To examine the relapse rate achievable in a real-life early psychosis treatment service.
Method: A 2-year longitudinal cohort study of consecutive admissions to an early psychosis programme which served the entire population in a catchment area. The primary outcome measure was relapse.
Results: One hundred and forty-eight consecutive consenting admissions were recruited, 124 (83.8%) of whom were followed for 1 year and 116 (78.4%) for 2 years. Relapse was assessed by clinicians using structured criteria. The 2-year relapse rate among subjects with complete data collection was 34/95 (35.7%, 95% CI 26.2-46.3). A Kaplan-Meier life table censoring subjects lost to follow-up yielded a comparable estimate of the proportion not relapsing: 68% (95% CI 58-76%).
Conclusion: These estimates compare favourably with a published range of 2-year relapse from 55% to 70% in the older literature and are comparable with the results in recent clinical trials.