Objective: To examine the diagnostic accuracy of the four-item self-report and parent-report versions of the CAGE-AID as a screen for substance-use disorders (SUD) among adolescents in mental health care. Method: Patients were 190 Dutch treatment-seeking adolescents (12-18 years old, mean age 15.5 years, 53% boys) and their parents who completed the CAGE-AID questionnaire during intake. DSM-IV clinical diagnoses were used as gold standard in receiver operating characteristic analyses to determine psychometric properties. Results: Psychometric properties for both the self-report and the parent-report versions showed the CAGE-AID's excellent diagnostic accuracy in predicting SUD. Conclusions: The CAGE-AID seems to be a brief and valid instrument for detecting SUD among adolescents in mental health care. However, denial may play a role in both the CAGE-AID and the clinical interviews. More research is needed to further establish the CAGE-AID's usefulness in other settings. The study's limitations have been noted.