Nezu 1991.

Methods	Randomised controlled trial Setting: study conducted in the USA; participants recruited from an outpatient clinic setting No study dates are given
Participants	Eligibility criteria: mild ID, concomitant mental health problem and maladaptive behaviour (e.g. anger control, verbal or physical aggression, destructive behaviour). No chnages in medication one month prior to participation in the study. Participants were excluded if they had organic brain syndrome, substance misuse, active psychotic sysmptoms or were receiving psychological therapies 28 adults (mean age 36 years) were included; 18 were men and most were White caucasian but two were Black. 9 were receiving antipsychotic medication. Baseline characteristics according to allocation group were not provided
Interventions	Treatment 1: Problem‐solving‐assertiveness (n = 9) Treatment 2: Assertiveness‐problem‐solving (n = 9) Treatment 3: waiting list (n = 10) The interventions were delivered by psychologists and were based on established treatment manuals. For the purposes of the review, the 2 arms of the intervention were grouped as 1: assertiveness training and problem solving The control group went into treatment at the end of the intervention In order to assess adherence to therapy manual, each therapy session was videotaped and independently assessed by 2 research assistants
Outcomes	Brief Symptom Inventory, Problem‐Solving Task, Adaptive Behaviour Scale‐Revised (part II), Role‐Play Test of Anger Arousing Situations, Subjective Units of Distress Assessments were carried out mid‐treatment, post‐treatment, and at three‐month follow‐up (treatment group only and therefore not included in the analysis)
Notes	Participants had comorbid mental disorders but data were available on the behavioural problems separately from the mental disorders. The authors used widely‐known instruments for measuring behaviour change and diagnosing mental illness No information in source of funding No declaration of conflicts of interest statement
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation using coin tossing (This information is not reported in the published paper, but communicated via email to reviewers, Nezu 2000 [pers comm]).
Allocation concealment (selection bias)	Unclear risk	Not reported in the paper; authors contacted but did not respond
Blinding of participants and personnel (performance bias) All outcomes	High risk	No blinding of participants and personnel
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcome assessors are reported as being unaware of the treatment status of participants
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Not reported in paper; authors contacted but did not respond
Selective reporting (reporting bias)	Low risk	We believe that all prespecified outcomes were reported
Other bias	Unclear risk	Effects of co‐morbidity and psychotropic medication on treatment response is unclear