Australian Institute of Criminology

Skip to content

Risk assessment by mental health professionals and the prevention of future violent behaviour

Trends & issues in crime and criminal justice no. 281

Bernadette McSherry
ISBN 0 642 53847 6 ; ISSN 0817-8542
Canberra: Australian Institute of Criminology, July 2004

Foreword | The concept of 'risk assessment' is of increasing relevance when dealing with the prevention of criminal behaviour. This risk can be examined from a number of viewpoints. In an earlier Trends & issues in crime and criminal justice paper (no. 280) the risk of reoffending by Indigenous male violent and sexual offenders was examined. This paper looks at the issue from a mental health perspective. It discusses the implications for sentencing decisions and examines the current legal and ethical background to risk assessment for the purpose of preventing future serious injury to others.

Toni Makkai
Acting Director


Risk assessment may be legally relevant to mental health professionals in two ways. First, in the forensic context, psychiatrists and psychologists may be called upon to assess the risk that their patient or client may be violent in the future. Risk assessment may be relevant in civil and criminal law as well as in professional conduct contexts. In the civil law field, risk assessment is necessary in relation to the involuntary commitment of those diagnosed with a mental illness or intellectual disability, detention to prevent the spread of infectious diseases, assessing the risk of child abuse in family law matters, child protection proceedings and workplace occupational health and safety. In the criminal law field, mental health professionals may also be asked to write reports in relation to the risk of an accused person reoffending, for the purposes of bail applications, sentencing and preventive detention, the disposition of offenders with mental disorders, and parole. Second, one of the most difficult questions for psychiatrists and psychologists concerns knowing when they should disclose a patient's confidential communication on the basis that the patient may be at risk of harming others. If they breach confidentiality, they may leave themselves open to a legal claim for negligence, breach of contract or breach of confidence by the patient, as well as professional disciplinary action. If the mental health professional does not breach confidentiality, there may be a risk of the patient committing a serious offence, engaging in self harm, or putting other people's lives and well-being at risk.

The development of concepts of risk assessment and risk management

Foucault (1988: 128) has pointed out that '[l]egal justice today has at least as much to do with criminals as with crimes'. In his view, since the 19th century, there has been a shift in focus from punishing criminal conduct to regulating the danger potentially inherent in the individual. Certainly, the assessment of 'risk' has become of such significance in recent years, that is has been viewed as a core organising concept of the Western world (Gray, Laing & Noaks 2002). Risk assessment and risk management now occupy a prominent position in virtually all forms of mental health practice (Mullen 2000).

However, assessing the risk of future violence is a notoriously difficult task (Mullen 2000). During the early 1980s, research suggested that mental health professionals tended to overpredict violence (McAuley 1993: 7) and one study concluded that it was rare for psychiatrists to predict future violence with a better that 33 per cent accuracy (Monahan 1981). During this time, the emphasis was on making clinical assessments of 'dangerousness' that did not provide a medical diagnosis, but involved 'issues of legal judgment and definition, as well as issues of social policy' (Steadman 2000: 266).

Between the mid 1980s until the mid to late 1990s, the focus shifted from assessing dangerousness to a focus on statistical or actuarial risk prediction. This shift to risk assessment and risk management has seen the rise of 'scientific' literature examining a range of risk factors that have a statistical association to a future event. The main limitation of this shift in approach is that actuarial judgments may ignore individual needs and individual differences, whilst focusing too much on historical variables. The main benefit of the rise of actuarial instruments to assess risk is that it has altered the focus from concepts of dangerousness to probabilistic thinking and ideas of graduated intervention as opposed to 'yes/no', 'in/out' dichotomies associated with the concept of dangerousness (Steadman 2000: 266).

Currently, risk assessment involves the consideration of risk factors, harm and likelihood. It combines both clinical and actuarial approaches to form what has been termed 'structural clinical judgment' (Heilbrun, Ogloff & Picarello 1999). Instruments such as the Psychopathy Checklist Revised (PCL-R) (Hare 1991), the Violence Risk Appraisal Guide (VRAG) (Quinsey et al. 1998) and the Historical/Clinical/Risk-20 (HCR-20) (Webster et al. 1997) focus on variables that are said to have been ascertained by actuarial studies. The Macarthur Study of Mental Disorder and Violence has also led to the development of a classification tree model referred to as an 'iterative' classification tree (ICT) (Monahan et al. 2001).

