Relatives play a key role in supporting people with psychosis at all stages of recovery, but this can be associated with high levels of distress. Family interventions, with an international evidence base, improve outcomes for service users but little is known about their impact on relatives' outcomes. This review of published evaluations aimed to assess whether family interventions are effective in improving outcomes for relatives of people with psychosis, to identify the key components of effective intervention packages, and to identify methodological limitations to be addressed in future research. Fifty studies were identified which evaluated an intervention to support relatives against a control group, and in which outcomes for the relatives were reported. Thirty (60%) studies showed a statistically significant positive impact of the intervention on at least one relatives' outcome category. Eleven key intervention components were identified across all 50 studies, but there was no evidence that the presence or absence of any of these key components reliably distinguished effective from ineffective interventions. Methodological quality of studies was generally poor with only 11 studies rated as adequate using the Clinical Trial Assessment Measure (CTAM). Recommendations to improve future research include larger samples; better defined interventions and controls; true randomisation and blind assessors; clearly specified primary outcomes; pre-published analysis plans that account appropriately for missing data and clustering of data; a consensus on the most relevant outcomes to assess and valid and reliable measures to do so. Alternative research designs need to be considered to evaluate more recent approaches which focus on family support, personalised to meet individual need, and offered as an integral part of complex clinical services.

The toolkit is a feasible and potentially effective intervention to improve outcomes for relatives. A larger trial is needed to reliably assess the clinical and cost-effectiveness of REACT, and its impact on longer-term outcomes.

No abstract