Risk assessment in forensic settings — structured clinical viva

Do not answer "is he safe?" with a binary. Safety is context-specific. Translate into: residual risk of which harms, to whom, under which conditions, with which mitigations, and with what residual uncertainty.[1][5][6]

SPJ vs actuarial. SPJ (HCR-style) organises Historical, Clinical, and Risk management information into scenarios and a plan. Actuarial tools (VRAG-class) may keep him in a high static band because history does not erase — that is expected and does not alone forbid progress if dynamic and protective factors have changed. Metas show moderate group accuracy only.[1][2][4]

Protective factors. Work programme, family contact, sustained abstinence, and improved clinical items are meaningful. Research links improvements in dynamic risk and protective factors during forensic treatment to lower post-discharge recidivism. Use these to justify graduated testing, not unrestricted freedom.[3]

Horizons. High-to-medium security is still institutional; short-horizon aggression tools remain relevant on the ward, while leave and later community release need longer SPJ and substance/housing planning.[4]

Scenarios. Example: arson or assault if cannabis relapse + command hallucinations return + unsupervised access to vulnerable housing. Mitigation: no unescorted leave until longer abstinence; drug screening; robust relapse plan; victim/neighbour considerations; multi-agency information sharing principles if community risk rises.[4][5]

Anti-stigma. Schizophrenia elevates relative odds of violence in meta-analysis but most patients are never violent; here risk is driven by history plus specific dynamic pathways (commands, substances, arson method), not diagnosis as destiny.[7]

Report language. State method, sources, scenarios, recommendations, limitations, and review triggers. Avoid inventing statute numbers. Avoid false precision from scores.[5][6]