/control groups were coded with respect to the number and types of outcome variables assessed, sources of outcome data, and methodological aspects of outcome assessment. Results: Although the number of outcome variables assessed in studies, on average, did not increase significantly over time, the number of different types of outcome variables did increase. An expected decrease in the assessment of categorical abstinence was not found, but another categorical variable, global ratings of drinking improvement, did decrease over time. More recent studies were more likely to assess such continuous variables as time abstinent, alcohol consumption, time drinking, dependence symptoms, and drinking-related problems. Physiological markers of drinking/alcohol misuse also were assessed more frequently in later years. Some aspects of outcome assessment methods exhibited improvement over time; validity data were more likely to be provided or cited, and self-reports of drinking behaviors were more likely to be corroborated in studies first published in more recent years. However, the percentages of studies that provided/cited reliability data for outcome measures, indicated that follow-up data collectors were not affiliated with treatment and were unaware of respondents' treatment conditions, and reported that respondents were alcohol-free at follow-up did not rise significantly over time. Conclusions: Although the methods of outcome assessment improved between 1968 and 1998, much room for improvement remains.