Journal Logo


Violence Risk Assessment in Clinical Settings: Enduring Challenges and Evolving Lessons

Pinals, Debra A. MD

Author Information
Harvard Review of Psychiatry: 1/2 2021 - Volume 29 - Issue 1 - p 90-93
doi: 10.1097/HRP.0000000000000279
  • Free


Scenario 1: An 18-year-old man with no psychiatric history is brought by police to the emergency department at the insistence of his family after he punched a hole in the wall and was yelling at voices that only he could hear.

Scenario 2: A 28-year-old woman on probation with a long history of depression and repeated appearances in jails, substance use facilities, and psychiatric inpatient units was admitted to a substance use program as an alternative to incarceration after overdosing on heroin. On admission, she frequently exhibited angry outbursts and once became aggressive. She has a history of early trauma, multigenerational criminal justice involvement, and homelessness.

Scenario 3: A 37-year-old woman was acquitted as not guilty by reason of insanity on a charge involving the killing of her baby six years prior. She has mild intellectual disability with a psychotic illness. At the time of the offense, she was experiencing delusional beliefs that her baby needed saving even as voices told her to drown the child. She is now being evaluated for discharge.

Scenario 4: A 58-year-old man works for a manufacturing company and is increasingly angry and sullen about a promotion that never came. He also is facing divorce and has been drinking more heavily. As his sense of injustice builds, he is asked to see a psychiatrist after he made some explicit threats toward his employer.

Societal views about an association between mental illness and violence cut across cultures and centuries.1 Modern research has attempted to decrease stereotypes and myths about this relationship and has illuminated many subtleties in the antecedents of violence. Data support that mental illness alone is not society’s greatest threat when it comes to violence. Violence resulting from mental disorders represents only 3% to 5% of violence in the United States.2 The risk of violence among persons with mental disorders increases, however, when combined with other factors, especially substance use.3 Variables such as early trauma and childhood conduct disorder are also linked to violence and further complicate the relationship between mental illness and violence, given that individuals with mental illness can be at greater risk of having some of these underlying problems.4

Notwithstanding the many factors that can play into a violent act or series of acts, decisions—for example, whether to discharge patients from the emergency room or seek to hospitalize them, to intervene with patients who reveal violent thoughts during psychotherapy sessions, or to conditionally release insanity acquittees or seek mandated outpatient civil commitment—require application of risk concepts to everyday clinical decisions. As an example, to effectuate a civil commitment it is necessary to determine, after a proper clinical assessment, whether a particular patient meets the state statutory criteria for commitment, usually by examining whether the symptoms of mental illness lead to sufficient risk of harm to justify the involuntary commitment intervention.5 Such risk assessment might also result in determinations that other persons need to be contacted for their own protection—a legally permitted breach of confidentiality.6 Initial and subsequent decisions related to risk management can change the outcome trajectory for all involved.

Today, as the above scenarios illustrate, psychiatrists are called upon to assess the risk of violence posed by individuals across a variety of clinical and community settings, and to help mitigate and manage the risks identified. The enduring real-world challenge for clinicians is that serious violent behavior is a rare, but costly, adverse outcome that is associated with many nonspecific risk factors; violence is both difficult to ignore and almost impossible to predict. The stakes are high, and mental health professionals are under increasing pressure to be soothsayers, fortune tellers, and agents of control in order to protect the public from all manner of harms—even those for which they may have no special power to predict or eradicate. This article reviews advances from research on assessing the risk of violence, and it highlights the ongoing need for incorporating thoughtful clinical formulations of violence risk into its management across settings.


