INTRODUCTION
How society should deal with the challenge that arises when crime is com- mitted by individuals with serve mental disorders has been a topic of concern for both the justice and health sectors for centuries. In this paper, serious mental disorders are limited to psychotic disorders, bipolar disorders, severe depression, and severe personality disorders. In recent years, there have been several serious individual incidents in Norway where patients who have been discharged from psychiatry have committed a serious act of violence or mur- der (NCIS 2022). There is broad agreement that neither the police nor the © 2024. Published by AHFE Open Access. All rights reserved. 19
20 Xas
health service can prevent all incidents. However, certain events always raise questions about whether the incident could have been prevented or whether the system may have failed for various reasons.
When serious incidents occur that indicate a system failure, it can often end up with the case being subject to scrutiny by the authorities. A com- mon feature of several of these incidents under investigation by authorities is either inadequate violence risk assessments or absence of such assessments. It is difficult to assess in all cases the significance of inadequate violence risk assessments or their absence on the outcomes of these cases. Nevertheless, this paper assumes that if violence risk assessments are implemented as a tool for preventing violence, the goal must also be to use these tools according to their intended purpose and, importantly, in accordance with procedures and professional practice for them to be effective.
This paper is part of a doctoral project focusing on the prevention of vio- lence and homicide committed by patients with severe mental disorders. The empirical data for the paper is reports from Norwegian oversight authori- ties. The oversight is conducted in cases where individuals discharged from psychiatric care have committed serious acts of violence or homicide in society.
Risk and Risk Assessment Risk is a concept that is defined and used differently in various fields. A well- established approach to risk has long been that risk is an objective char- acteristic of certain phenomena that can be calculated and quantified (Aven 2012). A widely used definition based on the approach based on objectivity is that risk is about the probability that an unwanted event will occur and the consequence of it if the event occurs (probability x consequence) (Renn 2008). Within medicine, risk has been linked to limiting contagion and dis- ease, and risk assessment has been underpinned by a belief in objectivity and safety (Dixon and Oyebode 2007; Renn 2008). However, over time there have been changes both in relation to how risk is understood and how it can be handled.
Violence is not a disorder or disease; it is a human action and a complex interaction phenomenon influenced by many factors (Douglas and Kropp 2002). It is difficult to estimate numerical probability (percentage) and pre- dict future unwanted human actions due to the difficulty in accounting for numerous conditions and external factors. In the context of violence and risk assessment of violence, we need an approach to the concept of risk that is adapted to the phenomenon of violence.
According to Aven (2007) Risk is combination of possible events/conse- quences (outcomes) and associated uncertainty (will the events occur, what will be the consequences) (Aven and Renn 2009). The definition assumes knowledge strength as a measure of uncertainty. While the field of medicine is still dominated by a positivist approach where universal and true predictions are made, social science allows for more uncertainty, in the sense that ’knowl- edge does not equate to truth and certainty’ (Van Asselt, 2000: p. 81 in Mandy Dixon 2007). An understanding of risk where the dimension of uncertainty
Challenges and Uncertainties in Violence Risk Assessment 21 is included acknowledges that it is unavoidable not to consider subjective judgments in a risk management process (Dixon and Oyebode 2007). When assessing the risk of a future violent event, the uncertainty dimension in the concept of risk considers the complexity associated with the phenomenon of violence and the subjective judgments.
Although medicine, including psychiatry, has previously been criticized for not keeping pace with the way uncertainty is included in a risk assessment process, there have also been changes in this field (Dixon and Oyebode 2007).
The prediction of violence among individuals receiving psychiatric and criminal justice services has developed from a basic, unstructured estimation of risk relying on clinical knowledge and intuition, to an ’actuarial’ method centered on static predictors of violence, and further to structured profes- sional judgment (SPJ) (Bjørkly et al. 2014). In SPJ, a compilation of static risk factors is assessed alongside dynamic and individual-specific idiosyncratic factors, guiding the formulation of an individual’s risk of violence (Bjørkly
et al. 2014; Constantinou et al. 2015). The aim of a SPJ is not to specify or quantify who will or will not be violent in the future, but to find concrete strategies that can help prevent an incident of violence from occurring or limit the damage from it if it does occur (Douglas and Kropp 2002). In
SPJ, it is assumed that risk factors associated with violence are expressions of violence at the group level and that they are dynamic; thus, uncertainty associated with these factors is considered, and the validity of assessments is contingent on up-to-date knowledge. Additionally, a risk formulation, risk scenarios, and risk management strategies are developed, rather than quanti- fying an estimate or the probability of violence risk (Norwegian Directorate of Health 2018).
