Cognitive skills programmes have been associated with improvements on psychometric measures and reductions in antisocial behaviour in mentally disordered offenders (MDOs). However, to date there have been no randomized controlled trials (RCTs) of such programmes with this population. In the first RCT of a cognitive skills programme with MDOs we aimed to determine if participation in the Reasoning and Rehabilitation (R&R) programme was associated with improvements in social–cognitive skills and thinking styles.

A total of 84 men with a primary diagnosis of psychotic disorder and a history of violence were recruited from medium-secure forensic units and allocated to receive R&R (n=44) or treatment as usual (TAU; n=40). At baseline and post-treatment interviews, participants completed questionnaires to assess social problem-solving, criminal attitudes, anger experience, blame externalizing and perspective-taking. Researchers were not blind to group status.

The R&R group demonstrated significant improvements on measures of social problem-solving relative to the TAU group, some of which were maintained at 12 months post-treatment. Only half of those allocated to receive R&R completed the full programme. In post-hoc analyses programme completers showed improvements in social problem-solving at the end of treatment and changes in criminal attitudes at 12 months post-treatment.

Among male MDOs, R&R participation was associated with improvements in social–cognitive skills, some of which were maintained for up to 12 months post-treatment. Our finding that programme completers do better may reflect pre-treatment patient characteristics. This study establishes that multi-site RCTs can be conducted in medium-secure forensic units.