Violence risk assessments across most clinical settings incorporate years of research and clinical advances. Early approaches utilized unstructured and less-informed clinical methods. Today’s routine comprehensive clinical mental health evaluations factor in an individual’s background, including understanding both the recognized “static,” or fixed, risk factors associated with violence risk and the “dynamic” variables that can be modified to reduce risk. Prototypical, clinically relevant violence-associated risk factors include the following: past history of violence, prior arrest history, young age at first arrest, history of substance use, prior fire setting, animal cruelty, risk-taking behavior, impulsivity, current mental state circumstances, nonadherence to treatment, access to weapons, dynamics in the household, viewing oneself as a victim, lack of empathy, overt intention to harm, and lack of concern over violent act.7 In acute settings, hostility and suspiciousness might signal risk of acute aggression.7

Research in the field of violence risk assessment has taken steps to examine approaches that go beyond a basic mental health evaluation. This effort has resulted in the promulgation of a range of violence risk assessment instruments (VRAIs) and scales to aid in the work. These have included highly structured actuarial approaches with algorithms for scoring risk factors that calculate stratified risk levels, as well as empirically driven structured guides to aid professional judgment.

The Violence Risk Appraisal Guide (VRAG-Revised [VRAG-R]), one example of an actuarial tool, is applicable to the assessment of risk of violence and sexual offending.8 Another actuarial tool that uses a rigorous analysis of weighted variables to generate a risk level is the Classification of Violence Risk (COVR). This tool was designed to examine the risk of violence after discharge of psychiatrically hospitalized patients. It does so by analyzing numerous research-supported risk factors gathered through a review of medical records and a brief interview of the patient. The patient’s risk information is entered into a software program that runs iterative algorithms to generate information about whether the patient presents a high or low risk of violence after discharge.9

With the advent of actuarial, statistically driven violence risk assessment measures and research on their predictive powers, it has been debated whether they are superior to clinical judgment in assessing violence risk. Some authors argue that clinical methods without structure are susceptible to bias, oversight of important risk factors, and lack of reliability and validity.10 Others have found actuarial assessments to be limited in that they yield risk stratification but without incorporating contextual and dynamic shifts relevant to an individual’s risk presentation.11 That said, the actuarial instruments were not intended to design interventions in treatment, and recent studies have highlighted that although these tools identify risk levels, they do not provide all that is needed for individualized treatment and release planning.11–13 Still, clinicians in specific settings, especially within correctional and forensic systems, might use actuarial-type VRAIs to add additional data in the risk assessments needed for decision making.

Another approach to violence risk assessment involves structured professional judgment (SPJ). This method calls for gathering data on empirically recognized risk factors associated with violence and for incorporating it into risk assessment and risk-management planning.14 One measure that follows the SPJ approach is the Historical Clinical Risk Management (HCR)–20, Version 3, which guides clinicians through historical, clinical, and risk-management variables known to be correlated with violence risk to help clinicians develop a violence risk formulation.15 In the HCR-20 Version 3, historical items include a history of violence and antisocial behaviors, relationship history, history of serious mental illness, substance use disorder, trauma history, and prior response to supervision, among other factors. Clinical items include insight, violent ideas and intent, current symptoms of major mental illness, and instability, while risk-management variables include items such as whether the individual will be connected to professional and personal supports, or have a problematic living situation. The third version of the HCR-20 also offers a scenario-analysis model to help the evaluator consider how risk might be higher or lower, depending on plans and stressors that the individual may face.

Although a full review of all available tools is beyond the scope of this overview, a range of other tools, in addition to the VRAIs described above, are designed for various uses. For example, the Dynamic Appraisal of Situational Aggression (DASA) is a brief tool used to help with short-term assessment of the risk of harm through analysis of key items that may signal more immediate concerns.16 Its brevity and design makes it useful in settings such as emergency departments. The Brøset Violence Checklist (BVC) has been shown to be helpful in examining the risk of inpatient violence;17 it can provide a more structured way to help staff assess patient progress and behavior. Other nuanced or specialized types of risk assessment tools include those that look at specific risk behaviors, such as stalking, sex offending, or domestic violence, or that examine violence risks among specific populations such as youth, older adults, veterans, or individuals in the criminal justice system. Although this column focuses on assessing the risk of violence toward others, much has also emerged in the discourse on suicide risk assessment,18 which is an important element in assessing a range of risks that a patient may present that could be associated with violence.

