Correctional Service Canada
Symbol of the Government of Canada

Common menu bar links


Warning This Web page has been archived on the Web.

Intervention with Sex Offenders, Toward a National Strategy

Correctional Service of Canada
Corporate Advisor
Sex Offender Programs
February 1996


Sex offenders make up an increasingly large proportion of the offender population and concomitantly their assessment, treatment, and management become high profile on both social and correctional levels.

The National Committee on Sex Offender Strategy was struck late in 1993. In order to increase its familiarity with current sex offender programs, this committee spent the greater part of 1994 visiting each Region and meeting with both federal and provincial service providers. These site visits together with the clinical expertise of the committee members, resulted in the development of the early drafts of the National Sex Offender Strategy, Standards for the Provision of Services to Sex Offenders. Consensus was reached albeit with considerable discussion.

The second phase in the development of National standards involved the National conference on Sex Offenders held from March 28 to March 30, 1995, in Toronto. This conference served two distinct purposes.

On the one hand, it provided a forum for eminent speakers to discuss issues related to the assessment, treatment, and follow-up of sex offenders. Over 425 attendees from Canada, US, and Europe took part.

The conference also provided an opportunity for approximately 100 federal corrections practitioners from across the country to review and provide feedback on the draft Standards, which was then incorporated into a subsequent draft.

The Conference met both goals. Speakers presented information on risk assessment (Monahan, Andrews, Quinsey) and examined a variety of treatment programs and issues related to service provision.. Topics included medical intervention (Hucker), intellectual/ developmental deficits (Boer), violent offenders (Serin), as well as the challenge of providing culturally relevant therapy (Ellerby) for aboriginal sex offenders. Both institutionally based (Marshall, Aubut) and community programs (Rouleau) were highlighted. Dr. Barbaree examined services for this high profile group. Outcome data related to treatment success and failure were provided (Marques, Khanna) and papers by Wormith and Gordon provided a more global perspective.

It was essential to bring treatment providers together from across Canada, for the express purpose of updating their knowledge base and discussion of the recently developed Standards. These proceedings provide the reader with an overview of the theoretical and practical issues addressed at the National conference: Intervention with sex Offenders - Towards a National Strategy.

Sharon Williams
Corporate Advisor
Sex Offender Programs
and Conference Chairperson


The conference on the National Sex Offender Strategy was held in Toronto from March 28 to 30 1995, at the Westin Harbour Castle. It consisted of two segments: a publicly attended conference and an internal CSC review of the National Strategy for Sex Offenders. More than 425 participants attended the conference over the three days.


Since January 1994, a CSC National Committee has been working on developing a National Strategy on Sex Offenders. This Working Group has a completed document and after consulting with provincial mental health and correctional agencies has developed a national consensus on issues such as assessment and treatment standards and guidelines. This initiative is supplemented by the work of the Task Force on High Risk Violent Offenders and the Action Committee on Corrections and Mental Health.

The National Committee’s proposed guidelines were presented for discussion at the National Conference on Sex Offender Strategy March 28 - 30.

Meeting in a subsequent closed session on the afternoon of March 30, a group of 100 CSC practitioners reviewed the standards which are part of the National Strategy on Sex Offender Treatment. This session was also attended by CSC Senior Management which acted as a listening panel.

There was agreement in principle on the standards, followed by further consultation with CSC staff and service providers. The Executive Committee of the Correctional Service of Canada has recently approved the Commissioner’s Directive and the Standards and Guidelines with minor modifications. An implementation plan is being drawn up, subsequent to the consultation process.


The main conference was opened by the Parliamentary Secretary to the Solicitor >General, Patrick Gagnon. Mr. Gagnon reaffirmed the Government’s commitment to the treatment of sex offenders and stated the various measures his Government was pursuing to address the issue. The Parliamentary Secretary was introduced by the Commissioner of Corrections, Mr. John Edwards.

Following the opening remarks, the Corporate Advisor, Dr. Sharon Williams addressed the assembly and introduced the keynote speaker, Professor John Monahan, School of Law, University of Virginia. The first day dealt with the prediction of violence. John Monahan gave an excellent and informative talk on violence and mental disorder. He has found that by using a multi-method approach to prediction, it is possible to predict with considerable precision who will become violent.

Vern Quinsey, Queens University, discussed more specific ways of accurately predicting sexual violence. For example, using the following information the likelihood of accurately predicting recidivism increases: the PCL(R), school problems, anti-social personality, age, failure on conditional release, separation from parents before the age of 16, history of violence and marital status. In other words, there is an orderly relationship between these predictors and the risk to re-offend violently. Dr. Quinsey has added two dynamic or changeable factors: procriminal sentiments and compliance with supervision.

