Findling 2000.

Methods	Randomised double‐blind control trial of risperidone and placebo
Participants	Setting: outpatients, 1 centre, inner city academic outpatient medical centre Sample size: 20; 10 active treatment, 10 placebo Sex: 19 male, 1 female Age range: 5 to 15 years Mean age: 9.2 (2.9) years (range 6 to 14 years) IQ range: IQ more than 70 Diagnosis: DSM‐IV of conduct disorder Inclusion criteria: Clinical Global Impression ‒ Severity score moderate severity; Child Behaviour Checklist aggression subscale T‐score 2 SD or more above mean for age and gender‐matched peers Comorbidity: moderate to severe ADHD excluded Withdrawn/dropouts: 11 withdrew (4 active treatment, 7 placebo) Other Interventions: psychosocial interventions not mentioned
Interventions	Intended dose: for under 50 kg, 0.25 mg increasing to 1.5 mg; and for over 50 kg, 0.5 mg increasing to 3 mg Mean dose at endpoint: 0.028 (± 0.004) mg/kg per day (range 0.75 to 1.50 mg/day)
Outcomes	Primary outcomes: Rating of Aggression Against People and Property Scale Secondary outcomes: Conners' Parent Rating Scale ‒ conduct problem subscale, Child Behaviour Checklist, Clinical Global Impression ‐ Severity, Clinical Global Impression ‐ Improvement Follow‐up interval: 10 weeks
Notes	Imputation method for incomplete data: unclear from the published study Funding/support This work was supported in part by Janssen Research Foundation (The Stanley Foundation) (p 1). Medication and placebo were supplied by Janssen (p 3).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	A random number list. The list was kept in the Center for Drug Research and not was not accessible to either PI or other study raters.
Allocation concealment (selection bias)	Low risk	A random number list. The list was kept in the Center for Drug Research and not was not accessible to either PI or other study raters.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Risperidone and placebo matched in appearance. The blind was not broken during the course of the trial.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not stated.
Incomplete outcome data (attrition bias) All outcomes	High risk	Of the 10 youths assigned to risperidone, 6 completed the entire study (40% attrition): "3 youths who were assigned to receive risperidone were withdrawn by their guardian because of lack of effect, and 1 youth who received risperidone was withdrawn from the study during week 4 because of the development of a rash" (p 511). Only 3 youths who received placebo finished the trial (70% attrition): "4 patients assigned to placebo were withdrawn from the protocol by their guardians because of lack of benefit, 2 more were withdrawn from the study by the PI because of non‐compliance with study procedures, and 1 youth randomly assigned to placebo was lost to follow‐up" (p 511).
Selective reporting (reporting bias)	Unclear risk	Protocol unavailable.
Other bias	High risk	Small sample size and therefore limited power to detect differences.