This study considers the question of whether relapse rates among schizophrenic patients can be reduced by means of relatives' groups. In a randomized, controlled intervention study, two therapeutic strategies (therapeutic relatives' groups, initiated relatives' self-help groups) were compared with each other and with a control group. Interventions were confined to the relatives, with the patients continuing their standard treatment. The study involved 151 relatives of 99 chronic DSMIII schizophrenics. Data were collected before and after a 1-year intervention phase and in a 2-year follow-up. No difference existed between the groups with respect to rehospitalization data. However, numerous differences recorded in the psychopathological findings and in living and working circumstances suggest that therapeutic work with relatives is of clinically significant benefit.