Screening for borderline personality disorder after self-harm episodes: validation of the self-rating KABOSS-BD, the borderline subscale of the Karolinska Affective and borderline symptoms scale

Background Borderline personality disorder (BPD) is overrepresented among individuals with deliberate self-harm, yet identification of patients requiring diagnostic evaluation remains challenging in emergency settings. Effective screening tools may facilitate earlier identification of symptoms, shorten diagnostic delay, and improve treatment planning. The present study sought to validate the borderline subscale of the Karolinska Affective and Borderline Symptoms Scale (KABOSS-BD) as a self-rating screening tool for BPD in patients presenting with self-harm, and to examine its correlation with clinical diagnosis. Methods Data were analysed from 671 patients presenting with self-harm (with or without suicidal intention) at emergency departments in three Swedish cities. We investigated the internal validity of KABOSS-BD and its correlation with clinically diagnosed BPD (ICD-10 Emotionally Unstable Personality Disorder) and determined optimal cut-off scores for women and men for screening purposes. Results The KABOSS-BD demonstrated good internal consistency (Cronbach’s α = 0.87) and acceptable diagnostic accuracy in identifying patients with BPD, with an Area Under the Curve (AUC) of 0.73. The optimal cut-off score was 32, yielding sensitivity of 68% and specificity of 68%. Women consistently scored higher than men on the KABOSS-BD, and gender-specific cut-offs improved diagnostic accuracy. The scale performed optimally for women with a cut-off at 33, while for men, the scale performed better at a lower threshold of 26. Conclusions The KABOSS-BD demonstrates good internal consistency and acceptable diagnostic accuracy for BPD screening in both women and men who present with suicide attempts and non-suicidal self-harm. The scale’s brevity and ease of administration make it suitable for emergency settings, potentially reducing diagnostic delay and improving treatment allocation for this high-risk population. Trial registration Not applicable.