Methods	Allocation: cluster randomised with 46 GP practices. Blinding: open study. Setting: community health centre, London, UK. Follow-up: over 12 months.
Participants	Diagnosis: psychosis (CAARMS*) N=113. Age: mean age 23 years. Sex: 81 M, 32 F. Excluded: participants with a history of contact with mental health services for psychosis for more than 6 months or antipsychotic treatment for more than a month. History: first-episode psychosis.
Interventions	1. GP education in early detection with access to LEO-CAT. N=23 practices, N=50 participants. 2. Usual care, with access to LEO-CAT. N=23 practices, N=63 participants
Outcomes	Hospitalisation. Referred to Mental Health Services by Accident and Emergency or Emergency Health services
Notes	*Comprehensive Assessment of At-Risk Mental States (CAARMS) (Yung 2005). LEO-CAT- Lambeth Early Onset Crisis Assessment Team. GP practices randomised into the intervention group received both the GP education training and direct access to the LEO CAT team for referrals.
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised, no further details.
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	High risk	Open study.
Incomplete outcome data addressed? All outcomes	High risk	Study attrition not reported.
Free of selective reporting?	Unclear risk	No details.
Free of other bias?	Unclear risk	No details.