Currier 2004.

Methods	Allocation: randomised. Blindness: single‐blind (rater‐blinded). Duration: 24 hours.
Participants	Diagnosis: DSM‐IV diagnosis of paranoid schizophrenia (N = 54), schizoaffective disorder (N = 36), bipolar disorder (N = 13), psychotic disorder not otherwise specified (N = 31), other (N = 28). N = 162. Age: range 18‐65 years. Sex: 105 males, 57 females. History: acute agitation. Excluded: pregnancy, delirium, epilepsy, learning disability, intoxication, symptoms of withdrawal from alcohol or other psychoactive substances, medical illness, treatment with any anti‐psychotic or benzodiazepine within 6 hours of screening, history of neuroleptic malignant syndrome, known hypersensitivity to any of the trial medications, treatment with depot injection, use of disallowed medication. Setting: multi‐centre, USA (24 sites)
Interventions	1. Haloperidol: dose 5 mg/IM + lorazepam 2 mg/oral (mean number of IMs not reported). N = 79. 2. Risperidone: dose 2 mg/oral + lorazepam 2 mg/oral (mean number of doses not reported). N = 83.
Outcomes	Tranquillisation or asleep. Leaving the study early. Global state: CGI‐S, additional lorazepam. Agitation: PANSS‐EC. Aggression: OAS. Adverse effects: BAS, SAS, EPS, heart monitoring, sedation.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Described as randomised ‐ method of randomisation is not reported.
Allocation concealment (selection bias)	Low risk	"A telephone‐based central service was used to randomly assign eligible patients to receive a single dose of oral treatment or IM injection." (p.387)
Blinding of participants and personnel (performance bias) All outcomes	High risk	Single‐blind study.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	"independent raters blinded to the treatment arm conducted efficacy assessments". (p.387)
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	A flow diagram is provided with the reasons for leaving the study early reported. However, the number of participants completing different assessments at the same time point vary and reasons are not given.
Selective reporting (reporting bias)	High risk	The authors only report the side effects that occurred in ≥ 5% of people.
Other bias	High risk	Sponsored by drug company (Janssen Phamaceutica).