VRAIs have advanced our understanding, but they are appropriately applied only to populations and contexts upon which they are normed, and they should be interpreted only within the limits of their reach. As noted by Buchanan and colleagues,7 they can help aid practitioners to take into account the myriad variables correlated with violence in particular situations, but no one instrument or technique will accurately identify only those who would act violently, and no one approach fits across all settings. That said, future research will continue to help elucidate strategies for risk assessments, especially by incorporating dynamic risks and approaches for mitigating those risks. Technology may also have a role in advancing our ability to assess violence risk blindly and in the absence of bias, such as through artificial intelligence or machine learning.

In the meantime, in everyday clinical practice settings, decisions about admission, discharge, outpatient commitment, and level of care needed to manage risk must rely upon clinical assessments of violence risk, generally (except in forensic-type settings) in the absence of in-depth, structured VRAIs. Therefore, training general mental health practitioners to understand the field of violence risk assessment and how to address factors associated with violence remains a priority. Most notably, general clinicians need to be able to talk to patients who may have violence histories and to ask questions that elucidate any violent thoughts and fantasies that the patients may harbor.


As noted above, the impact of decisions that might follow violence risk assessments are serious for the individual being assessed and for society. Thus, practitioners need to incorporate a number of considerations into their clinical approaches, including potential ethical and other challenges. Practitioners need to balance beneficence and nonmaleficence, preserve autonomy and the therapeutic alliance insofar as possible, and act to prevent harm when public safety could otherwise be compromised. The clinician’s actions that flow out of this balancing can have major consequences, such as compromised confidentiality, new or sustained compulsory hospitalization, or court-ordered outpatient treatment. Each of these actions can have further downstream effects such as disrupted relationships, termination from employment, removal from school, or arrest, to name a few.

Studies pointing to the accuracy of risk assessment across populations have shown that clinicians may do better than chance, especially in short-term predictions, but that there is a risk of overpredicting risk, with high false-positive rates or findings of high risk when the risk is actually not high.7,19 The complexity arises when one considers how incorrect prediction of risk might result in the involuntary commitment of individuals who would not have become violent or how the decision to release someone might result in an act of serious violence that could have been averted. Often, especially when risk assessment data are somewhat murky, the clinical risk analysis involves “erring on the side of safety” and making decisions that give more weight to preventing harm to others than to individual liberty interests of the patient; unfortunately, such a approach entails accepting some degree of error rate—which might result in an action counter to the patient’s wishes.7

Another social challenge in violence risk assessment involves principles of distributive justice that are potentially also implicated in the disproportional allocation of intensive mental health services to putatively violent patients, many of whom may not actually be violent in the long run. For example, one study found that individuals who were treated in the forensic system remained hospitalized longer despite being thought to be at no greater risk than a similar group treated in the civil system,20 resulting in the use of hospital resources that perhaps could have been more appropriately offered to those in greater need. Clinicians looking at the person level may not be tuned into the issues of distributive justice, given that they make decisions regarding individual cases and may not see the overarching systems considerations. It therefore remains possible that the resources used in relation to mental health violence risk assessments (including indications for, and use of, further measures or resource-intensive treatments) could potentially be better used to address other social issues or other populations.

Another serious concern related to violence risk assessment is the potential for biases, such as those regarding race, entering into risk determinations. For example, it is commonly recognized that structural racism is present in both the health care and criminal justice systems—which has recently come into particular focus worldwide. Research indicates that minority populations are more likely to be classified as high risk and thereby subject to outcomes that include being sent to more inherently coercive or secure settings.21 Ironically, even some structured risk assessment tools, developed and promulgated as “objective” and “unbiased,” are now being critiqued for being built upon a foundation of variables fraught with biases and other structural limitations.22 These challenges highlight the importance of keeping in mind the advantages and limitations of particular approaches to violence risk assessments and the consequences of these assessments.


