HCR-20: What It Measures and What the Evidence Shows
Author: Ukrainian Psychological Hub · Published: October 11, 2026 · Editorial Policy
HCR-20 is a structured professional judgment framework used by trained professionals to assess and manage the risk of future interpersonal violence in adults. The current major version, HCR-20 V3, organizes 20 areas of information into historical, clinical, and risk-management domains. Its central purpose is to connect a documented risk formulation with plausible future scenarios and practical risk-management actions. The framework does not provide a diagnosis, prove that a person committed a crime, or calculate a universally valid probability that an individual will be violent.
The evidence supports the value of structured assessment over an entirely impressionistic approach, while requiring careful interpretation of accuracy, population limits, and intended use. HCR-20 V2 has an appreciable validation literature; research on V3 is growing, but the two versions are not interchangeable. A major 2023 systematic review of violence risk instruments reported a pooled area under the curve (AUC) of 0.69 for V2 in the relevant independent validation studies, while V3 did not qualify for that particular meta-analysis. A 2026 systematic review of longitudinal HCR-20 studies found small changes in dynamic ratings and little evidence that improvements in those ratings independently predict less subsequent recidivism.
What does HCR-20 stand for?
HCR means Historical, Clinical, and Risk Management. The number 20 refers to the 20 risk-factor areas considered under the instrument. The historical domain contains ten factors, the clinical domain five, and the risk-management domain five. The version used in contemporary specialist practice is commonly called HCR-20 V3 or HCR-20 Version 3. The developers' overview explains that V3 was released in 2013 as a substantive revision of V2, which had been published in 1997.
The HCR-20 is best understood as a professional assessment framework with a structured sequence of inquiry and decision-making. Its developers describe it as a set of guidelines for violence-risk assessment and management, using the structured professional judgment (SPJ) model. This model is intended to discipline judgment through systematic evidence collection, explicit reasoning, consideration of future circumstances, and revision of management plans. See the official HCR-20 V3 description maintained by the developers.
What exactly does HCR-20 assess?
The outcome of interest is prospective interpersonal violence. Depending on the assessment context, this may include physical assaults or other acts covered by the governing operational definition of violence. The assessor must make the outcome, timeframe, setting, severity, and exposure opportunities explicit. A risk judgment about aggression within a secure ward over the next several months addresses a different question from the likelihood of a recorded violent offense after community discharge over several years.
The HCR-20 examines features of a person's documented history, current presentation, and anticipated circumstances that research or professional theory associates with violence-related outcomes. Such features may be risk markers, potentially modifiable conditions, or information useful for building intervention plans. A measured association with violence across a sample does not establish that a particular feature caused a particular act, and the presence of a factor does not establish that violence will occur.
Historical factors: the H domain
The ten historical areas address past circumstances and patterns relevant to violence assessment. Representative topics include the history of violence, previous behavioral or relational instability, relevant clinical or substance-use history, and responses to past supervision or treatment. The assessor examines accurate records, chronology, source quality, and the meaning of historical information for the present referral question.
Historical information often contributes predictive information because past behavior can be associated with subsequent behavior in comparable circumstances. Yet it also requires contextual interpretation. An allegation is not an adjudicated fact; an administrative record may undercount events; a criminal record is influenced by enforcement and detection; and conduct that occurred under very different life conditions may have different implications today. Historical factors generally remain recorded even when someone achieves substantial improvement.
Clinical factors: the C domain
The five clinical areas focus on more current, potentially changeable features. They concern the person's recent psychological functioning, attitudes, behavioral control, symptoms where relevant, and responses to intervention. A trained evaluator considers whether a concern is currently present, whether it plausibly matters to violence risk in this case, what evidence supports that conclusion, and whether the observed pattern has changed.
Clinical symptoms deserve especially careful interpretation. A psychiatric diagnosis is a distinct clinical construct and does not imply violence, criminality, or a particular level of future risk. Substance use, distress, suspiciousness, poor impulse control, or other features matter only to the extent that reliable case information and relevant evidence support their role in the specific risk formulation. A concise score cannot substitute for this explanation.
Risk-management factors: the R domain
The five risk-management areas concern anticipated circumstances and practical barriers or supports. These can include the realism of proposed plans, professional services, future environmental exposure, access to stability and support, and anticipated responses to stress. The relevant question is what could happen under a specified living, treatment, supervision, or institutional arrangement, including what might reduce or increase the likelihood or severity of violence.