This development of violence risk assessment sheds light on the importance of the dimension of uncertainty in these assessments. Some uncertainty will always be a natural part of these violence risk assessments; it is the type of uncertainty that cannot be controlled due to a lack of knowledge regarding risk factors and the future and uncertainty related to the subjective inter- pretations of the analyst. Beyond this, violence risk assessments also entail another form of uncertainty that is easier to control, but this depends on identifying that type of uncertainty. This uncertainty, for example, may arise from the misuse of violence risk assessment tools. To be able to reduce this type of uncertainty, it is crucial to understand the reasons why violence risk assessment tools are not used according to recommended procedures. This underscores the need to further examine the national recommendations and advice related to violence risk assessments. These recommendations and advice play a crucial role in harmonizing theory and practice.
Method and Analysis Empirical findings are collected though document analysis of mandates and directives for actors involved in handling individuals with serious mental dis- orders. In Norway, supervision of actors in the health service is primarily conducted by the Norwegian Board of Health Supervision. In addition to
22 Xas
the Norwegian Board of Health Supervision, there is also the State Admin- istrator, who also has the authority to supervise the healthcare organization that is subject to their responsibility. The Norwegian Board of Health Super- vision publishes several publicly available reports on their website, but the reports from the State Administrator are not available to the public without a request for access. Therefore, in the process of this paper, which is part of a larger project, requests have been made for access to reports from all state administrators in Norway, with several requests being denied. These reports are subject to confidentiality; hence, they have been redacted to con- ceal sensitive information, and the author has made further restrictions in the reproduction of the data to ensure the anonymity of the individuals involved in these cases. For this paper, thematic analysis has been used to analyse the data. The principles of Clarke & Braun’s (2014) analysis process consisting of several steps are used to identify themes (Braun & Clarke, 2006). The process first started with reading the reports and noting down themes that were recurring. As the themes were reviewed several times systematically, it became possible to create tentative codes. The coding generally focused on violence risk assessments, but later specific codes were additionally added such as for example ‘competence’, ‘use of violence risk assessments’, ‘lack of resources’, ‘training in violence risk assessment’. To ensure quality, parts of the documents were reviewed again after the themes and codes had been established.
The Norwegian Directorate of Health’s Recommendations and advice for Violence Risk Assessments Norwegian health authorities have their own national recommendations and advice for specialized health services for assessing violence risk. The national recommendations and advice by The Norwegian Directorate of Health recommend both the various steps in a violence risk assessment process and provide recommendations on which tools can be used (Norwegian Directorate of Health 2018). As a first step, a conversation between clinician and patient on the subject of violence is recommended. Whether a violence risk assessment should be conducted after the first step depends on identifying violence risk. If it is decided that the patient should be evaluated for violence risk, the authorities recommend the use of the V- RISK-10 tool (Violence risk screening-10), a tool primarily intended for the target group of adults over 18 years of age. The tool is intended for screening or initial mapping of risk, to clarify whether the patient should undergo a comprehensive violence risk assessment (Norwegian Directorate of Health 2018). As part of the risk assessment phase of violence, the authorities also recommend another tool called the BVC (Brøset Violence Checklist) which is also an instrument for screening or initial mapping of risk. This is a tool that can be used to predict violent behaviour in a short-term perspective (the next 24 hours) in adult patients over 18 years of age who are hospitalized. While one tool focuses on predicting violent behaviour within a limited and short time frame (BVC), the purpose of the other is to identify patients who need
Challenges and Uncertainties in Violence Risk Assessment 23 comprehensive violence risk assessment. Both of these two tools are less time- consuming than the tools used for comprehensive violence risk assessment (Norwegian Directorate of Health 2018).
