Aman 2002.

Methods	Double‐blind, randomised controlled trial of risperidone solution and placebo
Participants	Setting: outpatients, multicentre Sample size: 118; 55 active treatment (reported for 52 participants); 63 placebo Sex: 97 males, 21 females Age range: 5 to 12 years Mean age: active treatment 8.7 (SD = 2.1) years; placebo 8.1 (SD = 2.3) years IQ range: 36 to 84 Inclusion criteria: ≥ 24 (70th percentile) on the Conduct Problem subscale of the Nisonger Child Behaviour Rating Form Diagnosis: DSM‐IV diagnosis of conduct disorder, oppositional defiant disorder, or disruptive behaviour disorder not otherwise specified Comorbidity: 70 ADHD (33 active treatment; 37 placebo) Withdrawn/dropouts: 31 withdrawn (12 active treatment; 19 placebo) Other interventions Use of consistent doses of psychostimulants was permitted if the dose had been stable for at least 30 days before the start of the study. Behavioural therapy was permitted if it was initiated at least 30 days before the start of the study. No changes to psychostimulant use or behavioural therapy were allowed during the trial.
Interventions	Intended dose: risperidone solution 0.01 mg/kg increasing to 0.02 mg/kg on day 3 Mean dose at endpoint: 1.16 mg/day (mean 0.037 mg/kg per day)
Outcomes	Primary outcomes: Conduct Problem subscale of the Nisonger Child Behaviour Rating Form Secondary outcomes: Aberrant Behaviour Checklist, Behaviour Problem Inventory, Clinical Global Impression ‒ Severity Rating Scale, Clinical Global Impression ‒ Change Scores, Visual Analogue Scale of the target symptoms Follow‐up interval: 6 weeks
Notes	Imputation method for incomplete data: last observation carried forward (LOCF) (p 1339) Funding/support "Supported by Janssen Research Foundation" (p 1344). "Study medication was provided by Janssen Research Foundation" (p 1338).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Randomly assigned" (p 1338) ‒ insufficient information.
Allocation concealment (selection bias)	Unclear risk	Not stated.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Appearance and taste of solutions were identical. All trial medications were labelled with the protocol number, medication number, lot number and strata. A tear‐off label was provided on each box of study medication which contained the medication code. The label was placed in the Case Report Form on the appropriate page. The code should only be broken in case of an emergency (Loy 2011b [pers comm]).
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not stated.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	12 participants (22%) in the risperidone and 19 (30%) in the placebo group withdrew. 4 in the risperidone and 15 in the placebo group due to insufficient response, 2 in risperidone because of adverse events, 3 in risperidone due to non‐compliance, 1 in risperidone and 3 in placebo lost to follow‐up, 1 in risperidone and 1 in placebo withdrew consent and 1 in risperidone lost medication. No efficacy data were recorded for 3 patients in the risperidone group and hence they were not included in any efficacy analyses (but the authors stated to have used an intention‐to‐treat analysis).
Selective reporting (reporting bias)	Unclear risk	Protocol unavailable.
Other bias	Unclear risk	1‐week placebo run‐in "to rule out placebo responders" (p 1338).