Objective: To examine the diagnostic performance of shorter versions of Alcohol Use Disorder Identification Test (AUDIT), including Alcohol Consumption (AUDIT-C), in identifying risky drinkers in primary care settings using conventional performance measures, supplemented by decision curve analysis and reclassification table.
Study design and setting: A cross-sectional study of adult males in general outpatient clinics in Hong Kong. The study included only patients who reported at least sometimes drinking alcoholic beverages. Timeline follow back alcohol consumption assessment method was used as the reference standard. A Chinese translated and validated 10-item AUDIT (Ch-AUDIT) was used as a screening tool of risky drinking.
Results: Of the participants, 21.7% were classified as risky drinkers. AUDIT-C has the best overall performance among the shorter versions of Ch-AUDIT. The AUC of AUDIT-C was comparable to Ch-AUDIT (0.898 vs 0.901, p-value = 0.959). Decision curve analysis revealed that when the threshold probability ranged from 15-30%, the AUDIT-C had a higher net-benefit than all other screens. AUDIT-C improved the reclassification of risky drinking when compared to Ch-AUDIT (net reclassification improvement = 0.167). The optimal cut-off of AUDIT-C was at ≥5.
Conclusion: Given the rising levels of alcohol consumption in the Chinese regions, this Chinese translated 3-item instrument provides convenient and time-efficient risky drinking screening and may become an increasingly useful tool.