If a comprehensive violence risk assessment of a patient is needed, the authorities recommend using the HCR-20V3 tool, among others. The author- ities emphasize here that collaboration between specialized health services and municipal health and care services is required, in addition to the various levels of specialized health services. Furthermore, it is emphasized that col- laboration with the police is also required. HCR-20V3 is a time-consuming process that requires that the professional is qualified to use the tool either through further education or training and relevant guidance (Norwegian Directorate of Health 2018). What distinguishes HCR-20V3 from the meth- ods of risk assessment is that HCR-20V3 results in a risk management plan used to implement risk-reducing measures (Douglas et al., 2014). In the Directorate of Health’s recommendations, it is discouraged to use HCR-20V3 if there is not enough time or expertise to conduct an assessment following the procedures in the user manual (Norwegian Directorate of Health 2018). According to the Directorate of Health’s recommendations, it is the responsibility of the organization to ensure that healthcare professionals can conduct violence risk assessments in a responsible manner. It is emphasized that the organization must provide training and facilitation (Norwegian Directorate of Health 2018).
Although the guidelines appear clear for assessing violence risk, it is not necessarily easy to implement these recommendations in practice. By analysing individual incidents from the period 2019-2023, we will see how the importance and implementation of the guidelines and recommendations are challenged and how misuse of tools contributes to increased uncertainty in risk assessments.
DISCUSSION AND DATA Violence Risk Assessments in Practice
The mental health services in Norway are organized at two organizational levels (Ruud and Friis 2021). The municipalities are responsible for the pri- mary healthcare service, including general practitioners, team-based primary health substance abuse care. Hospitals and specialized mental health are oper- ated by 19 health trusts, governed by four regional health authorities (Ruud and Friis 2021). The recommendations and advice by The Norwegian Directorate of Health on violence risk assessments are targeted towards specialist health services. Although the primary healthcare services or the municipalities can conduct their own risk assessments of a patient, it is the responsibility of the specialist health services to carry out violence risk assessments.
In several reports, there are concrete examples indicating that the violence risk assessments conducted were deficient or that there was an absence of such assessments. At the same time, in one supervisory case, it was concluded that the violence risk assessment in question had been conducted in accor- dance with national recommendations and advice and recommendations. However, it is argued in this paper that adhering to national recommen- dations and advice does not necessarily equate to good quality when the
24 Xas
guidelines are poorly adapted to practice. In this specific case, the report high- lights that the BVC tool was used to assess the violence risk of an admitted patient. As mentioned earlier, the BVC is a tool used to predict short-term violence (24 hours) for inpatients, which was used while the patient was admitted. Because this patient had a history of serious violence shortly before discharge from residential care, according to the national recommendations and advice and recommendations, it would have been correct to carry out a survey of the risk of violence using the V-RISK-10 tool before discharge. This was not done, and the patient was discharged based on the conclusions of the BVC assessment and other clinical judgments. The patient commits a homicide shortly after discharge. The BVC tool does not provide the same opportunity to analyse the risk factors associated with violence over a long- term perspective, especially those factors associated with environments outside a residential institution. Therefore, relying on responses from a tool designed to predict short-term violence in an institution to make decisions about violence in a longer-term perspective and in other environments out- side the institution would be vulnerable. Such an assessment would entail high uncertainty because the risk-reducing measures implemented are based on incomplete information and a tool that does not consider the patient’s environments outside the institution. The national clinical recommendations suggest that context-independent risk assessments of violence should not be conducted. Several supervisory cases point out that clinicians conducting violence risk assessments have not involved other relevant stakeholders to gather information about the patient that may be useful in the risk assess- ment. Several Norwegian municipalities have criticized that the violence risk assessments conducted by specialist health services are context-independent in the sense that they do not take into account the municipality’s limitations in services and authority to manage a patient with violence risk (Hansen, Makrussen, and Bråthen 2023).
