Comparison of Alcohol Screening Questionnaires for Women

Despite widespread agreement that primary care physicians should screen for alcohol problems, there is no consensus on the optimal screening test, especially for women. This study compared the performance of the 5-question TWEAK (Tolerance, Worried, Eye-opener, Amnesia, Cut down), the 10-question AUDIT (Alcohol Use Disorders Identification Test), and the 3 consumption questions from the AUDIT (AUDIT-C) to an interview reference standard in 393 female veteran outpatients.

Of the veterans, 23% met criteria for hazardous drinking (i.e., amounts that placed them at risk for consequences) and/or for alcohol abuse or dependence. Ten percent met criteria for active alcohol abuse or dependence alone. At cut-offs that retained acceptable specificity for hazardous drinking and/or active alcohol abuse or dependence,

• the TWEAK identified 44% of patients (specificity of 89%);
• the AUDIT identified 70% at a cut-off of 3, well below the usually recommended score of 8 or greater (specificity of 86%);
• the AUDIT-C detected 81% (specificity of 86%).

At the same cut-offs for each test, sensitivity was greater for detecting active alcohol abuse or dependence than for detecting the whole spectrum including hazardous drinking, abuse, or dependence.

Comments:

Of the 3 questionnaires tested, the AUDIT-C appears to be the best for detecting hazardous drinking and alcohol use disorders in women. Before its widespread use can be recommended, these findings need to be replicated in other patient populations (e.g., men, non-veterans). With replicated validation, the AUDIT-C will have a notable advantage (its brevity) over the other tests, and therefore may help solve the greatest deficiency in screening in current practice—the failure to use any validated questionnaire to screen at all.

Joseph Conigliaro, MD, MPH

Reference:

Bush KR, Kivlahan DR, Davis TM, et al. The TWEAK is weak for alcohol screening among female veterans affairs outpatients. Alcohol Clin Exp Res. 2003;27(12):1971–1978.
(view abstract)