To return to the scenarios that opened this column:

Scenario 1: It may be useful to screen for acute violence risk. Hospitalization will likely be needed, and the clinician may need to pursue a civil commitment petition, followed by ongoing risk mitigation with family and supports. The work in the emergency room with this young man may set him on a positive trajectory for dealing with a lifelong psychiatric illness.

Scenario 2: For this woman, a risk assessment might reveal the need for trauma-informed therapeutic approaches and substance use treatment at the proper level of care, including medications to address her likely opioid use disorder and the provision of housing supports to help reduce the likelihood of environmental stressors that could increase her volatility.

Scenario 3: Through a clinical risk assessment that included the use of structured professional judgment tool, it was determined that this woman’s current risk was low. Her discharge will require supportive staff to foster medication adherence and mental status monitoring, with proper positive functional supports for managing whatever risks may emerge.

Scenario 4: A psychiatric assessment revealed that this man needed intensive therapy to help him cope. It also revealed that he owned firearms and had no real intent to use them, but in discussing his anger, he agreed that it was best for him to have his firearms removed. While identifying his sadness over his impending divorce, he desired to focus on positive goals such as reducing his alcohol use, maintaining income through other employment, and rebuilding a positive relationship with his adult children.

The case scenarios above had little in common other than an individual at the center who might have dormant or active mental illness and present some risk. No one violence risk assessment tool is capable of identifying all the nuances of their risk or of sorting through how to address that risk. Still, guides or instruments may be helpful in analyzing the risk and assisting in various clinical situations. Rather than assess risk for the purpose of predicting violence—and then investigating how such cases play out in the real world—clinical approaches need to focus on assessing patients and on then developing strategies to mitigate and manage risk. Any clinical situation can unfold differently, depending on approaches used, along with any related decisions and responses. Risk mitigation therefore requires evolving decisions informed by outcomes of prior decisions.

In the clinical care of patients and in assessing risk, clinicians should consider—in addition to asking about violent ideation and intent—an individual’s self-perception of his or her risk.23 Clinicians should also take into account other factors such as the individual patient’s mental capacity to make reasoned choices to engage in safe behavior versus their tendency toward impulsivity. Clinicians should assess, too, the individual’s access to weapons and intent to use them, the operative dynamics of anger, shame, desperation, and fear, and the motivation to refrain from acting, or to act, in a violent way. In approaching plans to address and mitigate risk, clinicians should ascertain the individual’s alliance with a treatment provider and available community supports. Taking all that information into account and then selecting the most effective intervention to reduce risk can be a challenge in real-world settings, where the options and resources may be limited.

The weighty task of mental health professionals guarding public safety while delivering care cannot be robotic; it must be informed by critical thinking in clinical contexts and by evolving research on violence risk assessment. Mental health clinicians faced with real-world situations involving violence risk assessment will continue to need to balance tensions between safety, coercion, and fairness as they incorporate ethically, clinically, and empirically sound approaches that will often pivot around statutory and regulatory obligations and limitations as applied to individual circumstances. Case-based scenarios can help practitioners learn to weigh contextual, environmental, personal, and cultural variables that might contribute to violence, as well as the clinical, societal, and ethical challenges involved in managing it. Given the important role of clinical violence risk assessment, discourse on these complicated concepts needs to continue assisting mental health professionals in mitigating the risk of harm to others while supporting individuals in care.

Declaration of interest: The author reports no conflicts of interest. The author alone is responsible for the content and writing of the article.