Within specialist health services in Norway, there is a specialized treat- ment sector known as forensic psychiatry. Forensic psychiatry is responsible for patients with aggression or violence issues. Here, you will find patients who are under compulsory treatment either because they have been sentenced by a court after committing a serious act, or because they have been civilly assessed for compulsion due to a high risk of violence (SIFER 2020). Because forensic psychiatry treats this patient group, they also have expertise in var- ious comprehensive violence risk assessment tools such as HCR-20V3. This expertise may not necessarily be as well-implemented in other parts of spe- cialist health services that also encounter patients who may have violence issues. Both the time aspect and expertise are, according to national recom- mendations and advice and recommendations, two crucial factors in using HCR-20V3, but if clinicians outside forensic psychiatry do not have the time or expertise to conduct a comprehensive violence risk assessment, there are no recommendations in the national recommendations and advice on what to do. Supervisory reports contain several examples of deficient violence risk assessments, deficient, for example, in terms of not including information about previous violence history in the risk assessments, or because the person responsible for the violence risk assessment has not contacted the police and
Challenges and Uncertainties in Violence Risk Assessment 25 the municipality to gather information (NBHS 2020, n.d.; The Norwegian Police University College 2022; Ukom 2023). A patient with severe mental disorders with a history of violence often has information recorded in pri- mary health care and with the police. In addition, the municipality typically has information on how this patient is functioning in society and what needs and challenges they have. In a survey conducted in 2023 among 42 Norwe- gian municipalities, only 17 municipalities responded that they are involved in violence risk assessments in specialist health services by obtaining informa- tion about the patient. Most of the municipalities have no experience with involvement (Hansen et al., 2023). An obvious consequence of not obtaining all relevant information when assessing violence risk is that the assessment will less knowledge strength than if the information were obtained. In this way, decisions may be made based on incorrect information.
In some cases, violence risk assessments have not been conducted despite a known history of violence (Ukom 2021). Additionally, in the reports anal- ysed, there are also examples of violence risk assessments being conducted, but no measures are taken to address the risk of violence. Such an assessment holds no value unless attempts are made to manage the risk. Another chal- lenge highlighted in the data analysis is related to the updating of violence risk assessments. Since the assessed risk factors are dynamic, the risk assessments must also be updated to remain effective. Failing to update risk assessments can, in the worst-case scenario, result in a patient receiving inadequate treat- ment or an increase in violence risk going unnoticed. Both deficient violence risk assessments and the absence of such assessments are often justified by a lack of competence. This may contribute to reinforcing a suspicion that there is more competence outside forensic psychiatry on violence risk screen- ing tools than on the tools used for comprehensive violence risk assessments, such as HCR-20V3. If this is the case, there may be a gap between national clinical recommendations and the practice in specialist health services outside forensic psychiatry. It is therefore legitimate to question whether the national recommendations and advice related to violence risk assessment are adapted to the daily practice of clinicians in specialist mental health services. Recently, this topic was problematized in an article published in the Journal of the Nor- wegian Medical Association, where it is also noted that the recommended tools are poorly adapted to the reality outside forensic psychiatry. Further- more, the authors advocate for a revision of the national recommendations and advice for violence risk assessment (Løvgren and Engebretsen 2024). We know based on a review of the national recommendations and advice that clear requirements are associated with collaboration, context-dependent assessments, and sufficient time when conducting comprehensive violence risk assessments such as HCR-20V3, but these requirements are absent for the violence risk screening tools. In addition, an HCR-20V3 results in a risk management plan to be used to prevent violence or reduce its consequences if it occurs. V-RISK-10 does not include such a risk management plan as the tool is meant to be used for assessing violence risk. At the same time, there are no recommendations in the national recommendations and advice on what to do if a clinician has conducted a V-RISK-10 and there is a need for further assessment of violence risk, but the clinician lacks the expertise and time.
26 Xas
CONCLUSION
A clinician operates within the framework of the competence and time avail- able, if the national recommendations and advice are not adapted to this, it will have consequences for both the individual clinician, the patient, and society. If the organization fails to facilitate clinicians in con- ducting violence risk assessments in a manner consistent with the professional recommendations, it is necessary to examine whether the national recommen- dations and advice and requirements are aligned with available resources and expertise. This paper argues that a thorough investigation is needed to deter- mine whether there is a gap between the national clinical recommendations and how clinicians outside forensic psychiatry practice the use of violence risk assessment tools. If this gap is real, either a revision of the national recommendations and advice should be considered, or measures should be implemented to enable clinicians to adhere to the recommendations. This could be a way to reduce the likelihood of clinicians misusing violence risk assessment tools and thereby decrease uncertainty associated with the assessments.
