Identifying depression among adolescents using three key questions: a validation study in primary care

Background: Depression in adolescents is a serious psychiatric illness. GPs play an important role in identifying adolescents with depression and those at risk of developing depression. Few validated tools are suitable for identifying adolescent depression in general practice.

Aim: To determine if three verbally asked key questions are valid for identifying depression in adolescents.

Design and setting: A cross-sectional, general practice multicentre, validation study was conducted in Oslo, Norway, and Aarhus, Denmark.

Method: A total of 294 adolescents answered three verbally asked key questions followed by a Composite International Diagnostic Interview (CIDI) for psychiatric diagnosis. Inclusion criteria were age (14-16 years) and fluency in the Norwegian or Danish language. The primary outcome was ROC curve statistics in terms of sensitivity and specificity, predictive values, and likelihood ratios of the three key questions. Secondary outcomes were Loevinger's H, Cronbach's α, and prevalence of depression.

Results: The three key questions met the criteria for construct and criterion validity for detecting depression among the adolescents. ROC curve statistics for the three key questions demonstrated an AUC of 0.79 for the answer 'yes' to either screening question and of 0.73 for the answer 'yes' to the help question. The positive predictive value was 31% and the negative predictive value was 97%.

Conclusion: The three key questions are useful for identifying depression in adolescents in primary health care.