The dangers of using diagnoses outside of established psychiatric nosology in the courtroom: Analysis and discussion of current Swiss legal precedent from a medical perspective.

Akin to many jurisdictions, Switzerland has a dual system of sanctions comprising sentences and measures. To order a therapeutic measure per Article 59 or 63 of the Swiss Criminal Code, the presence of a "severe mental disorder" must be determined. Before the new legal precedent, this required a medical diagnosis according to recognised classification systems like the ICD or DSM. The court then decided if a disorder was "severe" in the legal sense, thereby requiring such a therapeutic measure. However, in two 2019 rulings, the Swiss Federal Supreme Court concluded that a severe mental disorder could legally exist without a diagnosis according to the ICD or DSM, if it is based on offence- and risk-relevant personality-related factors amenable to risk-reducing therapy. We examine the details and context of the rulings, alongside their wider dangers. Specifically, we outline how undue influence could be exerted by non-ICD/DSM diagnostic systems, which were developed within individual theoretical schools of thought and lack empirical validation, like in these two court cases. Such non-manual diagnoses could make the presence of a severe mental disorder dependent upon whether an expert witness employs a particular diagnostic system, which would undermine principles of legality. Moreover, the Court's requirement that the disorder is based on personality-related risk factors amenable to risk-reducing therapy is problematic because research has highlighted the low effectiveness of treatment provided independently of a psychiatric disorder. Finally, broadening entry criteria may increase the number of offenders who require psychiatric treatment, thus endangering the quality of care for those with ICD/DSM-based diagnoses that are known to respond well to treatment (e.g. schizophrenia). In short, fulfilling the Court's request that any non-manual diagnoses are based on personality-related risk factors that are amenable to risk-reducing therapy is not possible for such non-manual diagnoses. Using unvalidated diagnoses could also render the system susceptible to ethical issues and hypothetical misuse, which may adversely affect society's most vulnerable people. To counter these dangers, we suggest that in order to be admissible in court, any diagnostic system must mandatorily fulfil sufficient scientific standards.