REFERENCES
Aven, Terje. 2012. ‘The Risk Concept—Historical and Recent Development Trends’. Reliability Engineering & System Safety 99:33–44. doi: 10.1016/j.ress.2011.11.006. Aven, Terje, and Ortwin Renn. 2009. ‘On Risk Defined as an Event Where
the Outcome Is Uncertain’. Journal of Risk Research 12(1):1–11. doi: 10.1080/13669870802488883. Bjørkly, Stål, Pål Hartvig, John Olav Roaldset, and Jay P. Singh. 2014. ‘Norwegian Developments and Perspectives on Violence Risk Assessment’. Criminal Justice and Behavior 41(12):1384–97. doi: 10.1177/0093854814547949. Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualita-
tive Research in Psychology, 3(2), 77–101. https://doi.org/10.1191/1478088706q p063oa Constantinou, Anthony Costa, Mark Freestone, William Marsh, Norman Fenton, and Jeremy Coid. 2015. ‘Risk Assessment and Risk Management of Violent Reof- fending among Prisoners’. Expert Systems with Applications 42(21):7511–29. doi: 10.1016/j.eswa.2015.05.025. Dixon, Mandy, and Femi Oyebode. 2007. ‘Uncertainty and Risk
Assessment’. Advances in Psychiatric Treatment 13(1):70–78. doi: 10.1192/apt.bp.105.002022. Douglas, Kevin S., Stephen D. Hart, Christopher D. Webster, Henrik Belfrage, Laura S. Guy, and Catherine M. Wilson. 2014. ‘Historical-Clinical-Risk Management- 20, Version 3 (HCR-20 V3): Development and Overview’. International Journal of Forensic Mental Health 13(2):93–108. doi: 10.1080/14999013.2014.906519. Douglas, Kevin S., and P. Randall Kropp. 2002. ‘A Prevention-Based Paradigm for
Violence Risk Assessment: Clinical and Research Applications’. Criminal Justice and Behavior 29(5):617–58. doi: 10.1177/009385402236735. Hansen, Inger Lise Skog, Tom E. Makrussen, and Kjetil Bråthen. 2023. I grenseland.
Samfunnsvern og psykiatri i et kommunalt perspektiv. 06. Fafo og KS. Nasjonalt kompetansenettverk for sikkerhets-, fengsel- og rettspsykiatri (SIFER).
2020. Sikkerhetspsykiatri i Norge 2019-En Statsurapport. Gaustad.
Challenges and Uncertainties in Violence Risk Assessment 27
Norwegian Directorate of Health. 2018. ‘Voldsrisikoutredning Ved Alvorlig Psykisk Lidelse.’ Retrieved (https://www.helsedirektoratet.no/faglige- rad/voldsrisikoutredning-ved-alvorlig-psykisk-lidelse). The Norwegian Police University College [Norwegian: Politihøgskolen; PHS]. 2022.
Evaluering av PSTs og politiets håndtering i tilknytting til Kongsberg-hendelsen 13. oktober 2021. Oslo, Bodø og Stavanger. The Norwegian Board of Health Supervision (NBHS) [Norwegian: Helsetilsynet].
2020. Tilsynsrapport Etter Alvorlig Hendelse-Svikt i Helsehjelp Ved Akkutinn- leggelse Av Psykotisk Pasient. Helsetilsynet. The Norwegian Board of Health Supervision (NBHS) [Norwegian: Helsetilsynet.
Tilsynsrapport Etter Alvorlig Hendelse. Svikt i Behandlingen Av Dom for Tvun- gent Psykisk Helsevern. Tilsynsrapport. The National Criminal Investigation Service (NCIS) [Norwegian: Kripos]. 2022.
Vold Begått Av Personer Med Alvorlige Psykiske Lidelser. Løvgren, Pia Jorde, and Sondre Sperle Engebretsen. 2024. ‘Nasjonale faglige råd for voldsrisikoutredning bør revideres’. Tidsskrift for Den norske legeforening. doi: 10.4045/tidsskr.24.0007. Renn, Ortwin. 2008. Risk Governance: Coping with Uncertainty in a Complex
World. London; Sterling, VA: Earthscan. Ruud, Torleif, and Svein Friis. 2021. ‘Community-Based Mental Health Services in Norway’. Consortium Psychiatricum 2(1):47–54. doi: 10.17816/CP43. Statens undersøkelseskommisjon for helse- og omsorgstjenesten (Ukom). 2021.
Hva Kan vi Lære Etter et Drap Begått i Psykotisk Tilstand? Helsehjelp Og Samfunnsvern for Pasienter Med Psykose Og Økt Voldsrisiko. 5. Statens undersøkelseskommisjon for helse- og omsorgstjenesten (Ukom). 2023.
Helsehjelp Til Personer Med Alvorlige Psykiske Lildelse Og Voldsrisiko.