For cut-off values of eight points and five points, 87.5% and 88.9%, respectively, of the AUDIT-positives, and 98.9% and 95.1%, respectively, of the AUDIT-negatives were identically identified at retest, with kappa =. On the item level, the reliability, measured by the intraclass correlation coefficient (ICC), ranged between. As the gold standard for detection of an alcohol-use disorder (AUD), we used the Munich-Composite International Diagnostic Interview (MCIDI), which is based on Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, criteria. This definition was supported by diagnostic interviews of 99 screen-negative patients from which no false negatives could be detected. Individuals who scored less than five points in the AUDIT and also tested negative in a second alcohol-related screen were defined as 'negative' (n=6003). For this study, 1109 screen-positive patients received a diagnostic interview. Sensitivity and specificity at a number of different cut-off values were estimated for the sample of alcohol consumers (n=8237). The retest reliability was tested on a subsample of 99 patients, with an intertest interval of 30 days. In total, 10,803 screenings were conducted. Participants recruited from general practices (GPs) in two northern German cities received the AUDIT, which was embedded in a health-risk questionnaire. Our goal was to analyze the retest reliability and validity of the Alcohol Use Disorders Identification Test (AUDIT) in a primary-care setting and recommend a cut-off value for the different alcohol-related diagnoses.

Reliability and validity of audit scale