Hyde 1987.

Methods	Allocation: randomised ‐ by tossing a coin, matched on their potential for rehabilitation, duration ill, diagnosis, psychological impairment, remedial problems. Blindness: open*. Duration: 2 years.
Participants	Diagnosis: schizophrenia (14 out of 22), remainder epilepsy, organic mental states, manic‐depressive illness (criteria not reported). N=22. Age: mean ˜ 42 years, range 16‐65 years. Sex: 12 M, 10 F. History: already in standard care hospital >6 months.
Interventions	1. Hostel ward: 24 hour care satellite ward distant from main hospital. N=11. 2. Standard care: District General Hospital psychiatric ward admission. N=11.
Outcomes	Unable to continue with placement. Unable to use: Global impression: PIRS (no SD). Mental state: Krawiecka, PSE (no numerical data). Service use: (no numeric data). Social functioning: Social Facilities Checklist (no numerical data), use of time (no SD). Behaviour: Remedial Problem Behaviour Scale (no data). Satisfaction: patient attitudes scales (no data for experimental group). Cognitive function: MMS (no data). Economic data: costs (no SD).
Notes	*ratings of Social Facilities Checklist ‐ blinded raters, all other outcomes were not blinded.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised, toss of a coin. Even numbers in each group.
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	High risk	Open study design.
Incomplete outcome data addressed? All outcomes	Unclear risk	Unclear.
Free of selective reporting?	High risk	Many outcomes were either not reported, or were rendered unusable due to missing standard deviations or N values.
Free of other bias?	Unclear risk	Unclear.

MMS ‐ Mini Mental State 
 PIRS ‐ Psychological Impairment Rating Scale 
 PSE ‐ Present State Examination