Diagnostic instability and depression.

Blind raters applied research criteria to the charts of patients whose discharge diagnosis changed through several hospital admissions from depression to schizophrenia and to the charts of matched control patients. Fewer diagnostically unstable patients than control patients satisfied research criteria for depression despite matching discharge diagnoses. The use of operational criteria, then, can be expected to reduce diagnostic instability in depression and thus improve prognostic accuracy. Some patients satisfied the research criteria for depression when they were initially admitted and for schizophrenia at their final admission. This suggests that diagnostic error may not account for all cases of major diagnostic instability.