Methods: Participants were 160 young women (ages 18–24) with clinically significant BPD symptoms (e.g., aggression and/or self-injurious behavior) recruited from a larger community-based longitudinal study (Pittsburgh Girls Study; PI: Loeber). Trauma and ED constructs were assessed using self-report measures at a baseline assessment (i.e., Childhood Trauma Questionnaire and Difficulties in Emotion Regulation Scale). BPD features were assessed dimensionally using the Personality Assessment Inventory-Borderline Features at a follow-up assessment six months after baseline.
Analyses: After examining initial descriptives, a series of linear models were employed to test the effects of ED and individual types of trauma (physical abuse, sexual abuse, emotional abuse, physical neglect, and emotional neglect) on BPD features, controlling for age, race/ethnicity and sexual orientation. In addition, a path analysis with three latent variables (trauma, ED and BPD symptomatology) was conducted using structural equation modeling in R, where trauma was measured by five individual trauma types as aforementioned, ED by five emotional constructs: non-acceptance, goals, impulse, awareness, strategies and clarity, and BPD by affective instability, identify problems, negative emotions and self-harm. Indirect effects were tested by a bootstrapping method, which does not make distributional assumptions.
Results: Controlling for multiple trauma types, sexual abuse (b=.33, t=2.12, p<.05) and ED (b=.32, t=9.06, p<.001) significantly prospectively predicted BPD symptomatology six months later. Path analysis revealed that three factors were well identified with good construct validity; further, there were significant direct effects of trauma on ED (b=.26, z =2.05, p <.05) and ED on BPD (b=.57, z =4.12, p <.001). After accounting for the indirect effect of trauma on BPD via ED (b=.15, z =2.02, p<.05), the direct effect of trauma on BPD remained significant (b=.33, z =3.08, p <.01).
Conclusion and Implications: Our study contributes to an improved understanding of associations between trauma, ED, and BPD features in a diverse group of young women. Trauma experience, especially sexual abuse, is critical in accounting for significantly more BPD features among young women, partially through the underlying emotion regulation difficulties. Implications: (1) Early screening of trauma symptoms should be integrated into BPD treatments, (2) emotional difficulties should be targeted when treating people with trauma experience, and (3) further research can utilize more repeated measures to assess trauma symptomology, as well as compare the differential effects between momentary emotional reactions and stable traits in exacerbating BPD symptoms after traumatic exposure.