Morton 1968.

Methods	Randomisation: random, no further details. Allocation: the hospital pharmacist was responsible for supplying placebo and active drugs to the ward, no one concerned with the care of patients knew which patients were started on placebo. Blinding: double, identical tablets, but in most cases nurses made correct forecasts on who was on drug and who was on placebo. Blinding was broken when a participant relapsed. Duration: 6 months. Design: parallel. Location: single‐centre. Setting: inpatients.
Participants	Diagnosis: chronic schizophrenia (clinical diagnosis by two psychiatrists). N = 20. Gender: 40 M. Age: 25‐55 years. History: duration stable ‐ maintenance doses of tranquillisers had been administered for at least 18 months, in six participants who had to change treatment no change in symptoms was noted during 6 weeks, duration ill ‐ n.i., number of previous hospitalisations ‐ n.i., but duration of current hospitalisation > 2 years, age at onset ‐ n.i., severity of illness ‐ n.i., baseline antipsychotic dose ‐ all but six participants were on chlorpromazine or trifluoperazine, dose n.i..
Interventions	1. Trifluoperazine: fixed/flexible dose n.i., allowed dose range n.i., mean dose n.i.. N = 14. 2. Chlorpromazine: fixed/flexible dose n.i., allowed dose range n.i., mean dose n.i.. N = 6. Rescue medication: n.i..
Outcomes	Response to treatment: general clinical impression of the raters. Relapse: worsening of global state. Unable to use: Behaviour: Wing Scale (incomplete data, analysed by single items).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Random, no further details.
Allocation concealment (selection bias)	Low risk	The hospital pharmacist was responsible for supplying placebo and active drugs to the ward, no one concerned with the care of patients knew which patients were started on placebo.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Double, identical tablets, but in most cases nurses made correct forecasts on who was on drug and who was on placebo.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Double, identical tablets, but in most cases nurses made correct forecasts on who was on drug and who was on placebo.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	It is unclear whether there were drop‐outs.
Selective reporting (reporting bias)	Low risk	No evidence for selective reporting.
Other bias	Unclear risk	Blinding was broken when a participant relapsed, but this is not a problem for our primary outcome response to treatment.