Diagnostic interview for borderline patients. A replication study.

The borderline diagnosis is widely used despite a lack of systematic research on its reliability and validity. The recent development of a structured interview incorporating diagnostic criteria for borderline disorders in a replicable format represents a necessary, but not sufficient, methodological step in testing the validity of the borderline concept. To our knowledge, this is the first replication of Gunderson's Diagnostic Interview for Borderlines in a clinical setting and population quite different from the original. Clinically defined borderline patients were compared with control groups of schizophrenic and nondelusional unipolar depressed patients. Of 29 scored statements on the diagnostic interview, borderlines differed significantly from schizophrenics on 19, from depressives on 16, and from both on 19. Stepwise discriminant-function analyses of borderline vs each comparison group gave substantial support to the reliability of the interview and the diagnostic criteria.