Waxmonsky 2016.

Methods	Design: RCT, parallel group
Participants	Country: USA Setting: outpatient Sample size calculation: not reported Sample size: 68 children* Sex: 39 = boys, 17 = girls (group one: 20/31 (65%) = boys; group two: 19/25 (76%) = boys) Age: (group one: mean = 9.3 years (SD = 1.6), mean = 9.4 years (SD = 1.5)) Ethnicity: % defined as racial/ethnic minority: group one: = 12 (39%); group two = 9 (36%) Socioeconomic status: group one: mean = 42.3 (SD = 15.2) on Socioeconomic Index**, group two: mean = 42.03 (SD = 12.8) both on Nakao and Treas Socioeconomic Index (Nakao 1994) IQ: group one: mean = 100.6 (SD = 15.4); group two: mean = 100.7 (SD = 10.6) ADHD diagnosis: not reported ADHD medication: prior to therapy, phase psychostimulant doses were optimised for all participants. Thus, all participants received pharmacological treatment (group one: mean entry stimulant dose = 0.90 mg/kg/day (SD = 0.40); group two: mean entry stimulant dose = 0.90 mg/kg/day (SD = 0.43)) both in methylphenidate equivalents on a mg/kg/day. Comorbidity: (group one: CD = 4 (13%), ODD = 29 (94%), anxiety/subthreshold anxiety = 9 (29%); group two: CD = 1 (4%), ODD = 24 (96%), anxiety/subthreshold anxiety = 11 (44%)) Medications for comorbid disorders: balanced between groups Inclusion criteria: seven to 12 years of age combined subtype of ADHD and severe mood disorder ADHD evaluations based on the Disruptive Behavior Disorders Structured Parent Interview Exclusion criteria: IQ below 80 prominent traits of autism spectrum disorder use of any nonstimulant psychotropic bipolar I/II, or psychoses children with suicidal ideation
Interventions	68 participants allocated to one of two groups Group one (n = 35): medication + AIM (ADHD plus Impairments in Mood) consisting of 11 parallel‐group sessions for parents and children with parent and child group run in parallel. Each session lasted 105 minutes. Sesssions focused on: emotion recognition in self and others; connections between emotions and cognitions (e.g. problem‐solving when upset); application of coping tools and problem‐solving skills at school and home. A contingency management system was implemented with points that could be exchanged for gift cards. Group two (n = 33): medication + community psychosocial care. Encouraged to engaged with local psychosocial providers. Referrals but not treatment were provided from project staff. Use of community psychosocial care during the project period: 15 (60%) received other mental health services during the project period; of these two (8%) received school‐based counselling only, 13 (52%) received individual sessions (for mixture of behaviour problems, anger management, and social skills issues). Attendance: % of attendance was required. Completers were participants attending at least six of the 11 sessions (n = 29). Mean attendance = 9.7 out of 11 group sessions. All but two participants attended at least half of the group or make‐up sessions.
Outcomes	Primary outcomes Social skills: Social Skills Rating Scale (SSRS), parent‐rated General behaviour: SSRS: Problem Behaviour subscale, teacher‐rated; DBDRS: ODD subscale ‐ teacher version, parent‐rated Secondary outcomes Core ADHD symptoms: DBDRS: ADHD Symptoms subscale, teacher‐ and parent‐rated Outcome assessment: assessment at following points (in weeks after baseline): mid‐intervention (6 weeks), post‐intervention (11 weeks) and group one at follow‐up assessment (17 weeks)
Notes	Study ID:NCT00632619 Sponsorship source: National Institute of Mental Health (MH080791; Principal Investigator: Waxmonsky) Year conducted: 2016 Duration of the study: 11 sessions Comments: the study was approved by governing institutional review boards (IRBs) at both sites. Lead author's name: James G. Waxmonsky Institution: Pennsylvania State University, Department of Psychiatry Email: jwaxmonsky@hmc.psu.edu Address: Pennsylvania State University, Psychiatry, 500 University Dr. Dept of Psychiatry H073, Hershey Medical Center, Hershey, PA 17033
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Comment: randomised using a computer‐generated permutated blocking procedure. Randomisation occurred before the medication phase so parents would be aware of therapy status prior to making decisions about medication.
Allocation concealment (selection bias)	Low risk	Comment: assumed that the use of a computer blocking procedure would conceal the allocation
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: blinding of participants and personnel not possible
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: the PICO measures included were rated by parents or teachers. Clinician‐rated assessments were completed by staff masked to therapy status, though this was not relevant for these measures.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: 3 dropped out of the CC group due to assignment. All dropout was well described and except the dropout due to assignment, the reasons did not indicate a bias.
Selective reporting (reporting bias)	Low risk	Comment: all primary and secondary outcome measures stated in protocol were reported in paper. SSRS was also reported in the paper but this measure was not described in the study registration.
Vested interest bias	High risk	Comment: potential vested interest due to relationships with pharmaceutical companies
Other sources of bias?	High risk	Comment: potential bias in the data collected. Selection of informants prone to monetary incentive. The referral may have increased the service received in the CC group during the intervention period and thus may have provided a potential bias. At study registration, the authors also mentioned as a limitation that the study therapy group had more contact with study staff, which may have impacted results. Additionally, the first author seemed to be involved in the development of the treatment programme used in the study and thus there might be bias of interests.