Hagerman 1988.

Methods	Double‐blind randomised crossover trial Study length: 3 weeks
Participants	15 participants 2 female, 13 male Mean (s.d.) age: 7.9 (2.4) years Age range: 3.8 ‐ 11.8 years Mean (s.d.) IQ: 58 (12.8) IQ range: 29 ‐ 77 (3 have IQ >70)
Interventions	1. Dextroamfetamine as Dexedrine Spansules 0.2mg/kg every morning + placebo capsule at midday (to give same dosing as methylphenidate) 2. Methylphenidate 0.3mg/kg twice daily 3. Placebo (lactose in capsule) twice daily Each intervention lasted one week
Outcomes	Primary outcome not stated. Outcome measures used: Conners Abbreviated Parent‐Teacher Questionnaire ADDH: Comprehensive Teacher Rating Scale (ACTeRS) Behavioural observation Large scale integrated sensor actometer Delay task Vigilance task
Notes	Sequence generation unclear Participants, personnel and outcome assessors were blinded Outcome data were complete (i.e. no withdrawals/drop‐outs) Possible selective outcome reporting: subscale scores presented where main scale not significant; blood pressure readings not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Not described. Investigators were contacted, but had no further information.
Allocation concealment?	Unclear risk	Not described. Investigators were contacted, but had no further information.
Blinding? All outcomes	Low risk	Medication and placebo tablets were described as identical.
Incomplete outcome data addressed? All outcomes	Low risk	It is unclear whether any participants dropped out (although this is unlikely given the brevity of the study).
Free of selective reporting?	High risk	Trial protocol unavailable. Where an outcome measure showed a non‐significant difference, results for subscales were presented. Blood pressure and pulse readings were not presented despite being described in the Methods section of the paper.
Free of other bias?	Unclear risk	Study was supported by the Children's Hospital of Kempe Research Center, Colorado (USA). No competing interests declared.
Washout	Unclear risk	There was no washout period between treatment and placebo.
Validity/reliability of outcome measures	Unclear risk	The primary outcome measure was the ACTeRS scale. This has been demonstrated to be reliable in children with ID, but has poor correlation with behavioural observation in this population. Actometer recordings may be sensitive to medication response but may correlate poorly with report measures. Use continuous performance tasks may not be sensitive to medication effects.