Borderline personality disorder (BPD) is a severe mental health condition typically identified in adolescence. We aim to investigate the factor structure, internal consistency, item analysis, convergent validity, and diagnostic accuracy of the Greek version of the Borderline Personality Features Scale for Children-11 (BPFSC-11). A sample of 112 early adolescents aged 11 to 14 years was recruited from an outpatient community child and adolescent mental health service. Confirmatory Factor Analysis was used to explore the fit of the unidimensional model to our sample. Reliability was assessed using the omega coefficient, and regression analysis was employed to evaluate convergent validity with the dysregulation profile of psychopathology, as measured by the Youth Self Report (YSR). Item analysis was assessed via Item Response Theory, while criterion validity was evaluated against clinical evaluation using the ROC curve. We found that BPFS-11 did not fit the data well, while a 9-item version provided a good fit (RMSEA = 0.058, CFI = 0.986, TLI = 0.982). Reliability was strong (ω = 0.81). Convergent validity was demonstrated, as adolescents with high scores on the dysregulation profile of the YSR had statistically significantly higher summed scores on the BPFS. The ROC curve analysis suggested high diagnostic accuracy (AUC=0.86). Sensitivity at the optimum cut-off point of 26 was found to be 0.88, while specificity was 0.75. Our study presents evidence for the reliability and validity of the BPFS and provides a cut-off point for its use in outpatient clinical settings to timely identify adolescents with high borderline personality traits.

KEYWORDS: Borderline personality, Community child and adolescent mental health service, Early adolescence, Psychometrics.

Iouliani Koullourou, Konstantinos Kotsis, Giovanni Abrahão Salum, Carla Sharp, Emmanouel Tsalamanios, Evangelia Karagiannopoulou, Thomas Hyphantis, Christos Mantas

Full article in pdf