Various drugs 1984b.

Study characteristics
Methods	Randomisation: random, no further details. Allocation: procedure not described. Blinding: double‐blind, drug appearance was made identical with respect to powder, colour, taste and volume by adding a gastric acid. Duration: one year. Design: parallel. Location: single‐centre. Setting: outpatient.
Participants	Diagnosis: schizophrenia (DSM‐III), in remission or residual state. N = 87. Gender: 53 men, 34 women. Age: mean 41 years. History: duration stable‐ n.i., but in remission, duration ill‐ mean 8.2 years, number of previous hospitalisations‐ mean 3.4, age at onset‐ mean 32.8 years, severity of illness‐ in remission or residual symptoms, baseline antipsychotic dose‐ n.i..
Interventions	1. Drug: haloperidol combined with biperidine and nitrazepam. Fixed dose: 1 mg, 3 mg or 6 mg/day.* N = 37. 2. Drug: propericiazine combined with biperidine and nitrazepam. Fixed dose: 10, 30 mg/day or 60 mg/day.* N = 37. 3. Placebo combined with biperidine and nitrazepam. Duration of taper: 0 days. N = 13. Rescue medication: not indicated, probably no additional antipsychotic medication allowed.
Outcomes	Examined Relapse: clinical judgement. Leaving the study early. Global state ‐ number of participants in sustained remission (study defined). Unable to use/Not included Prolactin levels (no predefined outcome of interest).
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)	Unclear risk	Procedure not described.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Double‐blind, drug appearance was made identical with respect to powder, colour, taste and volume by adding a gastric acid.
Blinding (performance bias and detection bias) Subjective outcomes	Unclear risk	Double‐blind, drug appearance was made identical with respect to powder, colour, taste and volume by adding a gastric acid.
Blinding (performance bias and detection bias) Objective outcomes	Low risk	Double‐blind, drug appearance was made identical with respect to powder, colour, taste and volume by adding a gastric acid.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	While in the placebo group and in the haloperidol group the rates of participants leaving early due to other reasons were low, 9 out of 12 participants in the propericiazine group discontinued due to overdose. It is questionable whether relapse rates could be accurately measured, because most participants did not reach the endpoint.
Selective reporting (reporting bias)	Low risk	No evidence for selective reporting.
Other bias	Low risk	No evidence for other bias.