A number of risk predictor variables have been identified. The following provides a brief overview of them.

Past violence

Currently, it appears that the best predictor of future violence is past violence (Tardiff 1992). In terms of risk management, mental health professionals need to assess the patient/client's current clinical state rather than simply relying on past history. Very few people are going to be violent at all times and in all situations.

Pre-existing vulnerabilities

Mullen (2000) includes in this category: being male, anti-social traits, suspiciousness, childhood marred by disorganisation and/or abuse, youth, impulsivity and irritability. Youth is generally associated with the risk of violence (Swanson et al. 1990) and men commit the majority of violent crimes across different cultures (Marzuk 1996). However, in the mentally ill, the difference between men and women with regard to violence is far less marked (Steadman et al. 1994). A childhood history of abuse and neglect or harsh and inconsistent parenting has also figured prominently in the literature as a risk factor (Faulk 1994). Early signs of persistent antisocial traits, difficulties in peer relationships and hostility toward authority figures are also key risk factors for later risk of violence (Melton et al. 1997).

Social and interpersonal factors

Mullen (2000) includes in this category: poor social networks, lack of education and work skills, itinerant lifestyle, poverty and homelessness. Swanson et al. (1990) found that those who were violent were more likely to come from low socio-economic groups. Similarly, Stueve and Link (1997) suggest that the link between mental illness and violence was stronger amongst those with less education.

Mental illness

Mullen (1997: 169) states that the mental illness most consistently associated with the increased risk of violent behaviour is schizophrenia. However, among homicide offenders, the incidence of depression at the time of the offence is relatively high. The general literature appears to suggest that mental illness, of itself, does not reliably predict violence (Mullen 1996). However, some symptoms of mental illness are related to risk. Mullen (2000) includes in this category: active symptoms, poor compliance with medication and treatment, poor engagement with treatment services, treatment resistance and lack of insight into the illness.

Substance abuse

The presence of substance abuse is a strong risk factor for violence (McCord 2001). However, while many studies recognise a link between serious criminality and alcoholism, there is less evidence that alcohol is a direct contributing factor to violence (McCord 2001). Bean (2001) suggests that setting may be more important than the pharmacology of the substance used. The co-existence of substance abuse with mental illness appears to significantly increase the risks of violent behaviour (Steadman et al. 1998).

State of mind

Mullen (2000) refers to the presence of anger or fear, delusions that evoke fear or provoke indignation or produce jealousy, clouding of consciousness or confusion, ideas of influence, and command hallucinations.

Situational triggers

Mullen (2000) points out that actuarial models often ignore situational triggers. They include loss, demands and expectations, confrontations, ready availability of weapons, and physical illness.

Personality constructs

Personality constructs are perhaps the most controversial of predictor variables for violence. Psychopathy has been said to be the best predictor of future offending (Hart 1998). However, the terms 'psychopath' and 'antisocial personality disorder' have been criticised as being social constructs (Cavadino 1998; McCallum 2001). Ellard (1996: 62) points out: 'If you are a rather disagreeable small-time thief with a bad temper you are likely to be described as suffering from Antisocial Personality Disorder. If without any contrition you waste millions of dollars of other people's money and achieve nothing but notoriety you will be called an entrepreneur. No one reaches for the DSM-IV'.

Hare (2002: 27) states that psychopathy is 'a personality disorder defined by a cluster of interpersonal, affective, and lifestyle characteristics that results in serious, negative consequences for society. Among the most devastating features of the disorder are a callous disregard for the rights of others and a propensity for predatory behaviour and violence'.

Hare (2002) opines that most offenders diagnosed with 'antisocial personality disorder' pursuant to the criteria in the American Psychiatric Association's Diagnostic and statistical manual of mental disorders (2002 4th ed text revision) are not psychopaths. His view is that psychopaths make up only about one per cent of the general population, but as much as one quarter of the prison population.