This domain is inherently conditional: a person may face different circumstances in a secure hospital, supported housing, an unsupervised community setting, or a different correctional placement. A sound assessment specifies the scenario to which a rating pertains. If housing, access to services, supervision, or other circumstances change, the formulation and management plan may need to change as well.
What is the difference between HCR-20 V2 and V3?
Version 3 preserves the three-domain structure while expanding attention to the relevance of individual factors, explicit case formulation, scenario planning, and actionable risk management. V3 also revised item content and assessment procedures. Crucially, a large body of studies using V2 cannot simply be represented as a large body of independent validation studies of V3. The development article by Douglas and colleagues describes the methodological rationale for the revision.
V3 does not require the Psychopathy Checklist–Revised to be administered as a mandatory part of the procedure. Nor should users assume that numeric sums, cut points, or published accuracy estimates derived from V2 are official V3 thresholds. A retrospective Australian follow-up study compared the versions in one forensic psychiatric sample and found V3 AUC values of approximately 0.70–0.77 for violent recidivism; the results did not establish a statistically significant advantage over V2. That study is informative, but its sample size, retrospective design, and setting limit generalization.
How a professional HCR-20 V3 assessment works
A qualified assessor begins by defining the referral question: for whom is the assessment intended, what kind of violence is at issue, where might it occur, over what time horizon, and which decision requires support? The assessor then reviews records and, where feasible and ethically appropriate, integrates interviews, behavioral observations, collateral information, and the perspectives of the person assessed. Information gaps and conflicting sources are documented rather than converted into false precision.
Next, the evaluator considers the presence of each relevant factor and its individual significance for the person under the specified circumstances. In V3, relevance judgments are central: a factor can be present without being important to the particular scenario, while another can be particularly consequential because of its interaction with opportunities, stressors, or available supports. Factors are then synthesized into a reasoned account of how and under what circumstances violence could occur.
The final stages develop plausible future scenarios, identify warning signs, specify protective arrangements and interventions, and record structured summary judgments where appropriate. Those judgments may address future violence, serious physical harm, or near-term violence depending on the version's guidance and referral circumstances. The results should identify both uncertainty and concrete management options. The published professional user guide listing and developers' explanation describe this as a qualified professional procedure, not a public self-test.
Is HCR-20 a test, checklist, or clinical diagnosis?
It is commonly called a tool or instrument, and it includes structured ratings, but its intended application is broader than ticking boxes. A clinical diagnosis is a classification of a condition according to diagnostic criteria. A behavioral observation records something seen or reported. A factor rating summarizes specific evidence about an item. An HCR-20 formulation is a professional account of relevant violence-risk circumstances. A forensic opinion applies assessment findings to a defined psycholegal question, whereas a legal finding is made by the authorized decision-maker under applicable law.
Those products have different epistemic and legal roles. A high concern rating does not establish mens rea, criminal responsibility, competency to stand trial, guilt, dangerousness under a particular statute, or the satisfaction of a civil-commitment threshold. Courts and tribunals apply jurisdiction-specific evidence rules and legal standards. The critical review of HCR-20 V3 in legal contexts explains why the tool's clinical popularity does not settle its admissibility, fairness, or validity for a particular legal decision.
Does HCR-20 generate a numerical violence probability?
HCR-20 V3 is not designed to deliver an automatically calibrated percentage such as “this person has a 73% chance of violence.” Some studies use summed item scores, but statistical research based on numerical totals and routine SPJ decision-making are related yet distinct activities. The probability of a future outcome depends on the target event, base rate, observation period, case mix, and local conditions. A probability produced in one study cannot simply be transplanted into an individual courtroom report.
The central distinction is between discrimination and calibration. Discrimination asks whether an instrument tends to assign higher values to people who later have the studied outcome than to people who do not. Calibration asks whether predicted probabilities correspond to observed frequencies. AUC is a discrimination metric; an AUC of 0.70 does not mean 70% of people assessed will be classified correctly, and it does not mean that an individual has a 70% probability of violence.
What does the predictive-validity evidence show?
