Test-retest reliability of self-reported age at onset of selected psychiatric diagnoses in general health care

Objective: To investigate reliability of self-reported age at onset of frequent mental disorders and its association with patient and disorder characteristics. Method: A total of 1031 primary care patients with at least one lifetime psychiatric diagnosis were asked to report age at onset of their disorders at baseline and after 1 year. Intraclass correlation coefficients (ICC) for age at onset information were calculated for individual disorders. Results: ICC were high and lay between 0.6790 (generalized anxiety disorder) and 0.7977 (dysthymia). Factors associated with reliability are different for different disorders: gender for depressive episodes, gross national product per year per inhabitant for dysthymia, age for pain disorder, years of formal education for dysthymia, generalized anxiety disorder, and agoraphobia, and number of lifetime diagnoses for depression and agoraphobia. Conclusion: Self-reported age at onset is reliable. Further research on factors associated with reliability should focus on interview conditions and subject parameters during interview.