To what degree does the Composite International Diagnostic Interview (CIDI) correctly identify DSM‐IV disorders? Testing validity issues in a clinical sample

Abstract

This paper compares diagnoses of the DSM‐IV algorithms of the computer‐assisted version of the Munich‐Composite International Diagnostic Interview (M‐CIDI) with clinical diagnoses made by treating physicians using the standard LEAD procedure. A random sample of 68 patients being treated in three psychiatric and one neurological ward of the Max Planck Institute of Psychiatry were first examined with the CIDI, using the M‐CIDI/DSM‐IV algorithms. Diagnostic findings were then compared with lifetime and cross‐sectional diagnoses assigned by the treating physician, who was blind to the CIDI findings, taking into account all available symptom and diagnostic information from current and previous charts as well as his own assessments. Clinicians were encouraged to use the DSM‐IV manual to assign multiple lifetime diagnoses and not to focus exclusively on primary diagnoses. To explore agreements and disagreements further, all discrepant cases were subsequently discussed with the treating physicians. There was generally good concordance between clinicians and interview DSM‐IV diagnoses, with the exception of psychotic disorders (kappa: 0.21), dysthymia (0.54) and somatoform disorders (0.50), with kappa values for the remaining categories ranging from 0.63 (any panic disorder) to 0.96 (any depressive episode). It is concluded that the M‐CIDI, in clinical cases, provides valid diagnoses for almost all non‐psychotic disorders and is sensitive for all disorders but might have reduced speci…