The best interpretation separates historical versions, research populations, outcomes, and research designs. In a 2023 systematic review and meta-analysis of forensic psychiatric populations, reviewers examined 50 studies and found that 49 were at high risk of bias. For violent recidivism, a pooled estimate for HCR-20 V2 was AUC 0.69, with a 95% confidence interval of 0.65–0.72, based on nine studies in the relevant external-validation synthesis. V3 did not meet the review's criteria for a pooled estimate. That is an important boundary on the evidence, rather than a finding that V3 is ineffective.
A 2020 Australian V3 study reported AUC values of about 0.70–0.77 for violent recidivism after discharge in a sample of 100 forensic psychiatric patients. A 2024 international inpatient study followed 210 male forensic inpatients with schizophrenia-spectrum diagnoses for 12 months and reported that the total-score performance was stronger for the first six months than afterward, with repeat assessment at six months improving the subsequent prediction period. Those findings pertain to selected clinical populations and outcomes; they do not justify applying the same estimates to a general community sample.
Recent work also illustrates how sharply domains and settings can differ. A 2026 single-site retrospective inpatient study reported an AUC of 0.826 for the Clinical domain and 0.748 for the total score in its specific male forensic inpatient sample, with weaker performance for other domains. These figures describe performance against one defined inpatient outcome. A locally optimized cutoff is not an externally validated universal threshold and should not be imported into another hospital or legal jurisdiction without independent evaluation.
Why one AUC is insufficient
AUC summarizes one aspect of ranking performance. It does not report how many people were incorrectly labeled high concern at a chosen decision threshold, whether probabilities are calibrated, whether an assessment changes clinical decisions for the better, or whether it reduces violence in practice. A defensible evaluation also examines sensitivity, specificity, positive and negative predictive values, confidence intervals, missing data, event prevalence, follow-up duration, and external validation.
The 2023 review found that calibration and several clinically important classification metrics were seldom reported, even when studies reported discrimination. Its authors called for attention to the quality of validation and the realities of implementation. Reliability and validity should be reported for the exact version, the intended decision, the intended population, and the outcome being evaluated.
False positives, false negatives, and base rates
A false positive occurs when an assessment classifies a person as above a selected threshold and the defined outcome is not observed within follow-up. A false negative occurs when a person is classified below that threshold and the outcome subsequently occurs. Both have serious implications, but their frequency depends on how the threshold, outcome, and observation period were defined. In supervised populations, an event may also be prevented by the management intervention itself, complicating interpretation of apparent false positives.
A simple hypothetical example illustrates base-rate sensitivity. Suppose 100 people are followed and 10 experience the outcome. With sensitivity of 80% and specificity of 80%, a binary classification would identify eight of those ten and incorrectly flag 18 of the 90 without an observed event. Only eight of 26 positive classifications would be true positives, a positive predictive value of about 31%. This is arithmetic for an imaginary instrument and population, not an HCR-20 performance estimate.
Across several instruments and studies, the 2012 BMJ systematic review reported a median positive predictive value of 41% for violent offending and emphasized that risk tools should not be sole determinants of detention or release. This was a multi-instrument synthesis rather than a V3-specific estimate. Its message remains crucial: risk labels should not be treated as precise individual forecasts.
Is HCR-20 reliable between different raters?
Interrater reliability measures agreement between evaluators applying the same procedures to the same material. The result can depend on the completeness of information, quality of training, interpretation of ambiguous factors, and the particular summary judgment. It is possible to obtain high agreement on whether an item is present while obtaining lower agreement on an individualized scenario or serious-harm judgment.
A study in an Argentine forensic psychiatric unit found moderate-to-excellent agreement across several ratings, but the independent double-rating sample included only 15 people and some summary judgment reliability results were statistically uncertain. A 2020 Australian study reported good-to-excellent reliability within its research design. These findings support qualified use and continuing measurement work; they do not imply that all trained assessors will agree in every real-world case.
Does HCR-20 detect improvement and changing risk?
This question matters because the C and R domains are potentially dynamic. Treatment, a change in symptoms, housing stability, social supports, institutional conditions, and supervision can alter the circumstances addressed in an assessment. Evaluators can reassess when clinically relevant changes occur, but a decrease in a rating is an observed measurement change that requires interpretation, not proof that future violence has become less likely.
The Campbell and colleagues 2026 systematic review identified 12 longitudinal studies. It found that Clinical and Risk Management ratings commonly decreased by small amounts over time and sometimes moved in parallel with other dynamic measures. The review found little evidence that a decline in HCR-20 change scores predicts a decline in later violent recidivism. This gap separates sensitivity to recorded change from demonstrated predictive validity of change and from evidence that assessment-led interventions prevent harm.
