Abikoff 2004.

Methods	Design: RCT
Participants	Country: USA and Canada Setting: outpatient clinics in two large medical centres in New York and Montreal Sample size calculation: not reported Sample size: 103 children Sex: 93% = boys, 7% = girls Age: range = 7‐9.9 years Ethnicity: white = 84%, African‐American = 13%, Hispanic = 2%, other = 1% Socioeconomic status: 84 children (81.2%) lived with both parents, 13 (12.6%) with one parent, and six (5.8%) with their mother and stepfather IQ: normal IQ (i.e. WISC‐R ≥ 85) ADHD diagnosis: subtypes not reported ADHD medication: all participants received psychostimulants Comorbidity: ODD = 53.4%, CD = 30%, anxiety disorder = 16.5% Medications for comorbid disorders: not reported Inclusion criteria: diagnosis of ADHD based on the DISC‐P2 conducted by a clinical psychologist. The diagnosis had to be confirmed by a child psychiatrist based on a comprehensive clinical interview with the child and parent and teacher reports. The children had to, on two different occasions, receive a mean teacher rating of at least 1.5 on the hyperactivity factor or the Hyperactivity Index of the Conners Teachers Rating Scale (Conners 1998). children had to be medication‐free for at last two weeks before evaluation. normal IQ (i.e. WISC‐R ≥ 85) living with at least one parent, with telephone access positive response to methylphenidate Exclusion criteria: children with diagnosable neurological disorders psychosis significant medical illness current physical or sexual abuse chronic tic disorder or Tourette's disorder DSM‐III‐R‐based developmental reading or arithmetic disorder, defined as a standard score in reading or mathematics on the Kaufmann Test of Educational Achievement of 85 or less children with a diagnosis of conduct disorder Baseline characteristics: no between‐group differences except on socioeconomic status, where there were differences between the group given methylphenidate alone and the group given methylphenidate + attention control treatment
Interventions	103 participants allocated to one of three groups Group one (n = 34): medical manual and methylphenidate (effort was made to give each child a maximal dose of methylphenidate). Five‐week open methylphenidate titration study before randomisation Group two (n = 34): methylphenidate + Multimodal Psychosocial Treatment (MPT) comprising parent training/family therapy, academic organisational skills training, individualised academic assistance, reading remediation (when necessary), social skills training, and individual psychotherapy. All treatment modules were fully manual‐based and the manual was developed before the start of the study. Each component was delivered once a week in the first year and once a month during the second year. Parents received parent management training and counselling. Daily report cards were completed by teachers and formed the basis for a home‐based reinforcement programme for targeted school behaviour and academic performance. Group three (n = 35): methylphenidate + attention control treatment (ACT) consisting of components parallel to those in MPT but excluding the therapeutic content. Delivered once a week in the first year and once monthly during the second year Attendance: 75% attendance required. 22 children (methylphenidate = 10, methylphenidate + MPT = 6, methylphenidate + ACT = 6)
Outcomes	Primary outcomes Social skills: SSRS, parent‐ and child‐rated; Social Interaction Observation Code General behaviour: CTRS, Conduct Problems Factor, teacher‐rated Secondary outcomes Core ADHD symptoms: CPRS, Hyperkinesis Index, parent‐rated; CTRS, Hyperactivity Factor, teacher‐rated; CTRS Hyperkinesis Index, observer‐rated Outcome assessment: end of treatment
Notes	Study ID: not reported Sponsorship source: not reported Year conducted: not stated Duration of the study: 2 years Comments: none Lead author's name: Howard Abikoff Institution: NYU Child Study Center, New York University School of Medicine Email: abikoh01@med.nyu.edu Address: NYU Child Study Center, New York University School of Medicine, New York
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Comment: we requested clarification from one of the study investigators. Howard Abikoff informed us in an email on 28 January 2011 (Abikoff 2011 [pers comm]) that they had used a block randomisation scheme with blocks of four children. The groups were balanced for age, sex, ODD, and ethnicity.
Allocation concealment (selection bias)	Low risk	Comment: we requested clarification from 1 of the study investigators. Howard Abikoff informed us in an email as above (Abikoff 2011 [pers comm]) that they had used sealed envelopes.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: no blinding
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: blinding on at least one of this review's primary outcomes but no blinding for the rest of the outcomes
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: 22 out of 103 children failed to complete the study.
Selective reporting (reporting bias)	Unclear risk	Comment: no prior statement of assessment tools. Design article published at the same time as study article
Vested interest bias	High risk	Comment: the study was based in two large medical centres and the centres have extensive previous experience with research focused on ADHD and behavioural treatment. Dr Klein is a member of a pharmaceutical board.
Other sources of bias?	Low risk	Comment: no other sources identified