THE AUDIT QUESTIONNAIRE - CHOOSING A CUTOFF SCORE

The Alcohol Use Disorders Identification Test (AUDIT) is a 10-item que stionnaire designed by the World Health Organization to screen for haz ardous alcohol intake in primary health care settings. In this longitu dinal study we examine its performance in predicting alcohol-related h arm over the full range of its scores using receiver operating charact eristic analyses. Three hundred and thirty ambulatory care patients we re interviewed using a detailed assessment schedule which included the AUDIT questions. After 2-3 years, subjects were reviewed and their ex perience of alcohol-related medical and social harm assessed by interv iew and perusal of medical records. AUDIT was a good predictor of both alcohol-related social and medical problems. Cut-off points of 7-8 ma ximized discrimination in the prediction a trauma and hypertension. Hi gher cut-offs (12 and 22) provided better discrimination in the predic tion of alcohol-related social problems and of liver disease or gastro intestinal bleeding, but high specificity was offset by reduced sensit ivity. We conclude that the recommended cut-off score of eight is a re asonable approximation to the optimal for a variety of endpoints.