How often should HCR-20 be repeated?
The reassessment interval depends on the referral question, setting, relevant changes, and professional judgment. A single research study's recommended interval is not a universal rule. The 2024 international inpatient investigation found evidence supporting reassessment around six months within its 12-month observation period and specific forensic inpatient population. In practice, a significant transition such as discharge, change in placement, altered access to support, or a major deterioration in functioning may justify an earlier review.
The NICE NG10 guidance recommends regularly reviewing violence-risk assessments and plans in relevant clinical contexts, with review frequency informed by risk and circumstances. It also addresses acute and short-term tools suited to inpatient monitoring. HCR-20 should not be assumed to replace more immediate instruments designed for monitoring short-term escalation over hours or days.
Does using HCR-20 actually reduce violence?
Predictive validity is not the same as demonstrated preventive benefit. An instrument may classify groups at different observed event rates without improving decisions, treatment engagement, discharge planning, safety, or outcomes when implemented. To establish preventive benefit, research needs to examine whether using the framework produces better decisions or fewer harmful events compared with a credible alternative under realistic conditions.
A 2021 review of 206 HCR-20 papers reported that studies overwhelmingly focused on scores and prediction, while the full SPJ sequence, especially scenario planning and management, had rarely been evaluated. The authors identified only one study testing whether HCR-20 use affected violence rates. A 2026 qualitative study of real-world practice examined how clinicians integrate the instrument into actual management decisions. Such implementation evidence is valuable but cannot substitute for controlled outcome evidence showing that the assessment itself prevents violence.
Can HCR-20 be used with women?
Applications to women require direct examination of validity and risk formulation, not an assumption that findings from male samples transfer automatically. A Dutch multicenter study of 78 women reported limitations in the prediction of violent recidivism, while a 2024 German study of 452 female forensic inpatients with substance use disorders found useful discrimination for HCR-20 V3 and no incremental predictive benefit from the Female Additional Manual in that sample.
A separate 2024 study of 525 discharged female forensic psychiatric patients found moderate-to-good performance among several instruments but cautioned that sensitivity was low and that no tool should be the sole basis for discharge decisions. The coexistence of these results is informative: conclusions depend on sample composition, event rate, outcomes and follow-up. A gender-responsive formulation must attend to relevant circumstances without turning stereotypes into individual predictions.
How do culture, ethnicity, and fairness affect its use?
Fairness requires more than reporting an overall AUC for a mixed sample. If validation studies underrepresent a group, the accuracy of the tool's classifications and meanings of its items in that group can remain uncertain. Differences in language, institutional exposure, access to treatment, policing, socioeconomic circumstances, and documented history may influence both what information appears in files and how it is interpreted. These are empirical and ethical issues requiring local examination.
For example, records of criminal justice contact can reflect both behavior and surveillance practices. The evaluator must distinguish verified acts from allegations and avoid converting structural inequality into presumed personal dangerousness. Subgroup performance should be checked for discrimination, calibration where probabilities are used, false-positive and false-negative rates, and measurement equivalence. Small subgroups may produce unstable estimates, and apparent parity on one metric does not ensure parity on all others.
The American Psychological Association specialty guidelines emphasize appropriate assessment methods, demonstrated reliability and validity for the assessed population, and attention to the examinee's language and circumstances. NICE NG10 likewise instructs practitioners to avoid assumptions based on culture, religion, or ethnicity, and to consider whether a perception of risk can be verified.
What are its population and setting limits?
HCR-20 V3 is intended for adult assessment. It has been applied in forensic, correctional, civil psychiatric, and community-related settings, but each setting changes both the observed outcome and the available opportunities for violence. A secure ward's reports of aggression, police-recorded violent recidivism, self-reported violence, and allegations reported to probation are not interchangeable outcomes. An assessment for adolescents requires a developmentally appropriate approach rather than casual extension of an adult framework.
General community screening is a different task from an individualized assessment following documented concern in specialist services. The instrument is not a casual workplace personality questionnaire, a substitute for emergency threat triage, or a validated profile of people who commit particular types of crimes. Transfer to a new country, legal system, language, age group, or clinical subgroup requires attention to the actual supporting evidence.
