Abstract
OBJECTIVE: To validate the AUDIT test for identifying women with excess alcohol consumption and/or dependency syndrome (DS). DESIGN: Descriptive study to validate a test. SETTING: Two primary care centres and a county drug-dependency centre. PARTICIPANTS: 414 women from 18 to 75 recruited at the clinic. Interventions. Social and personal details were obtained through personal interview, their alcohol consumption was quantified and the AUDIT and MALT questionnaires were filled in. Then the semi-structured SCAN interview was conducted (gold standard; DSM-IV and CIE-10 criteria), and analyses were requested (GGT, GOT, GPT, VCM). 186 patients were given a follow-up appointment three-four weeks later (retest). MAIN MEASUREMENTS: Intra-observer reliability was evaluated with the Kappa index, internal consistency with Cronbach s alpha, and the validity of criteria with indexes of sensitivity and specificity, predictive values and probability quotients. To evaluate the diagnostic performance of the test and the most effective cut-off point, a ROC analysis was run. RESULTS: 11.4% (95% CI, 8.98-13.81) were diagnosed with alcohol abuse (0.5%) or DS (10.9%). The Kappa coefficients of the AUDIT items ranged between 0.685 and 0.795 (P<.001). Internal reliability, with Cronbach s alpha, was 0.932 (95% CI, 0.921-0.941). Test sensitivity was 89.6% (95% CI,76.11-96.02) and specificity was 95.07% (95% CI, 92.18-96.97). The most effective cut-off point was at 6 points. CONCLUSIONS: The AUDIT is a questionnaire with good psycho-measurement properties. It is reliable and valid for the detection of risk consumption and DS in women.