The validity of an Australian modification of the AUDIT questionnaire

The Alcohol Use Disorders Identification Test (AUDIT) has been used widely and is reported to be superior to conventional questionnaires in detection of current hazardous and harmful alcohol use. We assessed the validity of a n Australian modification of the AUDIT (the AusAUDIT), which has been emplo yed widely in Australian and New Zealand early intervention programmes. We used a cross-sectional study of 370 subjects from the follow-up phase of a randomized controlled trial of early intervention to reduce hazardous alcoh ol consumption. Scores on the AusAUDIT were compared against 12-month ICD-1 0 diagnoses of harmful alcohol use and dependence, as determined by the Com posite International Diagnostic Interview, and against self-report of alcoh ol consumption exceeding Australian National Health and Medical Research Co uncil (NH&MRC) recommended limits. AusAUDIT had good internal consistency a nd discriminated significantly between persons meeting criteria for ICD-10 alcohol use disorders, and drinkers who did not. At currently recommended c ut-off scores, AusAUDIT detected more than 85% of people meeting criteria f or ICD-10 alcohol use disorders, or drinking over NH&MRC recommended limits , but its specificity was limited (29% in men, and 58% in women for drinkin g over NH&MRC limits). No subset of questions performed as well as the full AusAUDIT in detection of drinking problems, but the alcohol consumption it ems provided a reasonable screen for drinking over NH&MRC limits. We conclu de that AusAUDIT is effective in detecting problematic drinking, but positi ve cases should be confirmed by clinical assessment. The findings illustrat e the need for validation of questionnaire modifications, and the difficult y in increasing test sensitivity without reducing specificity.