Don Andrews, Carleton University, examined four major factors in predicting violence: 1) attitudes, values and beliefs, 2) antisocial associates, 3) history of antisocial behaviour and 4) levels of socialisation. He believes that dynamic factors are extremely important as a contribution to the prediction of violence.

Steve Wormith, Ontario Corrections, praised CSC's level of expertise and suggested that the development of a provincial-federal relationship was important. He stressed that appropriate treatment (cognitive behavioural) has resulted in reasonable reductions, especially when looking at groups of studies (i.e. using "Meta-analysis").

In the afternoon several National Committee Members discussed the standards for service providers, assessment guidelines, research and evaluation and accountability thus highlighting the document produced by the Working Group. The speakers provided numerous practical applications.

The second day dealt with treatment issues. Janice Marques described the California Outcome Study: she is contrasting a treated sex offender group (2 years of treatment, one year of maintenance in the community) and has a motivated but untreated control group, as well as a group who have declined treatment. The data, while finding little difference between groups at this time, still has approximately 5 years of follow-up to come. Differences were identified between offenders who knew their crime cycles and understood how to follow a relapse plan, from those who did not learn these skills.

Dr. Marques found that the dropouts from treatment are the worst risks; treatment reduces the severity of re-offending; certain types of offenders benefit more from treatment than others; (e.g. those high risk offenders who learn their crime cycle are less likely to recidivate), and she strongly suggested that CSC work on a database to be used in evaluation.

Howard Barbaree, Clarke Institute, discussed the Clément case, a high profile sex offender who recidivated violently after treatment in the Warkworth Sexual Behaviour Clinic (WSBC). He suggested documenting supervision of personnel, nationally approved standards for assessment, treatment and certification of practitioners.

Dr. Jocelyn Aubut, Institut Phillipe Pinel de Montréal, suggested that we need an eclectic approach to the complex issues involved in treating sex offenders and that we consider individual differences found in sex offenders. We need a multi-disciplinary, flexible and adaptable approach to treatment.

Lawrence Ellerby, Native Clan Counselling Services, discussed the need to integrate conventional sex offender treatment with traditional healing led by elders. Sweat lodges, fasts and other spiritual healing were described. As well, different needs should be targeted in aboriginal sex offenders (e.g. dealing with discrimination, loss and grief, family of origin, community re-integration and identity). He has found that more aboriginals drop out or are suspended due to violations, compared to non-aboriginals. When treatment is completed, both aboriginals and non-aboriginals have similar recidivism rates.

Dr. Ralph Serin, CSC NHQ, discussed the need for more systematic work with violent offenders including multi-method assessment, matching risk/need to treatment, use of measures more sophisticated than self-report and empirical studies of treatment programs. He found that sex offender assessment and treatment has advanced beyond strategies used with violent offenders.

Dr. Boer, Regional Psychiatric Centre (Pacific) CSC, discussed the low functioning offenders who often "fall between the cracks". He described the need for simplicity, short but focused intervention, behavioural contracts, self-monitoring, arousal reconditioning, adjunctive therapies and possible hormonal intervention.

Dr. Steve Hucker, Queens University, examined the medical interventions: castration, hormonal interventions (MPA, CPA, Lupron). He found that they have numerous side effects and that offenders normally do not choose to take them (out of 100 who started taking the drugs, only 11 were still in the study after 3 months). However, with proper offender motivation and medication, recidivism is reduced.

On Wednesday afternoon, the National Committee presented its draft guidelines on treatment standards. The standards were generally well received and were subject to considerable discussion in closed CSC practitioner meetings. Each considerable group presented it comments orally and in writing for further consideration.

Dr. William Marshall, Queens University examined the need to focus on intimacy within relationships, denial and minimization, victim harm and empathy, procriminal attitudes, fantasy and arousal as well as detailed relapse prevention plans. He agreed with the Draft National Standards, suggesting that risk and need be matched to appropriate treatment intensity. He also described several recidivism studies which suggested that treatment could reduce recidivism from 22% to 8%(incest) and from 43% to 13-18% for non-familial child molesters. He finished with a statement of costs and benefits which demonstrates that a reduction in recidivism would both directly impact on victim suffering and save taxpayers money by reduced police costs, court costs, victim services and the costs of otherwise incarcerating offenders. He also recommended less dependence on phallometrics and separate housing of sex offenders.