Overall, as Prins (1996) points out, there is no ideal, or even sophisticated, approach available to the assessment of risk. It would seem that risk assessment should vary according to the characteristics of the individual, situation and potential victim involved along with the number of cumulative risk factors experienced by the patient.

The use of risk assessment in sentencing and preventive detention

While Steadman (2000: 268) acknowledges that the research literature has made the shift away from the dichotomous thinking associated with dangerousness, he concedes that 'from the judicial perspective, [is it unclear] how much change has really occurred' when it comes to making the final decision.

The rise of preventive detention has led to the law requiring assessments of 'the kind of crime one might commit in the future' (Pratt 1997: 171; Freiberg 2000). It has been argued that, in order to protect society, there will always be the need for the courts to take some account of the risk of future violent behaviour in imposing sentences (Zimring & Hawkins 1986). The common law principle of proportionality in sentencing provides that the 'type and extent of punishment should be commensurate to the gravity of the harm and the degree of responsibility of the offender' (Fox 2000: 298). The rationale for this principle is to ensure sentences remain commensurate to the seriousness of the offence even where the court takes into account the protection of society.

The High Court has consistently affirmed this principle of proportionality in sentencing. The most notable discussions about the relationship of proportionality and risk can be found in the Veen cases: Veen (no. 1) (1979) 143 CLR 458; Veen (no. 2) (1988) 164 CLR 465. The majority in Veen (no. 2) confirmed that proportionality was paramount, but stated that this did not mean that public protection was irrelevant. The majority drew a distinction between merely inflating a sentence for the purposes of preventive detention, which is not permissible and exercising the sentencing discretion having regard to the protection of society among other factors, which is permissible.

In Veen (no. 2), the majority of the High Court (at 486) noted that it is possible for Parliament to set up a scheme for indefinite detention. This is precisely what has happened in recent years with the introduction of legislative provisions that enable indefinite terms of imprisonment on the basis that the offender is a serious danger to the community. In comparison, the High Court has ruled that legislation aimed at the indefinite detention of an individual offender (Community Protection Act 1994 (NSW) which was based on the Community Protection Act 1990 (Vic.)) is unconstitutional: Kable v DPP(NSW) (1996) 189 CLR 51. Generally, in jurisdictions with provision for indefinite sentencing, a court can order such a sentence on its own initiative, or upon application of the prosecution. The legislation also provides for periodical review of the appropriateness of the sentence. Table 1 sets out an overview of such legislation.

Table 1 : Indefinite detention provisions
Jurisdiction Statutory provision
ACT No equivalent provision
NSW No equivalent provision. Para 10.8 of a report on sentencing produced by the New South Wales Law Reform Commission (1996) expressed the view that provisions providing for indefinite detention should not be introduced in New South Wales.
NT Sentencing Act 1995 - Section 65 (violent offenders convicted of a crime for which a life sentence may be imposed can be sentenced to an indefinite term of imprisonment by the Supreme Court where the Court considers the prisoner to be a serious danger to the community)

Penalties and Sentences Act 1992 - Section 163 (violent offender who presents a serious danger to the community)

Dangerous Prisoner (Sexual Offenders) Act 2003 - Section 13 (prisoners who pose a serious danger to the community if released can continue to be held under a continuing detention order for an indefinite term)

SA Criminal Law (Sentencing) Act 1988 - Part 2, Division 3 - Section 22 (habitual criminal) and Section 23 (offender incapable of controlling sexual instincts)

Criminal Code - Section 392 (dangerous offender) repealed by Sentencing Act 1997 Schedule 1 which commenced on 1 August 1998

Sentencing Act 1997 - Section 19 (dangerous offender convicted of a violent crime)

Vic. Sentencing Act 1991 - Section 18A (offender convicted of a serious offence and high probability that offender is a danger to the community)
WA Sentencing Act 1995 - Section 98 (superior court may impose indefinite imprisonment in cases where if released, the offender would pose a danger to society)

Indefinite detention legislation has created tension between principles of proportionality and questions of risk or public protection. Such legislation has been criticised on the grounds that:

  • risk is afforded too much prominence;
  • the inclusion in legislation of risk/public protection is often a political response to media and public pressure;
  • the legislation fails to define the key terms in a coherent and consistent manner; and
  • ill-defined legislative notions of 'risk' cut across notions of proportionality, resulting in conceptual confusion (Morgan, Morgan & Morgan 1998: 25-26).

