Clinical cutoffs of the Alcohol Use Disorder Identification Test for identification of DSM-5 alcohol use disorder: A psychometric evaluation in treatment-seeking patients

Abstract

The AUDIT is considered a gold-standard self-rating instrument used in various settings for the identification of alcohol use disorders. Although the AUDIT is widely used, there is not yet a consensus concerning the clinical cutoff for the identification of an alcohol use disorder according to the DSM-5. The current study aimed to investigate the optimal gender-based cutoffs on both AUDIT and AUDIT-C for different levels of AUD severity according to DSM-5. Methods: A sample (n=271) was collected from two clinical trials including adult individuals with no AUD, and AUD recruited at a specialized addiction clinic, in Stockholm, Sweden. Receiver operating characteristic curves were estimated for AUDIT and AUDIT-C scores on AUD severity levels using dichotomized variables for each diagnostic level. Optimal cutoff scores were decided based on the measures of sensitivity, specificity and Youden’s Index. Results: The optimal cutoff score for both men and women was 7 on the AUDIT as the diagnostic threshold, i.e., mild AUD. For both moderate and severe AUD, the optimal cutoff score was 18/16 in men/women. On AUDIT-C, the optimal cutoff score was suggested to be 6/4 in men/women, and was suggested as one common cutoff score for all levels of AUD. Discussion: In a treatment-seeking sample, gender-based cutoffs on AUDIT were not motivated, instead seven was suggested as the common diagnostic threshold for a mild AUD. On AUDIT-C, a general cutoff for all levels of AUDs was suggested, and gender specific cutoffs were justified to achieve optimal sensitivity to detect an AUD.