How does HCR-20 compare with actuarial tools?
An actuarial instrument generally combines predefined variables using a specified statistical scoring rule. In SPJ, the structure specifies what to examine, while the trained assessor integrates evidence to formulate plausible scenarios and risk-management options. Both approaches may contain structured items and numerical research measures; they differ in the role assigned to professional judgment and the way decisions are formulated.
The decisive comparison is purpose specific. If the task is estimating group-level event frequency in a validated population, a well-calibrated model may supply useful information. If the task is explaining the contexts in which a specific individual might become violent and which supports could mitigate that risk, a transparent formulation may be more directly actionable. Neither label guarantees superior prediction, fair outcomes, or practical benefit. The evidence review of HCR-20 study designs shows why research on sum-score accuracy should not be counted as validation of every SPJ management step.
HCR-20 versus psychopathy measures, acute checklists, and protective-factor tools
Psychopathy instruments assess a specific constellation of personality and behavioral characteristics; they do not establish a crime, and they are not synonymous with violence-risk assessment. HCR-20 may consider relevant history and personality-related information where documented, but it addresses a prospective management question. A person can receive an HCR-20 evaluation without meeting criteria for any particular personality disorder.
Short-term tools such as the Brøset Violence Checklist or DASA-IV may be appropriate for acute inpatient monitoring under specific circumstances, as reflected in NICE NG10. Protective-factor approaches can supplement attention to stabilizing resources, strengths, relationships, and supports. Their measurement claims require independent validation; adding more forms does not automatically improve decisions. HCR-20 V3 itself anticipates risk-management resources, but it is not simply an inventory of vulnerabilities.
How should protective factors enter an HCR-20 formulation?
Protective circumstances can influence the scenarios an assessor considers and the interventions proposed. Relevant examples might include reliable access to treatment, practical support, a stable living plan, improved response to stress, and structured supervision when justified. Their significance depends on the person's actual circumstances and on whether they are accessible and durable, rather than on the mere assertion that support exists.
A strong plan describes who is responsible for each action, what early warning signs would trigger review, what alternatives are available when a service fails, and how changes will be monitored. Protective conditions may reduce exposure to risk-relevant situations without proving that an underlying historical factor has disappeared. The effectiveness of any specific arrangement must be evaluated rather than presumed.
How is HCR-20 used in courts and release decisions?
Clinicians may contribute HCR-20-based opinions to proceedings involving release planning, institutional management, conditional supervision, or other questions in which future violence is relevant. The permissible use depends on the applicable jurisdiction, the legal question, rules of expert evidence, and the quality and relevance of the research cited. A court retains responsibility for applying the law and evaluating the whole evidentiary record.
The 2022 critical reflection on HCR-20 V3 in legal contexts reviews strengths, admissibility concerns, and potential biases across several jurisdictions. Practitioners should explain whether published validation results match the person and decision at hand, which outcomes were studied, what was not tested, and whether the assessment adds reliable information beyond facts already in evidence. A risk opinion should be expressed as an opinion with reasons and limitations, not as a judicial finding.
Criminal responsibility, competency, and dangerousness
Criminal responsibility concerns the legal and mental-state standards applicable to conduct at a particular time. Competency to stand trial concerns legally specified present capacities for participation in proceedings. Risk of future violence is a separate prospective question. A high HCR-20 concern rating cannot determine whether a defendant understood wrongfulness, possessed required intent, or can currently assist counsel. Those decisions require distinct evidence and jurisdiction-specific criteria.
Dangerousness may be a term used in law, but its meaning varies across statutory schemes. In some contexts, risk information can be relevant to a legal inquiry without satisfying the inquiry on its own. The APA specialty guidelines require forensic evaluators to make the intended use and limitations of methods clear, and to stay within their professional competence.
What should an HCR-20 report contain?
An intelligible report states the referral source, questions, dates, version used, qualifications of the evaluator, information reviewed, attempts to obtain missing data, limitations, operational definition of violence, and relevant time horizons. It distinguishes directly observed behavior, verified records, self-report, collateral reports, and disputed allegations. Ratings should be tied to evidence and relevance rather than asserted as self-evident facts.
The report should then describe plausible scenarios, circumstances expected to aggravate or reduce risk, management recommendations, contingencies if circumstances change, and the rationale for any summary judgments. It should distinguish a rating based on records from an independent clinical finding and from a legal decision. It should communicate uncertainty in terms the recipient can understand, without a false aura of mathematical certainty.