The High Court in the case of McGarry v The Queen (2001) 207 CLR 121 has confirmed that indefinite detention may be legislatively sanctioned, but has signalled that there must be more evidence before the sentencing judge than a risk that the offender will reoffend before an order for indefinite detention can be made. In their joint judgment, Gleeson, Gaudron, McHugh, Gummow and Hayne stated (at 126) that section 98 of the Sentencing Act 1995 (WA) 'does not oblige a sentencing judge to make an order for indefinite imprisonment in every case in which the conditions specified in that sub-section are met'. The sentencing judge retains a discretion at all times. They went on to note (at 130) that the consequences of the commission of predicted future offences must be 'grave or serious for society as a whole or for some part of it' before the offender could be reckoned 'a danger to society'.

Kirby in a separate judgment, emphasised that imposing an indefinite sentence is a serious and extraordinary step that must be based on reports provided by those 'with psychiatric, psychological or similar qualifications'. Kirby also acknowledged the limitations experienced by judicial officers, parole officers and others in predicting dangerousness accurately and estimating what people will do in the future.

In R v Moffatt [1998] 2 VR 229 (at 255), Hayne observed that 'the fundamental proposition [is] that such powers [of indefinite detention] are to be sparingly exercised, and then only in clear cases'. This remains the situation following the High Court decision in McGarry's case.

Breaching confidentiality in the public interest

There are a number of ethical issues raised in relation to risk assessment and particularly in relation to breaching confidentiality (McSherry 2000, 2001). The Australian Psychological Society's Code of ethics (1999) permits disclosure of confidential information in circumstances where there is a 'clear risk' to others (general principles III(a)) and the Society's Guidelines on confidentiality (1999: preamble para 4) state that confidentiality is not absolute. The guidelines issued by the Royal Australian and New Zealand College of Psychiatrists (RANZCP) (1999) also permit disclosure at the discretion of the psychiatrist where a patient's intention is 'to seriously harm an identified person or group of persons' (annotation to principle 4 para 4.6).

In general, the utilitarian or consequentialist rationale for limiting disclosure rests on the presumed importance of the relationship between a health professional and patient. If confidentiality is not guaranteed there is the possibility that patients will be inhibited in their discussions and unable to receive the full benefit of the therapeutic relationship. Engelhardt (1986) for example, has argued that clients may withhold information if they know it can be disclosed to third parties. This may be particularly salient in circumstances where the client has been referred to a psychiatrist or psychologist as part of the criminal justice system. Such damage to the therapeutic relationship and hindrance to treatment may also be counter-productive to therapeutic prevention of criminal behaviour (Kottow 1986).

The law also does not consider confidentiality to be absolute. In Australia, statutory provisions exist that require a health professional to breach confidentiality in circumstances such as reporting child abuse or notifying the authorities of certain infectious diseases (McSherry 1998). However, a common law public interest exception to confidentiality has yet to fully develop.

Over 20 years ago, the Supreme Court of California held in Tarasoff v Regents of the University of California 17 Cal d 425; 131 Cal Rptr 14, 551 P 2d 334 (1976) that a duty to protect potential victims may override the confidentiality of the relationship between psychologist and patient. Tarasoff's case dealt with a situation where the patient had disclosed to a clinical psychologist working at a University student health centre that he was going to kill a woman who could be readily identified. (It has not been uniformly followed in the United States and a 1985 addition to the Californian Civil Code has substantially curtailed the scope of the duty (Cal. Civ. Code 1988, 43.92(a) and (b)).

