ABSTRACT
BACKGROUND: The expressed emotion (EE) construct has predicted clinical outcomes in schizophrenia and depression, but few studies have been conducted with bipolar patients. Moreover, there is a particular dearth of information regarding the prediction of depressive versus manic episodes in bipolar patients. Questions also remain about the utility of EE compared to other variables (perceived criticism, relationship negativity, and chronic strain in close relationships) that more directly evaluate interpersonal stress and about specific predictions of mania or depression. METHODS: Forty-seven outpatients with bipolar I disorder participated in a 1-year longitudinal study. A close collateral of the patient completed the Five Minute Speech Sample (FMSS) to assess EE, and participants completed perceived criticism and negativity ratings of collaterals. Clinical outcomes and chronic interpersonal stress were assessed by interview at 3-month intervals. RESULTS: High EE predicted depressive, but not manic recurrence. Other variables of close interpersonal relationships were not significant predictors of recurrence. LIMITATIONS: Participants nominated collaterals, and those who did not have such a confidant were excluded. CONCLUSIONS: The FMSS was sensitive to even mild negativity by the collateral that predicted later depressive episodes. This is the first study to demonstrate polarity-specific effects of EE on the prediction of recurrence in bipolar disorder.