The Historical Clinical Risk Management-20 (HCR-20) is a commonly used violence risk assessment measure belonging to the Structured Professional Judgment model of violence risk assessment that sets out 20 risk factors across three scales: Historical, Clinical, and Risk Management. Violence risk assessment plays a pivotal role within criminal and forensic psychology, in that it is used to determine the level of violence risk posed by persons typically under some form of legal supervision or incapacitation. For instance, it is commonly used at admission to, within, and discharge from prisons, forensic hospitals, and psychiatric facilities. This article provides an overview of the HCR-20, including a brief description and history, its purpose, its applications, its core assumptions, its contents, how it was developed, how it is used in practice, and the nature of its research support.
Brief Description and History of the HCR-20
The HCR-20 is a violence risk assessment measure intended to be used by professionals who are responsible for making decisions about risk of violence posed by people typically under some form of legal supervision. The 20 risk factors are broken down across the scales as follows: Historical (10 risk factors), Clinical (5 risk factors), and Risk Management (5 risk factors). The HCR20 User Manual details instructions for administering the measure, coding each risk factor, and making decisions about level of risk and necessary risk management strategies. The first version of the HCR-20 was published in 1995, replaced shortly thereafter with a slightly revised Version 2 in 1997. In 2013, a more substantially revised Version 3 was published. There have been 20 translations of Version 2, with 12 translations completed or underway for Version 3 as of late 2016.
The Purpose of the HCR-20
First, the HCR-20 is intended to comprehensively structure the violence risk assessment procedure from the gathering of information to the final estimate of risk level. Second, it is intended to yield summary judgments regarding a person’s level of risk posed. Third, it facilitates a formulation of violence risk (an integrative, coherent explanation of a person’s violence), which allows evaluators to answer the question of why a person has acted violently. Fourth, it leads evaluators through a systematic consideration of necessary risk management strategies required to mitigate violence risk. Ultimately, its purpose is to prevent violence.
Intended Target for HCR-20
The HCR-20 is commonly used for persons under legal supervision for whom violence risk is of some concern or must be determined as a matter of law or ethics. Some of the traditional contexts in which it is used include corrections, forensic mental health, and civil psychiatry. When a person enters any of these settings, determinations of risks posed and management needs often must be addressed. For persons residing within correctional facilities, or forensic or civil hospitals, it is used to estimate risk level, to form risk reduction plans, and to monitor risk and adjust management plans over time. It is also frequently used to inform release decisions in terms of degree of risk posed and management needs if a person were to be released into the community. It also can be used for people being supervised in the community, to establish management plans and monitor risk over time.
There are additional applications of the HCR20. For instance, law enforcement agencies may use it as part of their work managing cases. It has been used in workplace settings as well, under occupational health and safety laws and regulations that require risk assessments to be conducted if the health and safety of employees are threatened. Clinicians in private practice may encounter situations in which a client poses a serious risk of harming third parties, and hence, they may have a legal duty to protect those third parties. Risk assessment is a vital component of this process.
Core Assumptions of the HCR-20
There are a number of assumptions that have gone into the design and implementation of the HCR20. First, trained professionals with appropriate education and expertise, within a structured framework, can make reliable and informed decisions about risk factors, risk level, and risk management needs. Also, the majority of its risk factors are dynamic, or modifiable, and hence can serve as appropriate intervention opportunities. Additionally, the majority of its risk factors can act as causal agents of violence at the idiographic level, and hence, if they are targeted for intervention, risk will decrease. Finally, although in general the greater number of risk factors present in a given case, the higher the risk posed, risk is not necessarily a simple linear function of the number of risk factors. Hence, professionals who use the HCR-20 are assumed to be able to exercise appropriate discretion in determining risk level through not only the consideration of the number of risk factors present, but their relevance, salience, and interrelatedness at the idiographic level. As such, decisions of high risk are permissible even with few risk factors present, as are decisions of low risk in the face of many risk factors. These assumptions have been tested and supported by research.
