Borderline personality disorder: interrater reliability of the diagnostic interview for borderlines

Three interviewers, working in pairs and using the Diagnostic Interview for Borderlines (DIB), interviewed three samples of 10 adult psychiatric inpatients. The two members of each interviewer pair alternated in the role of interviewer. Each member of the interviewer pairs completed the interview schedule and scored the protocol according to the instructions provided by the authors of the DIB. Interrater reliability was determined for each interviewer pair using the Kappa statistic. Reliability values were at about the level reported recently by Spitzer, Forman, and Nee (1979) for DSM-III axis I diagnoses in general. Additional statistical analyses were performed to determine whether there were patient sex differences on DIB scores, whether the two male interviewers differed from the female interviewer in their ratings of male or female patients, and whether experience differences among the three interviewers were reflected in their rating of patients. Among the latter analyses, the only significant finding was that in the interviewer pair consisting of a very experienced male psychiatrist and a male first year psychology graduate student, the former rated male patients as less borderline than did the latter. No other difference was found among raters, or between sex of patient and interviewer sex or amount of clinical experience. It was concluded that the DIB is sufficiently reliable for use in clinical research.