The predictive capacity of AUDIT and AUDIT-C among adolescents in a one-year follow-up study.

AIM: The Alcohol Use Disorders Identification Test (AUDIT) has been validated for use with adolescents to screen their harmful alcohol consumption. How well AUDIT or its derivative consumption version AUDIT-C predicts the development of problematic alcohol use among adolescents remains unknown. The aim of our study was to examine the predictive capacity of AUDIT and AUDIT-C among adolescents in a one-year follow-up. METHODS: Finnish adolescents (N = 337) were examined at baseline with AUDIT and one year later with the Schedule for Affective Disorders and Schizophrenia for School-Age Children - Present and Lifetime version (K-SADS-PL) interview to assess alcohol problem use. Test characteristics and regression models were analyzed in predicting alcohol problem use. RESULTS: The sensitivity of AUDIT (cut-off ≥5) was 0.809 and specificity 0.621 in predicting alcohol problem use among adolescents one year later. The positive test posterior probability was 0.51. For those who screened negative at baseline, the positive test posterior probability was 0.13. With AUDIT-C (cut-off ≥3), the posterior probabilities were 0.47 and 0.12, respectively (sensitivity 0.855, specificity 0.529). The odds ratio was 6.95 for those screening positive with AUDIT and 6.59 with AUDIT-C at baseline to have alcohol problem use one year later. CONCLUSIONS: AUDIT has utility in screening youth at risk for developing alcohol problem use. It has significant predictive capacity in detecting risk especially among adolescents with depression.

AUDIT and AUDIT-C as screening instruments for alcohol problem use in adolescents.

BACKGROUND: The Alcohol Use Disorders Identification Test (AUDIT) is commonly used in adults to screen for harmful alcohol consumption but few studies exist on its use among adolescents. Our aim was to validate the AUDIT and its derivative consumption questionnaire (AUDIT-C) as screening instruments for the detection of problem use of alcohol in adolescents. METHODS: 621 adolescents (age-range, 12-19 years) were drawn from clinical and population samples who completed the AUDIT questionnaire. Psychiatric diagnoses were assessed using K-SADS-PL. A rating based on the K-SADS-PL was used to assess alcohol use habits, alcohol use disorders, screening and symptom criteria questions. Screening performance of the AUDIT and AUDIT-C sum scores and Receiver Operating Characteristic (ROC) curves were calculated. The diagnostic odds ratios (dOR) were calculated to express the overall discrimination between cut-offs. RESULTS: Comparisons of ROC between the AUDIT and AUDIT-C pairs indicated a slightly better test performance by AUDIT for the whole sample and in a proportion of the subsamples. Optimal cut-off value for the AUDIT was ≥5 (sensitivity 0.931, specificity 0.772, dOR 45.22; 95% CI: 24.72-83.57) for detecting alcohol problem use. The corresponding optimal cut-off value for the AUDIT-C was ≥3 in detecting alcohol problem use (sensitivity 0.952, specificity 0.663, dOR 39.31; 95% CI: 19.46-78.97). Agreement between the AUDIT and AUDIT-C using these cut-off scores was high at 91.9%. CONCLUSIONS: Our results for the cut-off scores for the early detection of alcohol problem use in adolescents are ≥5 for AUDIT, and ≥3 for AUDIT-C.