Contents of the HCR-20
The HCR-20 is 130-page professional manual or user guide. Chapter 1 of the user guide reviews the history and foundation of the HCR-20; situates the HCR-20 in the larger risk assessment field, including how it differs from an actuarial approach; and defines and explains key concepts, such as violence, risk, assessment, and management. Chapter 2 describes the revision process used for the development of Version 3 of the HCR-20. The third chapter includes detailed instructions on administering the HCR-20. Chapter 4 contains definitions and coding notes for each of the 20 risk factors. The Appendices of the HCR-20 contain samples of the HCR-20 Worksheets.
The risk factors are divided into three scales. The Historical Scale has 10 risk factors, denoted as a History of Problems with: (1) Violence, (2) Other Antisocial Behavior, (3) Relationships, (4) Employment, (5) Substance Use, (6) Major Mental Disorder, (7) Personality Disorder, (8) Traumatic Experiences, (9) Violent Attitudes, and (10) Treatment or Supervision Response. The Clinical Scale has 5 risk factors, denoted as Recent Problems with: (1) Insight, (2) Violent Ideation or Intent, (3) Active Symptoms of Major Mental Disorder, (4) Instability, and (5) Treatment or Supervision Response. The Risk Management Scale includes five risk factors, denoted as Future Problems with: (1) Professional Services and Plans, (2) Living Situation, (3) Personal Support, (4) Treatment or Supervision Response, and (5) Stress or Coping.
The Historical Scale captures past functioning, behaviors, and experiences. The Clinical Scale pertains to recent (past 6–12 months) affective, cognitive, and behavioral functioning, and the Risk Management Scale refers to future adjustment and experiences in the coming 6–12 months.
Development of the HCR-20
The HCR-20 shares development principles with all Structured Professional Judgment measures. Risk factors were selected for the HCR-20 (be it Version 1, 2, or 3) based on comprehensive literature reviews and consultation with content area experts. This method is referred to as the logical or rational item selection method. It is intended to produce an evidence-based, comprehensive set of risk factors that is generalizable across settings. This method stands in contrast to how some other measures are developed, using statistical methods (such as linear or logistic regression) to extract a set of risk factors from one or (less commonly) a small number of individual samples within a particular setting. The advantage of the logical or rational item selection method is that it minimizes the possibilities that certain important risk factors will be excluded because they happened not to be significant in a given sample or were not even measured in the first place.
The development of Version 3 of the HCR-20 was based on an updated systematic literature review (including a 300-page annotated bibliography) since the publication of Version 2 in 1997. This was done to determine those risk factors that should be dropped, added, or modified.
For the Version 3 development, additional steps were taken. There was an extensive consultation process with experienced agencies and users of the HCR-20 Version 2 to seek their opinions on potential limitations of Version 2 and areas for improvement. Based on this feedback, and the experiences of the HCR-20 authors themselves, they sought to clarify ambiguities or inconsistencies; to improve the applied utility of the instrument by emphasizing individual-focused administration procedures focusing on formulation, scenario planning, and risk management; and to ensure legal acceptability by avoiding prima facie objectionable risk factors (such as the personal demographic attributes of gender, age, or ethnicity). Once a draft of Version 3 was developed, it was presented to an additional 30 experienced users for systematic feedback, after focus group exercises. The draft was also beta tested in several different countries within agencies that had been using Version 2 for at least 10 years. In total, the beta testing included roughly 25 additional clinicians and the same number of evaluees. Beta testers provided extensive feedback, and revisions to this draft were undertaken as a result.
Finally, considerable effort was devoted to the empirical evaluation of the reliability and predictive validity of the penultimate draft of the HCR20 Version 3. This was done to ensure that the new HCR-20 would have psychometric properties at least as strong as its predecessor. Data were collected across eight sites and seven countries for the purpose of empirical evaluation of Version 3: Canada, England, Germany, the Netherlands, Norway, Sweden, and the United States. There were 782 participants (offenders and patients) included in this research, across correctional, forensic psychiatric, and civil psychiatric settings. Findings supported the reliability and validity of Version 3 of the HCR-20, and hence, it was published soon after findings from these studies became available. This development research also led to a special issue on the HCR-20 Version 3 in a 2014 issue of the International Journal of Forensic Mental Health.
