Risk stratification for the early diagnosis of borderline personality disorder using psychiatric co‐morbidities

• Published in: Early intervention in psychiatry Vol. 12; no. 4; pp. 605 - 612
• Main Authors: Shen, Cheng‐Che, Hu, Li‐Yu, Tsai, Shih‐Jen, Yang, Albert C., Chen, Pan‐Ming, Hu, Ya‐Han
• Format: Journal Article
• Language: English
• Published: Melbourne Wiley Publishing Asia Pty Ltd 01.08.2018; Wiley Subscription Services, Inc
• Subjects: Adolescent; Adult; Anxiety; Attention deficit hyperactivity disorder; Bipolar disorder; Borderline personality disorder; Borderline Personality Disorder - diagnosis; Borderline Personality Disorder - epidemiology; Case-Control Studies; Comorbidity; data mining; decision tree; Diagnosis; Early Diagnosis; Eating disorders; Female; Humans; Insurance; Intellectual disabilities; Male; Medical diagnosis; Mental disorders; Mental Disorders - epidemiology; Middle Aged; Model accuracy; Morbidity; National Health Insurance Research Database; Patients; Personality; Personality disorders; Prevalence; Regression analysis; Retrospective Studies; Risk Factors; Risk management; Sleep; Sleep disorders; Taiwan - epidemiology; Young Adult; Taiwan; early diagnosis; National Health Insurance Research Database; decision tree; data mining; borderline personality disorder
• ISSN: 1751-7885; 1751-7893
• DOI: 10.1111/eip.12364

Aim To develop a risk stratification model for the early diagnosis of borderline personality disorder (BPD) using Taiwan National Health Insurance Research Database. Methods We conducted a retrospective case–control study of 6132 patients (292 BPD patients and 5840 control subjects) who were selected from the National Health Insurance Research Database. Psychiatric co‐morbidities including depressive disorder, bipolar disorder, anxiety disorder, substance‐use disorder, personality disorders other than BPD, sleep disorder, eating disorder, autistic spectrum disorder, mental retardation and attention‐deficit hyperactivity disorder, which were diagnosed within 3 years before enrolment, were collected. A logistic regression was used to calculate the odds ratio of psychiatric co‐morbidities between subjects with and without BPD. The classification and regression tree method was used to generate a risk stratification model. Results The odds ratios for depressive disorder, bipolar disorder, anxiety disorder, substance‐use disorder, personality disorders other than BPD, sleep disorder, eating disorder, mental retardation and attention‐deficit hyperactivity disorder were greater for BPD patients than for the control subjects. Furthermore, the risk of BPD can be reliably estimated using age and psychiatric co‐morbidities including bipolar disorder, substance‐use disorder and depressive disorder. Conclusions Most psychiatric disorders were more common in BPD patients than in the control subjects. Using psychiatric co‐morbidities, we identified four variables as significant risk predictors of BPD and permitted identification of subjects with low, intermediate or high risk for BPD. The accuracy of the risk stratification model is high and can be easily applied in clinical practice.