Diagnosis of personality disorders (PDs) has been improved with the development of semi-structured interviews with a long-term perspective (Weertman et al., 2003 and Zimmerman, 1994). However, instability of symptoms in Borderline PD (BPD) requires an instrument to assess BPD in a short-term perspective. An instrument that can be used to reliably and validly assess the recent severity of BPD-manifestations is necessary for clinical evaluations and treatment-outcome research (Zanarini et al., in press). One instrument capable of assessing recent BPD-manifestations in a detailed way and yielding dimensional scores for each of the nine DSM-IV criteria, a total score, and cutoff points, is the Borderline Personality Disorder Severity Index (BPDSI) (Arntz et al., 2003). A recent review (Zanarini et al., in press) of instruments assessing BPD-severity discussed two interviews, the BPDSI, and the ZAN-BPD (Zanarini et al., 2003), and two self-report inventories, the BEST (Pfohl et al., 2009), and the BSL (Bohus et al., 2007). The BPDSI is the only instrument that assesses each DSM-IV BPD-criterion with multiple items and therefore offers reliable severity scores per criterion based on multiple indicators of each criterion. It therefore meets the need for an instrument assessing “detailed changes in specific domains of Borderline pathology” (Zanarini et al., in press).
Based on feedback by interviewers and on psychometric considerations, a new version was developed, the BPDSI-IV. The BPDSI-IV has now been used in at least six trials in two countries (Bellino et al., 2005, Bellino et al., 2006, Giesen-Bloo et al., 2006, Rinne et al., 2002, Rinne et al., 2003 and Verheul et al., 2003), was translated into seven languages, and is used in at least three running trials in Norway, the Netherlands, and the USA. The published trials show high sensitivity for detecting change. The aim of the present study is to document the instrument’s psychometric properties.
Many combinations of the DSM-IV criteria can lead to a BPD-diagnosis. With a categorical view, this leads to an overwhelming number of BPD-variants. With dimensional views on PDs the picture becomes simpler. Although recent taxometric studies support dimensional models of BPD (Arntz et al., 2009), the issue remains whether BPD should be viewed as one-dimensional or multidimensional (Adams et al., 2001, Arntz, 1999, Clarkin et al., 1993, Fossati et al., 1999, Hurt et al., 1990, Livesley and Schröder, 1991, Morey, 1991, Sanislow et al., 2002 and Zanarini et al., 1989). We detected 7 multidimensional models in the literature that we tested with our BPDSI-IV data (DSM-IV criteria-numbers listed after each factor): Zanarini et al. (1989) found four scales of borderline pathology: affect (6,7,8), cognition (9), impulsivity (4,5) and interpersonal relationships (1,2). Hurt et al. (1990) identified three criteria-sets: identity (1,3,7), affect (2,6,8) and impulsivity (4,5). Morey (1991) found four dimensions: affective instability (6,8), identity problems (3), negative relationships (1,2) and self-damaging behaviour (5). Livesley and Schröder (1991) found three factors, instability/disorganisation (1,2,3,6,7,9), interpersonal exploitation (no DSM-IV criteria applicable), and self-damaging behaviour (4,5,8). With the eight DSM-III-R criteria, Clarkin et al. (1993) found three factors: uncertainty about self & interpersonal difficulties (1,2,3,7), affect & affect regulation (5,6,8), and impulsivity (4). Discussion emerged on the value of anger/hostility (8 separately from affect and affect regulation) as a fourth independent factor (Clarkin et al., 1993). Sanislow et al. (2002) tested the DSM-IV BPD-criteria as a unitary construct but also as a three-factor model comprising disturbed relatedness (2,3,7,9), behavioural dysregulation (4,5) and affective dysregulation (1,6,8).
To summarize, the present study aimed to assess the psychometric properties of the BPDSI-IV. Interrater reliability, internal consistency, validity, and underlying factor structure were assessed. For the last, we compared a one-factor model (assuming one underlying dimension), a nine-factor model (nine DSM-IV criteria representing different factors), and seven multiple factor models derived from the literature, with confirmatory factor analysis. As to validity, we expected higher scores in the BPD-group than in Axis-I, -II and non-patient controls on all BPDSI-items, subscales, and total score, as well as significant correlations with other BPD-markers, but not with healthy, psychotic, or neurotic markers (Arntz et al., 2003).