Using the HCR-20 in Practice
The HCR-20 is intended to structure a comprehensive violence risk assessment. Users must have appropriate experience, education, and knowledge, as described in the user manual. The administration of HCR-20 Version 3 is intended to represent the natural flow of activities and decision-making that professionals undertake not only for risk assessment but also for most assessment tasks. That is, professionals must gather information, identify important and relevant facts, make sense of that information, consider the implications of the information for the future, and recommend future steps to best handle concerns. In the violence risk assessment context, evaluators gather information, assign this information to risk factors, try to make sense of why a person has acted violently in the past, consider how and under what circumstances they might act violently in the future, and recommend management strategies to reduce risk.
Because human beings generally make better decisions when the decision-making process is structured, the general decision-making process outlined above is structured into seven steps. The purpose of providing structure to the entire risk assessment (decision-making) process is to avoid, as much as possible, reliance on heuristics, bias, consideration of task-irrelevant information, and ignoring task-relevant information. The seven administration steps of the HCR-20 Version 3 are as follows:
Step 1: Case Information
This step involves gathering and documenting enough relevant information to inform a comprehensive risk assessment. The HCR-20 Version 3 authors recommend that if at all possible the minimum information base should consist of a thorough review of case records (consisting of correctional, police, health, employment, and school files, as available) and an interview with the person of interest. Under some circumstances, information will be limited, or unavailable, in which case the evaluator must decide whether to use the HCR-20 Version 3. Where possible and indicated, other sources of information include direct observation and interviews with collateral sources such as family members, victims, or staff members who know the person well.
Step 2: Presence of Risk Factors
Each risk factor on the HCR-20 is defined within Chapter 4, and coding notes are provided. Together the definition and coding notes are used to rate whether the risk factor is present, absent, or possibly/partially present. Some of the broader or more complicated risk factors include subitems, which are optional. In addition to the definition and coding notes, each item comes with a list of indicators, which are possible ways in which the risk factor might manifest at the idiographic level. These are examples and are not meant to be coded. In addition to rating the 20 standard risk factors, evaluators can specify case-specific risk factors, so long as they can identify case facts to justify doing so.
Step 3: Relevance of Risk Factors
One of the novel features of the HCR-20 Version 3 is rating the idiographic relevance of risk factors. The 20 standard risk factors on the HCR-20 Version 3 have support at the sample or population level. However, this does not mean that, when present, they are equally important for elevating risk of all people who have them. As such, if a risk factor is determined to be present (or possibly/partially present) at Step 2, it is treated as a hypothesis that the evaluator must explore by asking, “Is this risk factor relevant to understanding this particular person’s violence?” This step is intended to optimize the individualization of the assessment and to bridge the nomothetic and idiographic levels of analysis. It is also a mechanism for evaluators to determine what risk factors really are most important in understanding a given case.
Relevance is defined as whether a risk factor is causal at the individual level (i.e., did it drive or motivate violence, or disinhibit a person, such as violent ideation or intoxication might?), whether it impairs a person’s decision-making about violence (i.e., does it impair or destabilize careful decision-making, such as intellectual impairment or executive dysfunction might?), or whether it is otherwise critical to manage (i.e., does it allow other risk factors to persist or perpetuate, such as poor treatment plans or lack of social support might?). This process helps to answer the question why a person did what he or she did and to make sense of the case.
Step 4: Risk Formulation
Risk formulation furthers the process of making sense of a case. Like formulation in other contexts (i.e., therapy), risk formulation involves the application or theory or conceptual models to integrate and condense case information into a set of smaller conceptual units that explain why, in this context, a person has been violent. As such, rather than having, say, 15 risk factors to deal with, the evaluator may now have three interconnected themes that explain a person’s violent behavior. The purpose is to develop an individual theory of violence, so that we may understand why a person might be violent in the future and what we can best do to mitigate risk.
Step 5: Risk Scenarios
This step turns to the future and asks evaluators to specify plausible concerns for the future with respect to a person’s violence. If this person will be violent again, what might the person do? What might the warning signs be? Scenario planning has been used for well over a century in other disciplines, such as military planning, and more recently for emergency preparedness, engineering, business, law enforcement, and public health. It is ideal for planning in the face of uncertainty, which aptly describes the risk assessment and management context. Typically, a small number of scenarios is specified (say, a repeat, twist, and worst-case scenario), a process that leads logically to the question, “What are we going to do to identify and manage risk under these plausible scenarios?”
