DSM-III melancholia: do the criteria accurately and reliably distinguish endogenous pattern depression?

The Structured Clinical Interview for DSM-III (SCID), Newcastle Endogenous/Reactive Index, Feinberg-Carroll Discriminant Index, and Hamilton Depression Scale were used to assess 70 depressed patients in order to determine similarities and differences in symptom structure and severity in those patients with and without endogenous/melancholic depression. All patients with melancholia according to DSM-III had definite endogenous major depression by the Research Diagnostic Criteria (RDC), but only 20 out of 35 patients with RDC definite endogenous depression were DSM-III melancholic. There was a greater difference in symptom pattern between those patients with definite endogenous depression and those with probable or non-endogenous depression than there was between the melancholic and non-melancholic definite endogenous depressives. A prerequisite for the valid delineation of a nosological category is the establishment of good reliability for diagnostic criteria. Using SCID ratings of audiotaped interviews of 9 patients (5 with major depression), the 8 raters in this study achieved a kappa coefficient of 0.79, suggesting that the use of a structured interview can improve the reliability of DSM-III diagnoses. Interrater reliabilities for most of the individual DSM-III major depressive episode and melancholia items were reasonable, but some were low. The low reliabilities could be improved by redefinition of the items to reduce ambiguity and by development of a SCID glossary.