OBJECTIVE: The purpose of this study was to perform a systematic review of the diagnostic accuracy of the Alcohol Use Disorders Identification Test (AUDIT) for detecting at-risk drinking. METHOD: The MEDLINE, PsycINFO, Science Citation Index Expanded, BIOSIS Previews, Cumulative Index to Nursing and Allied Health Literature (CINAHL), MEDION, and Cochrane Library databases were searched for relevant studies. The criteria for inclusion were a valid reference standard, AUDIT consisting of 10 items, avoiding bias that may result from how the reference standard was obtained, and when and how many participants were tested. Data were extracted independently by two reviewers. Data synthesis was performed by applying direct pooling of proportions and random effects model for likelihood ratios and diagnostic odds ratio. RESULTS: Twenty-three studies were included in the systematic review, 19 of which were included in the meta-analysis. With a cutoff of 8 points, sensitivity ranged from .31 to .89 and specificity ranged from .83 to .96 across the eight studies conducted in primary care. A single trial in general hospital inpatients found a sensitivity of .93 and a specificity of .94; another trial in emergency-department patients found a sensitivity of .72 and a specificity of .88. A study in university students found a sensitivity of .82 and a specificity of .78. Three studies in elderly patients found sensitivities between .55 and .83 at a pooled specificity of .96. There was large heterogeneity between study results, which could only partly be explained by setting diversity. The analysis of results with population-specific cutoff points led to similar findings. CONCLUSIONS: Findings on the diagnostic performance of the AUDIT proved to be largely heterogeneous. Its use should be restricted to primary care populations, inpatients, and elderly patients.
OBJECTIVE: The purpose of this study was to perform a systematic review of the diagnostic accuracy of the Alcohol Use Disorders Identification Test (AUDIT) for detecting at-risk drinking. METHOD: The MEDLINE, PsycINFO, Science Citation Index Expanded, BIOSIS Previews, Cumulative Index to Nursing and Allied Health Literature (CINAHL), MEDION, and Cochrane Library databases were searched for relevant studies. The criteria for inclusion were a valid reference standard, AUDIT consisting of 10 items, avoiding bias that may result from how the reference standard was obtained, and when and how many participants were tested. Data were extracted independently by two reviewers. Data synthesis was performed by applying direct pooling of proportions and random effects model for likelihood ratios and diagnostic odds ratio. RESULTS: Twenty-three studies were included in the systematic review, 19 of which were included in the meta-analysis. With a cutoff of 8 points, sensitivity ranged from .31 to .89 and specificity ranged from .83 to .96 across the eight studies conducted in primary care. A single trial in general hospital inpatients found a sensitivity of .93 and a specificity of .94; another trial in emergency-department patients found a sensitivity of .72 and a specificity of .88. A study in university students found a sensitivity of .82 and a specificity of .78. Three studies in elderly patients found sensitivities between .55 and .83 at a pooled specificity of .96. There was large heterogeneity between study results, which could only partly be explained by setting diversity. The analysis of results with population-specific cutoff points led to similar findings. CONCLUSIONS: Findings on the diagnostic performance of the AUDIT proved to be largely heterogeneous. Its use should be restricted to primary care populations, inpatients, and elderly patients.