Step 6: Management Strategies
This step flows naturally from the previous steps and should address the most important risk factors (present and relevant—Steps 2 and 3), why a person acts violently (Step 4), and what the person might do in the future (Step 5). Management strategies are specified under four categories: monitoring (how often, by whom, where?), supervision (how restrictive should conditions be?), treatment (what ameliorative or skills-based programming makes sense?), and victim safety planning (what can be done to help potential victims enhance their safety?). Not only should strategies be specified but so too should their details (if anger management is recommended, what specific program should it be?) and logistics (can it be ensured that the timing of this program does not conflict with other programming, or with employment?). Risk management intensity typically should be commensurate with risk level.
Step 7: Conclusory Opinions
The final step is simply for the evaluator to communicate his or her summary opinions about level of risk for which several such judgments are called. A judgment of future risk, or case prioritization, is a statement that reflects the evaluator’s opinion about the level of risk (how likely is it, based on the preceding steps, that this person will commit some form of violence in the next 6–12 months?) and concomitant degree of effort required to mitigate risk. A simple judgment of low, moderate, or high is provided. Evaluators are also asked, using the same logic, to make a risk judgment for serious (life-threatening) and imminent violence (in the coming hours, days, or weeks). These summary risk ratings have been studied extensively in research.
The Nature of Research Support for the HCR-20
The HCR-20 has been subjected to approximately 350 empirical evaluations across more than two dozen countries, the vast majority of which have been conducted by independent researchers. Most research is on Version 2, given that Version 3 was not published until 2013. Meta-analyses indicate that the HCR-20 performs at least as well as other risk assessment instruments, typically yielding moderate-to-large effect sizes. The judgments of low, moderate, or high risk have routinely been found to add incremental predictive validity to summing up the risk factors or to other risk assessment measures including actuarial ones. There has been evidence in a number of studies for dynamic predictive validity (i.e., increases in risk factors leading to increases in subsequent violence and decreases in risk factors leading to decreases in subsequent violence). Research has shown comparable predictive validity for men and women, and across countries, and settings of
References:
- Douglas, K. S., Hart, S. D., Groscup, J. L., & Litwack, T. R. (2014). Assessing violence risk. In I. Weiner & R. K. Otto (Eds.), The handbook of forensic psychology (4th ed, pp. 385–441). Hoboken, NJ: Wiley.
- Douglas, K. S., Hart, S. D., Webster, C. D., & Belfrage, H. (2013). HCR-20 (Version 3): Assessing risk for violence, user guide. Burnaby, Canada: Mental Health, Law, and Policy Institute, Simon Fraser University.
- Douglas, K. S., Hart, S. D., Webster, C. D., Belfrage, H., Guy, L. S., & Wilson, C. (2014). Historical-Clinical-Risk Management-20, Version 3 (HCR-20 V3): Development and overview. International Journal of Forensic Mental Health, 13, 93–108. doi:10.1002/ bsl.2134
- Hart, S. D., Sturmey, P., Logan, C., & McMurran, M. (2011). Forensic case formulation. International Journal of Forensic Mental Health, 10, 118–126. doi:10.1080/14999013.2011.577137
- Heilbrun, K., Douglas, K. S., & Yasuhara, K. (2009). Violence risk assessment: Core controversies. In J. L. Skeem, K. S. Douglas, & S. O. Lilienfeld (Eds.), Psychological science in the courtroom: Consensus and controversy (pp. 333–357). New York, NY: Guilford.
- Litwack, T. R., Zapf, P. A., Groscup, J. L., & Hart, S. D. (2006). Violence risk assessment: Research, legal, and clinical considerations. In I. B. Weiner & A. K. Hess (Eds.), The handbook of forensic psychology (3rd ed., pp. 487–533). Hoboken, NJ: Wiley.
- Singh, J. P., Grann, M., & Fazel, S. (2011). A comparative study of violence risk assessment tools: A systematic review and metaregression analysis of 68 studies involving 25,980 participants. Clinical Psychology Review, 31, 499–513. doi:10.1016/j.cpr.2010.11.009
- Yang, M., Wong, S. C. P., & Coid, J. (2010). The efficacy of violence prediction: A meta-analytic comparison of nine risk assessment tools. Psychological Bulletin, 136, 740–767. doi:10.1037/ a0020473