(PS7-54) A Principal Component Analysis of Borderline Personality Disorder with a Nationally Representative Epidemiological Sample of Non-institutionalized, Civilian Adults in the US

Abstract Text: Borderline personality disorder (BPD) is characterized as a serious mental health condition composed of unstable affect, interpersonal relationships, and self-image and engagement in impulsive behavior (American Psychiatric Association, 2013; Grant et al., 2008). There have been some disagreements concerning the conceptualization of the disorder (Trull & Widiger, 2013), which complicates its assessment and treatment. Thus, we examined the factor structure of BPD with a large nationally representative epidemiological sample of adults.

Participants were 36309 non-institutionalized US civilians aged 18-90 years old who participated in the National Epidemiologic Survey on Alcohol and Related Conditions-III. Diagnostic criteria for BPD were measured using the Alcohol Use Disorder and Associated Disabilities Interview Schedule-5 (AUDADIS-5), which is a structured interview administered by trained interviewers. Participants were evaluated with the nine DSM-5 BPD criteria assessed with the 30 AUDADIS-5 items: unstable relationships (3 items), affective instability (3 items), frantic efforts to avoid abandonment (4 items), inappropriate anger (3 items), unstable sense of self (4 items), chronic emptiness (2 items), paranoid ideation/severe dissociative symptoms (3 items), self-harm/suicide behaviors (2 items), and impulsivity (6 items). We used a principal component analysis with a varimax rotation with Kaiser normalization to examine the factor structure of BPD.

Analyses accounted for 52.24% of the variance and identified six components. The first component (labeled “self”) represented the largest proportion of the variance (29.05%) and consisted of nine items related to identity, perception of reality, and feelings of emptiness; the second (labeled “affective instability and anger”; 6.28%) consisted of six items related to emotional regulation and anger; the third (labeled “impulsivity”; 5.01%) consisted of six items related to impulsivity and risky behaviors; the fourth (labeled “efforts to avoid abandonment”; 4.39%) consisted of four items related to relationships and abandonment fears; the fifth (labeled “unstable relationships”; 4.02%) consisted of three items related to relationship instability; and the sixth (labeled “suicidality”; 3.49%) consisted of two items related to self-harm and suicidal ideation.

The 30 BPD items from the AUDADIS-5 fell into six distinct symptom clusters: self, affective instability and anger, impulsivity, efforts to avoid abandonment, unstable relationships, and suicidality. Surprisingly, the component of the self (composed of items regarding identity, perception of reality, and feelings of emptiness) accounted for the largest proportion of the variance, which seemingly contradicts the biosocial model. The model conceptualizes BPD to be rooted in the interplay of biological vulnerabilities (the genetic predisposition for emotional vulnerability) and invalidating environments (which exacerbate emotion dysregulation and are linked to problem behaviors). Future examination is needed to confirm this factor structure with other populations of BPD (e.g., a clinical sample), as well as by using other assessments of BPD (e.g., the 18 items from the Wave 2 NESARC).