BackgroundWe set out to assess the appropriateness of current placement of mentally disordered offenders allocated by the courts in Italy to REMS or to forensic community residences. We hypothesised that as in other countries, the match between a standardised assessment and the decision of the court would be imperfect.MethodsThe DUNDRUM Toolkit was translated into Italian. The translation had good psychometric properties. In order to compare the current level of therapeutic security with a calculated safest current placement, we compared the DUNDRUM-1 triage security assessment of need for therapeutic security prior to treatment, with evidence for progress made in treatment (DUNDRUM-3) and forensic recovery (DUNDRUM-4). The more conservative of these two would be taken as the safe current level of need for therapeutic security.ResultsThe Italian translation of the DUNDRUM Toolkit had good internal consistency and mean scores had a Reliable Change Index less than one unit. 3.7% of those in REMS (medium security) were assessed as needing high security and 38% were ready to move to a less secure place. In low secure places, 56% were assessed as needing a higher level of therapeutic security and 6% could have moved to open non-secure places.ConclusionsThe Italian translation of the DUNDRUM Toolkit allows an assessment of the current working of the model of care for forensic psychiatry following the reforms of 2015. Most patients are safely placed. A small but important proportion needed high secure places that are not currently available. (3.7% of 604 nationally, 95% Confidence Interval 1.2% to 8.4%, 7 to 50). A greater use of such measures would enable better health gains and safer outcomes.Trial registration ClinicalTrials.gov ID: NCT06018298 Unique Protocol ID: ITAL-EE-REMS.