[Standardized diagnosing: computer-assisted (CIDI) diagnoses compared to clinically-judged diagnoses in a...

Within the past two decades, there has been an increasing trend towards the use of empirically-based, standardized instruments to diagnose mental disorders. The purpose of this study was to investigate the congruence between clinically-derived and standardized computerized diagnoses using the Composite International Diagnostic Interview (CIDI) in a psychosomatic setting.

N = 230 inpatients treated at the Department of Psychosomatic Medicine, Charité Berlin, were diagnosed for mental disorders by experienced clinicians and by the CIDI according to the diagnostic criteria of ICD-10 and DSM-IV.

Congruence between computerized and clinically-derived diagnoses for all mental disorders was fair to poor (kappa = 0.0 - 0.33). The diagnostic congruence of somatoform (F45), anxiety (F40 - 41), depressive (F32 - 34), eating (F50) and dissociative disorders (F44) and disorders due to psychoactive substance use (F10 - 19) was fairly low (kappa = 0.19 - 0.33). Diagnostic congruence of diagnoses of reactions to severe stress (F43) and obsessive compulsive disorders (F42) was particularly poor (kappa = 0.0 - 0.01).

Overall diagnostic congruence between computerized and clinical diagnostics was unsatisfactory. Thus, the validity of computer-assisted standardized interviews - like the CIDI - is questionable in a psychosomatic setting.