In England, New Zealand and Canada, the courts have recognised a common law public interest exception to confidentiality (McSherry 2000, 2001). For example, Bingham of the English Court of Appeal stated in W v Egdell [1990] 1 All ER 835: 848 that 'the law treats [confidentiality] not as absolute but as liable to be overridden where there is held to be a stronger public interest in disclosure'. It remains unclear what this 'public interest' means in a legal context.

The Supreme Court of Canada case, Smith v Jones (1999) 132 CCC (3d): 225, has significantly broadened the public interest exception to enable disclosure where there is a potential risk to a class of victims. Cory in Smith v Jones set out three factors to be considered in weighing up breaching confidentiality in the interest of public safety:

First, is there a clear risk to an identifiable person or group of persons? Second, is there a risk of serious bodily harm or death? Third, is the danger imminent?

An assessment of these three factors may aid in deciding whether or not to breach confidentiality. However there are difficulties with the test set out in Smith v Jones. It may be that an appropriate legal test in what is an essentially discretionary area is impossible to formulate, but at least Smith v Jones may be a step toward greater certainty in a problematic area of professional practice.

Confidentiality of information disclosed by clients to mental health professionals has never been protected by case law in Australia from disclosure in courts. However, the courts have had discretion in this regard and would not necessarily compel disclosure: A-G (UK) v Mulholland [1963] 2 QB 477. This remains the position in relation to criminal proceedings in all states and territories in Australia. However, a statutory privilege exists in Victoria, Tasmania and the Northern Territory that protects confidential information imparted to medical practitioners in relation to civil proceedings: Evidence Act 1958 (Vic.) s 28(2); Evidence Act 1910 (Tas.) ss 87, 94, 96 and 101; Evidence Act 1939 (NT) ss 9(6) 10 12. More general legislation exists in New South Wales that may privilege confidential information disclosed in the course of any professional relationship: Evidence Amendment (Confidential Communications) Act 1997 (NSW) Div 1A.

In the late 1990s, there was an increasing practice of defence counsel in rape trials seeking access to counselling records made between alleged victims and their therapists (Bronitt & McSherry 1997). New South Wales, Victoria and South Australia enacted legislation protecting these confidential communications: Evidence Act 1995 (NSW) ss 126A-126F; Evidence Act 1929 (SA) ss 67D-67F; Evidence Act 1958 (Vic.) ss 32B-32G. Mendelson (2002) has pointed out that these provisions have been criticised by the courts as lacking precision.

At that time also, the police began to apply for search warrants to obtain psychiatric files on accused persons. The legitimacy of such conduct was considered by Cummins in Clifford v Victorian Institute of Forensic Mental Health and Anor unreported, [1999] VSC 359. On 2 March 1999, the Chief Magistrate of Victoria refused to approve a search warrant to obtain a psychiatric file from the Victorian Institute of Forensic Mental Health. The file was thought to contain an admission relating to a homicide case. Detective Senior Constable Clifford sought judicial review of that order in the Supreme Court.

Cummins found that 'public interest immunity' applied to the material and agreed with the Chief Magistrate in his opinion. In Sankey v Whitlam and Ors (1978) 142 CLR 1: 38, Gibbs stated: 'the general rule is that the court will not order the production of a document, although relevant and otherwise admissible, if it would be injurious to the public interest to disclose it'.

Cummins stressed the importance of the effective operation of the therapeutic and protective regime established under the Mental Health Act 1986 (Vic.). He found that it would be injurious to the public interest if the effective operation of that regime were not preserved and that allowing access to confidential documents would undermine that regime.

If a client has admitted to past criminal conduct, then a mental health professional may still be compelled by the courts to disclose this information: R v Lowe (1997) 2 VR 465. However, there is no legal duty to report to police a client who discloses past criminal offences except in Queensland. In that state, a medical practitioner is guilty of professional misconduct if he or she does not disclose to police any information received concerning an attempted or completed crime, or if he or she attends an injured victim or perpetrator of a criminal act and does not report the incident: Medical Act 1939 (Qld) s 35.

If the information deals with possible future harm, it would seem that public interest immunity could very well apply and mental health professionals should not grant access to such files.


