Waxmonsky 2010.

Methods	Design: RCT
Participants	Country: USA Setting: outpatient Sample size calculation: not reported Sample size: 56 children Age: range = six to 12 years (group one: mean = 8.3 years (SD = 1.6), group two: mean = 8.9 years (SD = 1.5)) Sex: 45 (80%) = boys, 11 (20%) = girls (group one: 24 (82.8%) = boys, 5 (17.2%) = girls; group two: 21 (77.8%) = boys, 6 (22.2%) = girls) Ethnicity: white = 80.4%, African‐American = 10.7%, mixed = 8.9% Socioeconomic status: Nakao and Treas Socioeconomic Index (Nakao 1994): group one: mean = 61 (SD = 17), group two: mean = 53 (SD = 13) IQ: WISC‐III IQ score > 75(group one: mean = 101 (SD = 16), group two: mean = 97 (SD = 13)) ADHD diagnosis:DSM‐IV‐TR; combined type = 48 (85.7%), inattentive type = seven (12.5%), hyperactive/impulsive type = one (1.8%) (group one: combined type = 24 (82.8%), inattentive type = four (13.8%), hyperactive/impulsive type = one (3.4%), group two: combined type = 24 (88.9%), inattentive type = three (11.1%), hyperactive/impulsive type = zero (0%)) ADHD medication: stimulant naive at baseline (group one = 13 (44.8%), group two = eight (29.6%)), all patients received psychostimulant medications as part of the study Comorbidity: CD = 22 (39.3%), ODD = 24 (42.9%), no comorbidity = 10 (17.9%) (group one: CD = 11 (37.9%), ODD = 15 (51.7%), no comorbidity = three (10.3%), group two: CD = 11 (40.7%), ODD = nine (33.3%), no comorbidity = seven (25.9%)) Medications for comorbid disorders: not reported Inclusion criteria: ADHD based on DSM‐IV Exclusion criteria: current or past history of seizures (not including benign febrile seizures) other physical conditions that precluded administration of atomoxetine (for example, marked cardiac conduction delay) documented failed study of atomoxetine, defined as three weeks or more on treatment with at least 0.8 mg/kg/d, or a documented inability to tolerate this dose serious forms of psychopathology other than ADHD such as autism, bipolar disorder, schizophrenia, or any other psychopathology requiring urgent treatment with psychotropic medication any history of major depression requiring treatment, or any past history of self‐harm or serious suicidal ideation an IQ of less than 75 (based on WISC‐III) no evidence of ADHD‐related impairment at school Baseline characteristics: no significant between group differences in mean doses of atomoxetine
Interventions	56 participants allocated to one of two groups Group one (n = 29): medication (as described in group two) + behaviour therapy in eight‐week intervention with three components: Parent group: received two‐hour sessions once a week for eight weeks, following the manual of Community Oriented Parent Education Program (COPE). COPE uses the principles of social learning theory to help parents develop skills to target their children's behaviour and lack of impulse control. Group leaders were advanced graduate students or doctoral level clinicians. Child group: participated in a social skills training (SST) programme consisting of a two‐hour session once a week for eight weeks. Group leaders were graduate students in clinical psychology. It was unclear if the child group intervention also was based on a manual. Teacher group: completed a daily report card which was sent to the parent daily. Parents were taught to monitor the report cards and provide appropriate consequences for their child based on the report cards (i.e. reward for positive performance and loss of privileges for negative performance). Group two (n = 27): medication; all medication was dosed openly using atomoxetine. A weight‐based protocol similar to previous studies using atomoxetine was used. Attendance: 62% of parents attended eight sessions, 62% attended six or more sessions. The children's attendance in the SST group was not reported; seven children (12.5%) discontinued the study (five from group one (medication + behaviour therapy) and two from group two (medication alone)).
Outcomes	Primary outcomes Social skills: SSRS, parent‐ and teacher‐rated General behaviour: CGI, observer‐rated Secondary outcomes Core ADHD symptoms: DBDRS, parent‐ and teacher‐rated Academic performance: Academic Performance Rating Scale, teacher‐rated Satisfaction with treatment: treatment satisfaction, parent‐rated Outcome assessment: post‐treatment
Notes	Study ID:NCT00918567 Sponsorship source: investigator‐initiated trial funded entirely by Eli Lilly and Company Year conducted: conducted between 2008 and 2013 Duration of the study: 11 weeks Comments: none Author name: James G Waxmonsky, MD Institution: Center for Children and Families Email: jgw@buffalo.edu Address: 106 Diefendorf Hall, 3435 Main St., Bldg 20, Buffalo, NY 14214
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Comment: clarification requested from one of the study investigators and Dan Waschbusch informed us in an email on 22 June 2011 that they had used a computer‐generated randomisation process (Waschbusch 2011 [pers comm]).
Allocation concealment (selection bias)	Low risk	Comment: clarification requested from one of the study investigators and Dan Waschbusch informed us in an email on 22 June 2011 that the clinicians did not know the treatment assignment before it was assigned (Waschbusch 2011 [pers comm]).
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: no blinding
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: We requested clarification from one of the study investigators and Dan Wascbusch informed us in an email that participants were dropped if there was not sufficient information. Scores in indexes were computed if at least 50% of the items in the index were answered; if not, they were counted as missing. Dan Wascbusch also informed us that they had essentially complete data at pre‐treatment and nearly complete data at post‐treatment. They had a lower response rate for teachers. They included whatever they had in the analyses and dropped participants when there was insufficient information, repeating this for each analysis (Waschbusch 2011 [pers comm]
Selective reporting (reporting bias)	High risk	Comment: protocol published in Clinicaltrials.gov after study had been conducted. Publication was not consistent with the report in Clinicaltrials.gov
Vested interest bias	High risk	Comment: funding from, and collaboration with, Eli‐Lilly
Other sources of bias?	Unclear risk	Comment: co‐medication not specified