Expressed emotion: current status

Almost three decades of controversy have surrounded the concept of'expressed emotion'. This fascinating saga has been carefully chronicled in the recent book by Leff and Vaughn, Expressed Emotion in Families (1985). The original work derived from the unexpected finding that long-stay patients survived longer in community settings and maintained better social functioning when they were discharged to lodgings, siblings, or distant relatives, than when they returned to parental or spousal households, or to professionally staffed hostels (Brown et al. 1958). The emotional response of key household members to the patient's behavioural disturbance appeared to be the best predictor of successful transfer to the community. Criticism of the crude methodology employed in this study and a second study (Brown et al. 1962) led to further replications (Brown et al. 1972; Vaughn & Leff, 1976 a) which employed standardized interview and rating techniques, the Camberwell Family Interview (Brown & Rutter, 1966; Vaughn & Leff, 19766). These studies both replicated the earlier findings that an index of the emotional response of household members to patients' disturbed behaviour during the three months prior to a hospital admission for an episode of florid schizophrenia was the best single predictor of the recurrence of florid episodes in the 9 months after discharge back to the same household. The predictive power of the ' expressed emotion' index remained intact when prognostic variables such as symptom patterns, work impairment, severity of behavioural disturbance, duration of illness, age and pre-morbid social adjustment were partialled out in a series of contingency analyses. Although past impairment of work function and disturbed behaviour were strongly associated with both relapse and high 'expressed emotion' (EE), these associations largely disappeared when examined in conjunction with the relatives' emotional responses. It was concluded, therefore, that the association between relapse and disturbed behaviour was mediated through the high EE response of household members to that behaviour and not through the behaviour itself. The care taken in these replication studies is exemplary. But before drawing definitive conclusions which link the emotional responses of household members to the course of schizophrenia a number of issues must be considered. These include: (i) What factors provoke high EE responses? (ii) What is the nature of the association between high EE responses and recurrence of florid episodes of schizophrenia? (iii) What factors moderate the negative predictive power of the EE index? (iv) Is EE a specific predictor of schizophrenic phenomenology?