Dr. Arunima Khanna, Regional Treatment Centre (Ontario) CSC, examined factors relating to relapse. She studied 30 sex offenders who had been treated at the RTC(Ontario) and who had been reconvicted for a new sexual/homicide offence. It was found that a majority of the failures had problems with living conditions, work, lack of friends, negative/unsupportive relationships, inadequate use of leisure time, and most (80%) had been drinking/drug-abusing before the offence. Most had experienced negative emotions such as anger, stress or depression prior to the offence. The majority did not use the coping strategies which they had learned in therapy.

Dr. Joanne Rouleau, Université de Montréal, discussed the outcome data from her Montreal clinic for released sex offenders. She found that 37 or 25 % had reoffended, and that 10 of these or 25% had committed new sexual offences. Child molesters accounted for most of the recidivists. Dr. Rouleau recommended better training for parole officers, more intensive community supervision, more structure in half-way houses and more co-operation between therapists.

Dr. Art Gordon, Washington State, commented on the inappropriate use of the term "cure" which results in expectations of an all or nothing outcome. He also examined the punitive model, political ramifications and public perceptions of sexually assaultive criminal acts. Dr. Gordon recommended that we target high risk offenders with treatment aimed at criminogenic needs. He stressed that offender responsivity is an important variable and that we use professional judgement in interpreting the individual needs of offenders as well as outcome.

Dr. Gordon, discussed the need to "have fun" in therapy and to carry out treatment using firmness, fairness, caring and enthusiasm - variables which are not specified in treatment manuals. Staff must have strong positive attitudes to help shape offenders. He strongly supported the development of National Standards, and has found that Canada has taken a lead role in the development of assessment and treatment programs for sex offenders.

The summation of the Public portion of the Conference was undertaken by Mr. Andrew Graham, Senior Deputy Commissioner on behalf of the Commissioner of Corrections. Mr. Graham affirmed the Correctional Service’s Commitment to the problem of Sex Offenders. In his view:

"Sex offenders continue to be one of the most pressing public safety concerns, both within the general public and the correctional community. It is the Service's view that this is an issue of enormous significance facing correctional authorities."

Further Mr. Graham stated that CSC hopes to build on the consensus to be achieved on the standards presented at this conference.

Final comments were made by Dr. Sharon Williams, Corporate Advisor and Chairperson of the Conference: thanking the speakers, participants and working group members. She noted that in this rapidly expanding and high profile field, it is necessary to develop assessment and treatment strategies which are rigorous and well-evaluated.

JOHN MONAHAN (School of Law, University of Virginia)

Mental Disorder and Violent Behaviour

Perceptions and Evidence

Throughout history and in all known societies, people have believed that mental disorder and violence were somehow related. The consensus of modern scholarly opinion, how ever, has been that no such relationship exists. Recent epidemiological studies cast doubt on this no - relationship position. Evidence now indicates that mental disorder may be a consistent, albeit modest, risk factor for the occurrence of violence. Denying that mental disorder and violence may be in any way associated is disingenuous and ulti mately counterproductive. Dire implications for mental patient advocacy, for mental health law, and/or the pro vision of mental health treatment need not follow from candidly acknowledging the possibility of a limited connection between disorder and violence.

Is there a relationship between mental disorder and violent behaviour? Few questions in mental health law are as empirically complex or as politically controversial. On the one hand, the general public and their elected representatives appear firmly committed to the view that mental disorder and violence are connected. On the other hand, many social science researchers and the patient advocates who cite them seem equally convinced that no such connection exists. Although I have long been in the latter camp (e.g., Monahan, 1981), I now believe that there may be a relationship between mental disorder and violent behaviour, one that cannot be fobbed off as chance or explained away by other factors that may cause them both. The relationship, if it exists, probably is not large, but may be important both for legal theory and for social policy. In this article, I lay before you the evidence and the inferences that have persuaded me to modify my views. I first consider the relationship between mental disorder and violence as it has been perceived by public and professional audiences, and then present an epidemiological framework within which the question can be empirically addressed.

Editor's note. Articles based on APA award addresses that appear in the American Psychologist are scholarly articles by distinguished contributors to the field. As such, they are given special consideration in the American Psychologist's editorial selection process. This article was originally presented as a Distinguished Contributions to Research in Public Policy award address at the 99th Annual Convention of the American Psychological Association in San Francisco in August 1991.

Author's note. My work on this topic has been supported by the MacArthur Research Network on Mental Health and the Law. The views expressed here are my own. I am grateful to the members of the Network and to Lawrence Fitch, S. Ken Hoge, Deidre Klassen, Bruce Link, Lee Robins, Joan Roth, Jeffrey Swanson, Linda Teplin, and Simon Wessely for their comments on the manuscript.

