Family interaction and the course of schizophrenic illness. Results of a multivariate prospective study.

The aim of the present study was to assess the prognostic relevance of relatives' interactive behaviour towards the patient, as covered by the Münster Family Interview (MFI), to the further course of the schizophrenic illness. The MFI is a family interview (of the whole family, including the patient) designed to record the emotional family atmosphere based on the concept of expressed emotion (EE). The ratings take place directly after the interview on five scales (criticism, hostility, overinvolvement, resignation and warmth), the resignation scale being added to the 'classic' EE scales. Ninety-nine families of outpatients diagnosed with schizophrenia according to the DSM-III were examined with the MFI during a home visit. The patients were seen 1 and 2 years after the first examination. The target criteria selected for the prognostic significance of the interaction measurements were: rehospitalisation within 2 years; extent of symptoms after 1 year, and psychosocial skills after 1 year. The significance of the interaction dimensions was verified in regression models. The control variable used in the regression models was the Strauss-Carpenter scale. Regression models were produced for the total group and for a subgroup of moderately ill patients. All target criteria yielded serviceable prediction models. The most important variable for prediction was the control variable, the Strauss-Carpenter scale. However the interaction variables made additional contributions to the prognosis, especially in the subgroup of moderately ill patients. The best MFI scale for all the outcome criteria was resignation; criticism predicted only the symptomatology, and emotional overinvolvement the level of social functioning after 1 year. In conclusion, practical work with families of schizophrenic patients should emphasise the protective function of relatives towards patients more strongly.