Mental health professionals are called upon to assess the risk of violence presented by those with mental disorders in a range of legal areas, from decisions concerning civil detention to determinations related to bail, sentencing, probation and parole. The issue of risk is also a topic of immediate relevance to public policy and health care delivery. Risk management is an essential component of the day-to-day treatment of many patients (Lidz & Mulvey 1995).

There is some degree of consensus that well-trained clinicians should be able to predict a patient's short-term potential for violence using assessment techniques analogous to the short-term prediction of suicide risk (Tardiff 2001: 118). In particular, mental health professionals need to take into account the current literature on risk predictor variables such as past violence, pre-existing vulnerabilities, social and interpersonal factors, mental illness, substance abuse, state of mind, situational triggers and, more controversially, personality constructs.

There are many areas of the law in which courts rely on risk assessment by mental health professionals. In the criminal law field, this has been particularly important in the area of sentencing and preventive detention. There is reason to believe, however, that the courts will take a cautious approach to making decisions relating to preventive detention.

In relation to breaching confidentiality in the public interest, while there are strong ethical justifications for preserving confidentiality, it appears that both ethics and law hold that confidentiality is relative rather than absolute. The developing common law in England, New Zealand and Canada on the public interest exception to confidentiality has set out some guidelines in the forensic setting that may also be appropriate in the therapeutic context.


