Storebø 2012.

Methods	Design: RCT, parallel group
Participants	Country: Denmark Setting: clinical Sample size calculation: a sample size of 26 children in each group was needed based on a sample size calculation of 80% power in detecting a clinical relevant change of four points on the primary outcome measure of hyperactivity and impulsivity. Sample size: 56 children* Sex: 39 (71%) = boys, 16 (29%) = girls (group one: 19 (67.8%) = boys, nine (32.2%) = girls; group two: 20 (74%) = boys, seven (26%) = girls) Age: (group one: mean = 10.6 years (SD = 1.29); group two: mean = 10.2 years (SD = 1.34)) Ethnicity: Danish 100% Socioeconomic status: not reported IQ: both verbal and nonverbal IQ > 80 (group one: WISC verbal mean = 93.9 (SD = 15.7), group two: WISC non‐verbal mean = 94.8 (SD = 19.0). ADHD diagnosis: group one: inattentive = 10 (35.7%), hyperactive‐impulsive = 0 (0%), combined = 16 (57.2%), not otherwise specified = 2 (7.1%); group two: inattentive = 6 (22.2%), hyperactive‐impulsive = 2 (7.4%), combined = 16 (59.2%), not otherwise specified = 3 (11.1%) ADHD medication: after assessment and confirmation, the family was offered medical treatment for the child following a medication protocol. The children had never previously received medical treatment for ADHD. Comorbidity: (group one: ODD = four (33.3%), anxiety disorder = four (33.3%), depressive disorder = one (8.3%), tics and OCD = zero (0%), enuresis = two (20%), stuttering = one (5%); group two: ODD = four (40%), anxiety disorder = two (20%), depressive disorder = one (10%), tics and OCD = 1 (10%), enuresis = two (20%), stuttering = zero (0%)) Medication of comorbid disorders: not reported Inclusion criteria: ADHD diagnosis according to the Diagnostic and Statistical Manual of Mental Disorders (DSM‐IV, 1994) eight to 12 years of age at the time of the start of assessment parents willing to take part in study and give consent for medical treatment of child and for child's participation in study Exclusion criteria: schizophrenia or all the autism diagnoses according to DSM‐IV violent and criminal children both verbal and nonverbal (IQ) below 80 previously medicated for ADHD resistance against participating *baseline characteristics were reported for 55 children as data from the child that withdrew from group two was not allowed to be analysed. Baseline characteristics: no significant difference in baseline demographics between the 2 groups
Interventions	56 participants were allocated to one of two groups Group one (28 participants): SOSTRA which consisted of social skills training plus parental training combined with standard treatment (medication). The children were offered 90‐minute, weekly social skills training sessions for a total of eight weeks, during which, the parents attended parental training. Social skills training aimed to improve and maintain the individual’s social skills. The children were taught how to adjust their verbal and nonverbal behaviour in their social interactions. The training also included efforts to change the child’s cognitive assessment of the ‘social world’ and generally focused on teaching the children to ‘read’ the subtle cues in social interaction such as learning to wait for their turn. The standard treatment offered encompassed the normal practice regarding ADHD patients at the Child Psychiatric Clinic in Holbaek: after assessment and confirmation of the ADHD diagnosis, the family was offered medical treatment for the child following a medication protocol; the children had never previously received medical treatment for ADHD. Group two (28 participants): normal practice, as reported above for the intervention group Attendance: one participant in each group did not receive the allocated intervention and one participant in group two was lost to follow‐up.
Outcomes	Primary outcomes Social skills: CBRS, social problems subscale, teacher‐rated Emotional competencies: CBRS, emotional subscale, teacher‐rated General behaviour: CBRS, aggressiveness subscale, teacher‐rated Secondary outcomes Core ADHD symptoms: Conners 3, hyperactivity‐impulsivity subscale, teacher‐rated Performance and grades in school: CBRS, academic score subscale, teacher‐rated Outcome assessment: post‐intervention and follow‐up at three and six months after end of intervention
Notes	Study ID:NCT00937469 Sponsorship source: the SOSTRA study was financially supported by Region’s Zealand University Hospital (RESUS), Region Zealand Research Foundation, and Psychiatric Research Unit, Region Zealand. Funding was also received from the Fru C. Hermansens Foundation, Slagtermester Max Wørzner and Inger Wøzners Foundation, and TrygFonden. Year conducted: 2012 Duration of the study: 8 weeks Comments: the study obtained approval from the Regional Ethics Committee of Zealand (SJ‐85), was registered at the Danish Data Protection Agency DO50892, and registered at clinical trials.gov/NCT00937469. Lead author's name: Ole Jakob Storebø Institution: Child Psychiatric Clinic, Child and Adolescent Psychiatric Department, Region Zealand, Denmark Email: ojst@regionsjaelland.dk Address: Birkevænget 3, 4300 Holbæk Denmark
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Comment: adequate method was used. Randomisation was conducted using computer‐generated, permuted randomisation sequences in blocks of four with an allocation ratio of 1:1 stratified for sex and comorbidity.
Allocation concealment (selection bias)	Low risk	Comment: adequate method was used. All data that could be used to identify the allocation before data entry was hidden and block size was unknown to the investigators.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: it was not possible to ‘blind’ participants, parents, treating physicians, or personnel in the Child Psychiatric Clinic in Holbaek.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Comment: the involved parties were instructed not to inform the teachers, who rated the primary and secondary outcome measures, of the intervention allocated. The outcome assessors were thus kept unaware of group allocation.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: one child from each group dropped out after the randomisation. Outcome assessment was still obtained from the child allocated to group two. Another child from group two was lost to follow‐up.
Selective reporting (reporting bias)	Low risk	Comment: matched study protocol. All outcome measures outlined in protocol were reported.
Vested interest bias	Low risk	Comment: no apparent source of bias. The funders of the study did not have a role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Other sources of bias?	Unclear risk	Comment: no other apparent sources of bias