Background: People with schizophrenia comprise the majority of patients with severe mental illness recruited to recent mental health service studies of new teams (e.g. assertive outreach, crisis resolution). Reduction in hospitalisation has been the most consistent outcome measure in these studies, but results are inconsistent.
Aims: To understand inconsistency of results from studies using hospitalisation as an outcome measure.
Method: The advantages and disadvantages of hospitalisation are explored, including the ways in which it is recorded. Regional variation in outcomes and the impact of control services are reviewed.
Results: Hospitalisation has face validity as an outcome but translates poorly between differing healthcare contexts. These variations can be exploited positively to distinguish potentially effective ingredients in community care (outreach, combined health and social care, team structure) from redundant components.
Conclusions: Hospitalisation is a good proxy outcome measure in schizophrenia care in randomised controlled trials, but the dangers of extrapolating to new contexts require care.