  • American Psychiatric Association 2002. Diagnostic and statistical manual of mental disorders. 4th ed text revision. Washington: American Psychiatric Association
  • Bean P 2001. Violence and substance abuse. In G-F Pinard and L Pagani (eds) Clinical assessment of dangerousness: empirical contributions Cambridge: Cambridge University Press: 216-237
  • Bronitt S & McSherry B 1997. The use and abuse of counseling records in sexual assault trials: reconstructing the 'rape shield'? Criminal law forum: an international journal 8(2): 259-291
  • Cavadino M 1998. Death to the psychopath Journal of forensic psychiatry 9(1): 5-8
  • Ellard J 1996. Personality disorder or the snark still at large. Australasian psychiatry 4(2): 58-64
  • Engelhardt HT 1986. Foundations of bioethics. New York: Oxford University Press
  • Faulk I 1994. Basic forensic psychiatry 2nd ed. London: Blackwell Scientific
  • Fox R 2000. Victorian criminal procedure. Melbourne: Monash Law Book Co-Operative Ltd.
  • Foucault M 1988. The dangerous individual. In LD Kritzman (ed) Politics, philosophy and culture New York: Routledge: 125-151
  • Freiberg A 2000. The politics of sentencing and imprisonment. In M Massina & R White (eds), Beyond imprisonment conference Hobart: Conference Proceedings, TasCOSS and Criminology Research Unit, University of Tasmania: 24-28
  • Gray N, Laing J & Noaks L 2002. Criminal justice, mental health and the politics of risk. London: Cavendish Publishing Ltd
  • Hare RD 1991. The Hare psychopathy checklist - revised. Toronto: Multi-Health Systems
  • --- 2002. Psychopathy and risk for recidivism and violence. In N Gray, J Laing and L Noaks (eds) Criminal justice, mental health and the politics of risk London: Cavendish Publishing Ltd: 27-47
  • Hart SD 1998. The role of psychopathy in assessing risk for violence: conceptual and methodological issues. Legal and criminological psychology 3(1): 21-137
  • Heilbrun K, Ogloff JRP & Picarello K 1999. Dangerous offender statutes in the United States and Canada: implications for risk assessment. International journal of law and psychiatry vol 22: 393-415
  • Kottow MH 1986. Medical confidentiality: An intransigent and absolute obligation. Journal of medical ethics 12(1): 17-122
  • Lidz CW & Mulvey EP 1995. Dangerousness: from legal definition to theoretical research. Law and human behavior 19(1): 41-48
  • Marzuk PM 1996. Violence, crime and mental illness: how strong a link? Archives of general psychiatry vol 53: 481-486
  • McAuley F 1993. Insanity, psychiatry and criminal responsibility. Dublin: The Round Hall Press
  • McCallum D 2001. Personality and dangerousness: genealogies of antisocial personality disorder. Cambridge: Cambridge University Press
  • McCord J 2001. Alcohol and dangerousness. In G-F Pinard and L Pagani (eds) Clinical assessment of dangerousness: empirical contributions Cambridge: Cambridge University Press: 195-215
  • McSherry B 1998. Dangerousness and public health. Alternative law journal 23(6): 276-279
  • --- 2000. Breaching confidentiality in the public interest: guidance from Canada? Monash bioethics review 19(3): 28-34
  • --- 2001. Confidentiality of psychiatric and psychological communications: the public interest exception. Psychiatry, psychology and law 8(1): 12-22
  • Melton GB, Petrila J, Poythress NG & Slobogin C 1997. Psychological evaluations for the courts 2nd ed New York: Guildford Press
  • Mendelson D 2002. Judicial responses to the protected confidential communications legislation in Australia. Journal of law and medicine 10(1): 49-60
  • Monahan J 1981. The clinical prediction of violent behaviour. Washington DC: US Department of Health and Human Services
  • Monahan J, Steadman HJ, Silver E, Appelbaum PS, Robbins PC, Mulvey EP, Roth LH, Grisso T & Banks S 2001. Rethinking risk assessment: the Macarthur study of mental disorder and violence Oxford, NY: Oxford University Press
  • Morgan F, Morgan N & Morgan I 1998. Risk assessment in sentencing and corrections. Perth: University of Western Australia
  • Mullen P 1996. The dangerousness of the mentally ill and the clinical assessment of risk. In W Brookbanks (ed) Psychiatry and the law Wellington: Brooker's: 93-116
  • --- 1997 Assessing risk of interpersonal violence in the mentally ill. Advances in psychiatric treatment vol 3: 166-173
  • --- 2000. Dangerousness, risk and the prediction of probability. In MG Gelder, JJ Lopez-Ibor & N Andreason (eds), New Oxford textbook of psychiatry London: Oxford University Press: 2066-2078
  • Pratt J 1997. Governing the dangerous. Sydney: The Federation Press
  • Prins H 1996. Risk assessment and management in criminal justice and psychiatry. Journal of forensic psychiatry vol 7: 42-62
  • Steadman HJ, Monahan J, Applebaum P, Grisso T, Mulvey E, Roth L, Clark Robbins P & Klassen D 1994. Designing a new generation of risk assessment research. In J Monahan and H Steadman (eds) Violence and mental disorder Chicago: University of Chicago Press: 297-318
  • Steadman HJ, Mulvey EP, Monahan J, Clark Robbins P, Appelbaum PS, Grisso T, Roth LH, & Silver E 1998. Violence by people discharged from acute psychiatric inpatient facilities and by others in the same neighbourhoods. Archives of general psychiatry vol 55: 393-401
  • Steadman HJ 2000. From dangerousness to risk assessment of community violence: taking stock at the turn of the century. Journal of the American academy of psychiatry & the law 28(3): 265-271
  • Swanson JW, Holzer CE, Ganju VK & Jono RT 1990. Violence and psychiatric disorder in the community: Evidence from the epidemiologic catchment area surveys. Hospital and community psychiatry 41(7): 761-770
  • Tardiff K 1992. The current state of psychiatry in the treatment of violent patients. Archives of general psychiatry vol 49: 493-499
  • --- 2001. Axis II disorders and dangerousness. In G-F Pinard and L Pagani (eds) Clinical assessment of dangerousness: empirical contributions Cambridge: Cambridge University Press: 103-120
  • Webster C, Douglas KS, Eaves K & Hart DS 1997. HRC-20: Assessing the risk of violence. Version 2 Burnaby BC: Simon Fraser University, Forensic Psychiatric Services Commission of British Columbia
  • Zimring FE & Hawkins G 1986. Dangerousness and criminal justice. Michigan law review vol 85: 481-509


This paper is taken from the report of research undertaken with the assistance of a grant from the Criminology Research Council.

CRC final report

Risk assessment by mental health professionals and the prevention of future violent behaviour

Dr Bernadette McSherry is an Associate Professor of Law at Monash University

Disclaimer: This research paper does not necessarily reflect the policy position of the Australian Government