Comparing the personality disorder interview for DSM-IV (PDI-IV) and SCID-II borderline personality disorder scales: an item-response theory analysis

Abstract

One-hundred sixty-nine psychiatric outpatients and 171undergraduate students were assessed with the Personality Disorder Interview-IV (PDI-IV; Widiger, Mangine, Corbitt, Ellis, & Thomas, 1995) and the Structured Clinical Interview for DSM-IV Axis II disorders (SCID-II; First, Gibbon, Spitzer, Williams, & Benjamin, 1997) for borderline personality disorder (BPD). Eighty individuals met PDI-IV BPD criteria, whereas 34 met SCID-II BPD criteria. Dimensional ratings of both measures were highly intercorrelated (rs = .78, .75), and item-level interrater reliability fell in the good to excellent range. An item-response theory analysis was performed to investigate whether properties of the items from each interview could help understand these differences. The limited agreement seemed to be explained by differences in the response options across the two interviews. We found that suicidal behavior was among the most discriminating criteria on both instruments, whereas dissociation and difficulty controlling anger had the 2 lowest alpha parameter values. Finally, those meeting BPD criteria on both interviews had higher levels of anxiety, depression, and more impairments in object relations than those meeting criteria on just the PDI-IV. These findings suggest that the choice of measure has a notable effect on the obtained diagnostic prevalence and the level of BPD severity that is detected.