RESUMO
OBJECTIVE: To identify features differentiating bipolar disorder (BP) from borderline personality disorder (BPD) and with each condition variably defined. METHOD: Participants were assigned a BP or BPD diagnosis on the basis of DSM criteria and, separately, by clinical judgment, and undertook a diagnostic interview and completed self-report measures. RESULTS: Predictors of BPD status varied according to diagnostic decisions, but with the most consistent items being childhood sexual abuse, childhood depersonalization, personality variables relating to relationship difficulties and sensitivity to criticism, and the absence of any BP family history. Across diagnostic groups, personality measure items alone predicted diagnostic allocation with an accuracy of 81-84%, the refined study variables other than hypo/manic features improved the classification rates to 88%, and when the presence or absence of hypo/manic features was added, classification rates increased to 92-95%. CONCLUSION: Study findings indicate that BPD can be differentiated from BP with a high degree of accuracy.