The report's audience matters. A clinical team may need a detailed intervention plan; a judge may need clear explanations of methodology and the limits of transfer from research samples; a person assessed should receive an appropriately explained account of the findings and available review procedures. The evaluator must also respect confidentiality, disclosure obligations, informed-consent requirements where applicable, and local professional standards.
What are the most common mistakes in HCR-20 interpretation?
One mistake is treating the 20 items as a mechanical diagnostic checklist and presenting a total as an official probability. Another is citing V2 evidence as though it validates every V3 process, using AUC as if it measured calibration, or importing a study-specific threshold to another population. A third is calling a historical risk marker a proven cause of offending. These errors confuse evidence about groups with conclusions about individuals.
Additional failures include giving disputed allegations the status of verified behavior, treating psychiatric diagnosis as synonymous with dangerousness, neglecting protective conditions, omitting the future setting and timeframe, failing to document uncertainty, and applying the assessment to a legal question outside its design. Even a carefully completed form can provide little benefit if its findings are never connected to a realistic and accountable risk-management plan.
Can someone complete HCR-20 on their own?
The framework is a specialist professional instrument requiring appropriate training, access to detailed information, and responsible interpretation. Publicly available explanations can help a person understand why historical information, current functioning, future circumstances, and management planning are considered. They do not equip an untrained person to generate a defensible formal HCR-20 assessment, nor should a family member, journalist, or online commentator use an informal checklist to label another person dangerous.
When an HCR-20 report affects liberty, services, employment, supervision, or clinical care, the relevant questions include who assessed the person, which version was used, whether the referral and outcome were defined, whether records and alternative interpretations were considered, and whether the assessment has appropriate validation for the decision. An independent qualified professional may be able to review methodological questions, subject to applicable procedures.
What the evidence establishes—and what it leaves open
The strongest supported conclusion is that HCR-20 provides a systematic language and workflow for examining violence-related information, and that several versions and applications have shown useful but imperfect group-level discrimination in selected forensic populations. Independent validation, particularly for V3 and its full scenario-to-management sequence, remains less complete than is sometimes implied by the size of the broader HCR-20 literature. Performance varies across populations, outcomes, periods, and methods.
The second conclusion concerns decision quality. HCR-20 is most responsibly used as one part of careful, revisable, transparent professional reasoning. Its value in a particular case depends on the relevance and accuracy of the evidence, what the decision actually requires, available management interventions, and fair interpretation. A structured assessment can make reasoning more explicit, but its score or category alone cannot decide an individual's future or resolve a legal standard.
Frequently asked questions
What is HCR-20 used for in forensic psychology?
HCR-20 V3 is used to organize information relevant to future interpersonal violence and to help trained professionals formulate scenarios and management plans, especially in forensic or related clinical settings. It supports assessment and planning; it is not evidence that a crime occurred.
What is a high HCR-20 score?
In published studies, researchers sometimes examine summed ratings. In V3 practice, however, item presence, individual relevance, risk formulation, scenarios, and structured summary judgments all matter. A numerical total alone does not translate into a universal individual probability or legal classification.
What are the 20 factors divided into?
They are organized into ten historical, five clinical, and five risk-management areas. The domains direct attention to past information, current clinical or behavioral concerns, and anticipated future circumstances. The official guide controls specific item definitions and rating procedures.
How accurate is HCR-20 at predicting violence?
Accuracy differs by version, population, timeframe, and outcome. The 2023 systematic review pooled an AUC of 0.69 for HCR-20 V2 in its relevant external-validation synthesis; it did not obtain a qualifying pooled estimate for V3. Individual V3 studies report varying results, and none provides a universal person's probability of violence.
Is HCR-20 evidence-based?
It has a substantial research background and a growing V3 literature, including independent cohort studies and systematic reviews. The evidence is stronger for some predictive or reliability questions than for the claim that complete HCR-20-led management programs themselves reduce violence. The quality and applicability of evidence should be evaluated for the specific use.
Does HCR-20 diagnose psychopathy or mental illness?
No. It is a violence-risk assessment framework. Psychiatric and personality-related information may be considered when relevant and verified, but diagnostic judgments follow separate clinical standards and are not equivalent to violence predictions.
