Brook 2000.

Methods	Allocation: randomised. Blindness: open. Duration: 7 days.
Participants	Diagnosis: diagnosis of schizophrenia, schizoaffective disorder, bipolar disorder, brief psychotic disorder, schizophreniform disorder, delusional disorder, or psychotic disorders not otherwise specified as defined by DSM‐III‐R. N = 132. Age: range 19‐66 years. Sex: 123 males, 9 females. History: recently hospitalised with acute psychosis. Excluded: substance‐induced psychosis (confirmed by urinalysis), psychosis related to organic origin, clinically relevant medical illness, abnormal ECG, imminent risk of suicide or homicide, history of substance abuse or dependence in the previous 2 months. Setting: Multi‐centre (7 countries).
Interventions	1. Haloperidol: flexible dose 2.5 mg/IM to 10 mg/IM (mean number of IMs not reported), maximum 4 doses in 24 hours for 3 days. Followed by haloperidol: flexible dose 10 mg/oral to 80 mg/oral during 24 hours for 4 days. N = 42. 2. Ziprasidone: dose 10 mg/IM, maximum 4 doses in 24 hours for 3 days (mean number of IMs not reported). Followed by ziprasidone: flexible dose 80 mg/oral to 200 mg/oral in 24 hours for 4 days. N = 90.
Outcomes	Mental state: BPRS. Global state: CGI‐S. Leaving the study early. Adverse events: COSTART, BAS, modified SAS, investigators assessment of severity, blood pressure, laboratory tests, concomitant medication, ECG, urinalysis, others in ≥ 10 of participants. Unable to use Specific behaviours: BPRS agitation items (not published). Global state: CGI‐I (SE/SD not reported). Sedation score: 5‐point categorical scale (not published, present mean/SD rather than binary data).
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Computerized randomization assigned".
Allocation concealment (selection bias)	Unclear risk	Not stated.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Open trial.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Open trial.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No evidence of incomplete outcome data. A table is provided which presents a summary of discontinuations.
Selective reporting (reporting bias)	High risk	The number of participants reported as discontinuing due to adverse effects is reported inconsistently (see Brook 1998c table 2 and Brook 2000 table 4). It is not clear whether all adverse events are reported. Brook 2000 table 4 only reports adverse events occurring in ≥10 of people, however there are more adverse events mentioned throughout the text.
Other bias	High risk	Sponsored by drug company.