1. Monahan J. “A terror to their neighbors”: beliefs about mental disorder and violence in historical and cultural perspective. Bull Am Acad Psychiatry Law 1992;20:191–5.
2. Swanson JW. Mental disorder, substance abuse, and community violence: an epidemiological approach. In: Monahan J, Steadman H, eds. Violence and mental disorder. Chicago: University of Chicago Press, 1994:101–36.
3. Steadman HJ, Mulvey EP, Monahan J, et al. Violence by people discharged from acute psychiatric inpatient facilities and by others in the same neighborhoods. Arch Gen Psychiatry 1998;55:393–401.
4. Elbogen EB, Johnson SC. The intricate link between violence and mental disorder: results from the National Epidemiologic Survey on Alcohol and Related Conditions. Arch Gen Psychiatry 2009;66:152–61.
5. Pinals DA, Mossman D. Evaluation for civil commitment. Best practices in forensic mental health assessment. New York: Oxford University Press, 2012.
6. Appel JM. Trends in confidentiality and disclosure. Focus 2019;17:360–4.
7. Buchanan A, Binder R, Norko M, Swartz M. Resource document on psychiatric violence risk assessment. 2012 [online supplement to Buchanan Alec, Binder R, Norko M, Swartz M. Psychiatric violence risk assessment. Am J Psychiatry 2012;169:340].
8. Harris GT, Rice ME, Quinsey VL, Cormier CA. Violent offenders: appraising and managing risk. 3rd ed. American Psychological Association, 2015.
9. Monahan J, Steadman HJ, Appelbaum PS, et al. The classification of violence risk. Behav Sci Law 2006;24:721–30.
10. Harris GT, Rice ME, Cormier CA. Prospective replication of the Violence Risk Appraisal Guide in predicting violent recidivism among forensic patients. Law Hum Behav 2002;26:377–94.
11. Douglas KS, Kropp PR. A prevention-based paradigm for violence risk assessment: clinical and research applications. Crim Justice Behav 2002;29:617–8.
12. Fazel S, Singh JP, Doll H, Grann M. Use of risk assessment instruments to predict violence and antisocial behavior in 73 samples involving 24,827 people: systematic review and meta-analysis. BMJ 2012;345:e4692.
13. Wong SCP, Gordon A. The validity and reliability of the Violence Risk Scale: a treatment friendly violence risk assessment tool. Psychol Public Policy L 2006;12, 279–309.
14. Melton GB, Petrila J, Poythress NG, et al. Psychological evaluations for the courts: a handbook for mental health professionals and lawyers. 4th ed. New York: Guilford, 2018.
15. Douglas KS, Hart SD, Webster CD, Belfrage H. HCR-20V3: assessing risk of violence—user guide. Burnaby, BC, Canada: Mental Health, Law, and Policy Institute, Simon Fraser University, 2013.
16. Ogloff JRP, Daffern M. The dynamic appraisal of situational aggression: an instrument to assess risk for imminent aggression in psychiatric inpatients. Behav Sci Law 2006;24:799–813.
17. Abderhalden C, Needham I, Dassen T, Halfens R, Haug HJ, Fischer J. Predicting inpatient violence using an extended version of the Brøset-Violence-Checklist: instrument development and clinical application. BMC Psychiatry 2006;6:17.
18. Pinals DA. Liability and patient suicide. Focus 2019;17:349–54.
19. Kisely SR, Campbell LA. Compulsory community and involuntary outpatient treatment for people with severe mental disorders. Cochrane Database Syst Rev 2014;(12):CD004408.
20. Levitt GA, Vora I, Tyler K, Arenzon L, Drachman D, Ramos G. Civil commitment outcomes of incompetent defendants. J Am Acad Psychiatry Law 2010;38:349–58.
21. Pinals DA, Packer IK, Fisher W, Roy-Bujnowski K. Relationship between race and ethnicity and forensic clinical triage dispositions. Psychiatr Serv 2004;55:873–8.
22. Singh JP, Grann M, Fazel S. A comparative study of violence risk assessment tools: a systematic review and metaregression analysis of 68 studies involving 25,980 participants. Clin Psychol Rev 2011;31:499–513.
23. Skeem JL, Manchak SM, Lidz CW, Mulvey EP. The utility of patients’ self-perceptions of violence risk: consider asking the person who may know best. Psychiatr Serv 2013;64:410–5.

precision medicine; psychiatric settings; risk of harm; violence risk assessment

© 2021 President and Fellows of Harvard College