Can HCR-20 be used for teenagers?
HCR-20 V3 is intended for adult assessment. Adolescent violence and developmental risk require age-appropriate methods and evidence. The suitability of a specific juvenile instrument is a separate assessment question.
Is HCR-20 suitable for US courts?
It may be introduced or considered in some US legal contexts, depending on the issue and applicable rules. Its admissibility and weight are case- and jurisdiction-specific. A professional must show relevance, reliability, appropriate use, and transparent limitations; an HCR-20 rating does not decide the legal issue.
Can a person change their HCR-20 rating?
Dynamic clinical and risk-management ratings can change as circumstances and functioning change. A 2026 systematic review found that changes in ratings are often small and that evidence connecting their decline to lower subsequent violent recidivism remains limited. An updated assessment is needed when relevant circumstances materially change.
Does an HCR-20 high-risk judgment mean someone will commit violence?
No. Such a judgment is a structured professional opinion concerning risk under specified circumstances. It is inherently uncertain. It cannot establish guilt, justify an automatic restriction, or substitute for an individualized legal decision under applicable standards.
Related articles
For the general field and the distinction between risk opinion and legal evidence, see Forensic Psychology: What It Is and What Forensic Psychologists Do and Forensic Psychological Assessment: Purpose, Methods, and Limits.
For broader concepts that have their own canonical search intents, continue with Violence Risk Assessment: What It Can Predict and Where It Fails, Static and Dynamic Risk Factors: What They Mean in Forensic Psychology, and Risk Assessment and Recidivism: Prediction, Calibration, and False Positives. The clinical intersection is treated separately in Schizophrenia and Violence: What Research Really Shows.
Related Articles
References
American Psychological Association. (2013). Specialty guidelines for forensic psychology. American Psychologist, 68(1), 7–19. Professional guideline.
Brookstein, D. M., et al. (2020/2021). Predictive validity of the HCR-20 V3 in a sample of Australian forensic psychiatric patients. Psychiatry, Psychology and Law, 28(3), 325–342. DOI.
Campbell, D., Cohen, M., Macpherson, G., & O'Rourke, S. (2026). The longitudinal properties of the HCR-20: A systematic review. Aggression and Violent Behavior, 86, Article 102111. DOI.
Challinor, A., et al. (2021). The empirical evidence base for the use of the HCR-20: A narrative review of study designs and transferability of results to clinical practice. International Journal of Law and Psychiatry, 78. PubMed.
Challinor, A., et al. (2026). Violence risk assessment and management with the HCR-20 in real-world clinical practice: A qualitative interview study. The Journal of Forensic Psychiatry & Psychology, 37(1), 75–97. DOI.
Douglas Research Lab, Simon Fraser University. (n.d.). HCR-20 (Version 3). Developers' instrument description.
Douglas, K. S., Hart, S. D., Webster, C. D., Belfrage, H., Guy, L. S., & Wilson, C. M. (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.
Dudeck, M., et al. (2024). Evaluation of whether commonly used risk assessment tools are applicable to women in forensic psychiatric institutions. Comprehensive Psychiatry, 135, Article 152528. DOI.
Fazel, S., Singh, J. P., Doll, H., & Grann, M. (2012). Use of risk assessment instruments to predict violence and antisocial behaviour in 73 samples involving 24,827 people: Systematic review and meta-analysis. BMJ, 345, e4692. DOI.
Minoudis, P., et al. (2025, online; 2026 issue). Evaluation of a modified procedure for rating the presence of HCR-20 V3 risk factors. Legal and Criminological Psychology, 31(1), 64–84. DOI.
National Institute for Health and Care Excellence. (2015). Violence and aggression: Short-term management in mental health, health and community settings (NG10). Official guidance.
Ogonah, M. G., et al. (2023). Violence risk assessment instruments in forensic psychiatric populations: A systematic review and meta-analysis. The Lancet Psychiatry. PubMed.
Picchioni, M., et al. (2024). The predictive validity and temporal characteristics of the HCR-20 V3 for inpatient violence in forensic inpatient settings: An international study. Psychiatry Research, 339, Article 116079. DOI.
Wolf, V., et al. (2024). Predicting violence in female forensic inpatients with substance use disorders: The utility of a gender-responsive assessment. Frontiers in Psychiatry, 15, Article 1346815. DOI.