Willner 2013.

Methods	Cluster‐randomised controlled trial of 30 centres (15 experimental and 15 control) Setting: study conducted in the UK; day centres for people with intellectual disability Study was carried out between December 2009 and January 2012 Centres were randomised according to "minimisation" with a random component set at 80% Centres were balanced on 3 key variables (service users' mean self‐reported Provocation Inventory score, number of participants, and mean number of hours spent by service user with at least 1 lay therapist outside of sessions)
Participants	Centres were eligible if: there were anger control problems amongst at least 4 service users had at least 2 members of staff willing to be "lay therapists" no current anger management interventions 179 participants with mild or moderate intellectual disabilities were included in study (90 in intervention arm and 89 in control arm) Participants were excluded if they required urgent psychological treatment for anger or aggression Intellectual ability assessed using WASI and BPVS Median full‐scale IQ 57 Median age 38 71% men
Interventions	Treatment group: manualised group cognitive‐behavioural anger management therapy Training (3 sessions over the course of 1 day), based on a manual developed for the intervention, was provided to "lay therapists" (e.g. support staff), who delivered the therapy Lay therapists were supervised by a psychologist (fortnightly supervision) 52 sessions were monitored for adherence to the manual and inter‐rater reliability was assessed (83%). Cohen's kappa of 0.65 is reported overall Control group: "treatment‐as‐usual", waiting list control group
Outcomes	Primary outcomes: Self‐reported Provocation Index (PI) Score Secondary outcomes: PI completed by staff Profile of Anger Coping Skills completed by service user, key worker, and home carer Aberrant Behaviour Checklist, completed by key worker and home carer Modified Overt Aggression Scale, completed by key worker and home carer Controllability Beliefs Scales completed by key worker Other secondary measures included quality of life, depression, anxiety, self esteem, and service resource utilisation (using CSRI)
Notes	Relatively large study incorporating a large number of centres Funding by the NIHR Health Technology Assessment (HTA) programme Declaration of conflicts of interest: nil
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was based in 'minimisation' with a random component set at 80%
Allocation concealment (selection bias)	Low risk	Not an issue as clusters were all randomised at the same time
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants and study personnel were not blinded to the allocation group. Key worker reports were provided by staff who had not been the 'lay therapist' but this was not possible in all cases (n = 14)
Blinding of outcome assessment (detection bias) All outcomes	Low risk	The assessors were blinded to the allocation group but the authors report "although this may have been compromised by incidental comments from respondents during data collection". However, the authors believe that this risk was small
Incomplete outcome data (attrition bias) All outcomes	Low risk	Number of participants who withdrew from study are reported. At 10 months retention rate was 81% (projected 80%) with same number of dropouts in both arms (17 dropouts). Results were analysed according to intention‐to‐treat. 80% of key workers retained at 10‐month follow‐up but only 60% of home carers were retained at follow‐up
Selective reporting (reporting bias)	Low risk	All prespecified the outcomes were reported
Other bias	Unclear risk	Authors had a strong allegiance to the intervention and their desire for the intervention to succeed could have affected data collection, but the likelihood of this impacting the results is unclear

BPVS: British Picture Vocabularly Scale.
 CSRI: Client Service Receipt Inventory.
 IQ: intelligence quotient.
 MHA: Mental Health Act.
 NHS: National Health Service.
 NIHR: National Institute for Health Research.
 SD: standard deviation.
 WAIS: Wechsler Adult Intelligence Scale.
 WAIS‐III: Wechsler Adult Intelligence Scale, Third Edition.
 WAIS‐R: Wechsler Adult Intelligence Scale, Revised.
 WASI: Wechesler Abbreviated Scale of Intelligence.