Comparison of Short Questionnaires on Alcohol Drinking Behavior in a Nonclinical Population of 36-Year-Old Men and Women

Abstract

Little is known on the diagnostic characteristics of brief questionnaires on alcohol drinking behaviors. This report investigates the determinants of three short alcohol questionnaires and investigates their diagnostic utility as screening tools for alcohol-related problems in a general population from The Netherlands. This report uses cross-sectional data obtained in the year 2000 from 36-year-old healthy male (N = 166) and female (N = 165) volunteers who reported to drink alcohol at least occasionally. Since they were 13-years-old these volunteers have been members of the Amsterdam Growth And Health Longitudinal Study, which started as a school-based study in 1977. Among many other variables, quantity–frequency questions (QF), the CAGE questionnaire, and a question on the highest number of alcoholic units consumed on one occasion during the previous month (MAX) were asked. The sensitivity, specificity, Cohen's kappa, and diagnostic odds ratio of QF, CAGE, MAX, and combinations of these three brief questionnaires were calculated using a 7-item questionnaire on alcohol-related problems as reference. Both in women and men, the prevalence of most alcohol-related problems and of a high QF, CAGE, and MAX was low. QF, CAGE, and MAX, as well as all possible combinations of the three questionnaires, were poor in detecting last-year alcohol-related problems. The CAGE appeared to perform worse than the even shorter and easier-to-interpret QF and MAX. In this healthy population of 36-year-old men and women, using the QF, MAX, and especially the CAGE questionnaire as screening instruments for alcohol-related problems resulted in many false positive and false negative classifications.