Correspondence concerning this article should be addressed to John Monahan, School of Law, University of Virginia, Charlottesville. VA 2290 1.

Mental Disorder and Violence: Public and Professional Perceptions

Beliefs that mental disorder is linked to violent behaviour are important for two reasons. The first is that such beliefs drive the formal laws and policies by which society attempts to control the behaviour of disordered people and to regulate the provision of mental health care. Coherent theories of mental health law can be constructed that are not premised on the assumption that the mentally disordered are more prone to violence than is the rest of the general population (e.g., "Developments in the Law," 1974). But there can be little doubt that this assumption has played an animating role in the prominence of dan gerous to others as a criterion for civil commitment and the commitment of persons acquitted of crime by reason of insanity, in the creation of special statutes for the extended detention of mentally disordered prisoners, and in the imposition of tort liability on psychologists and psychiatrists who fail to anticipate the violence of their patients (Appelbaum, 1988; Grisso, 1991).

The second and perhaps more important reason why beliefs in the violence potential of the mentally disordered are important is that they not only drive formal law and policy toward the mentally disordered as a class, but they also determine our informal responses and modes of interacting with individuals who are perceived to be mentally ill. An ingenious study by Link, Cullen, Frank, and Wozniak ( 1987) vividly makes this point. These researchers investigated the extent to which a person's status as a former mental patient fostered social distance on the part of others, measured by questions tapping the willingness of the respondent to have as a co-worker or neighbour someone described in a vignette as having once been a patient in a mental hospital. Consistent with much prior research (e.g., Gove, 1980), Link et al. (1987) found no main effect of the former-patient label. But when they desegregated their subjects—adults drawn from the open community—by means of a "perceived dangerousness scale" into those who believed that mental disorder was linked to violence and those who did not, strong labelling effects emerged. Remarkably, people who believed that there was no connection between mental disorder and violence exhibited what might be called an affirmative action effect: They responded as if they were more willing to have as a co-worker or neighbour someone who had been a mental patient than someone who had never been hospitalised. People who believed that the mentally disordered were prone to violence, however, strongly rejected and wished to distance themselves from the former patient.

Before considering the contemporary nature of public and professional perceptions of the relationship between mental disorder and violence, it may be useful to briefly set the topic in historical and cultural perspective.

Perceptions in Other Times and Other Places

From the very origins of Western civilisation, most people's experience with the mentally disordered have led them to assume that there was a connection of some kind between mental disorder and violence (Monahan, in press-a). References in Greek and Roman literature to the violence potential of the mentally disordered date from the fifth century before the Christian era began. As the historian George Rosen (1968) noted, in the ancient world "two forms of behaviour were considered particularly characteristic of the mentally disordered, their habit of wandering about and their proneness to violence" (p. 98). Plato, for example, in "Alcibiades II," records a dialogue between Socrates and a friend. The friend claimed that many citizens of Athens were "mad." Socrates refuted this claim by arguing that the rate of mental disorder in Athens could not possibly be very high because the rate of violence in Athens was very low.

How could we live in safety with so many crazy people? Should we not long ago have paid the penalty at their hands, and have been struck and beaten and endured every other form of ill usage which madmen are wont to inflict? (cited in Rosen, p.100)

Likewise, Plautus, in a play written about 270 B.C., titled Casina, wrote of a maid who had taken up a sword and was threatening to murder a lover. One character describes the situation: "She's chasing everyone through the house there, and won't let a soul come near her; they're hiding under chests and couches afraid to breath a word." To this, her lover asks, "What the deuce has gotten into her all of a sudden this way?" The answer he received seemed to suffice for an explanation: "She's gone crazy" (cited in Rosen, p. 99). Advice to those responsible for the care of the mentally disordered in Greece and Rome often made reference to their dangerousness and to the necessity of keeping them in restraints, lest their caretakers be injured.

It is important to emphasise that even in ancient times, the public perception was not that all or most or even many of the mentally disordered were violent, just that a disproportionate number were. The Roman philosopher Philo Judaeus, for example, divided the mentally disordered into two groups. The larger one was made up of disordered people "of the easy-going gentle style," and the other, smaller one, consisted of those "whose madness was . . . of the fierce and savage kind, which is dangerous both to the madmen themselves and those who approach them" (cited in Rosen, 1968, p. 89).

Such public attitudes persisted throughout the Middle Ages and the Renaissance. Care of the disordered was left to family and friends; "only those considered too dangerous to keep at home . . . were dealt with by communal authorities" (Rosen, 1968)

p. 139). An early form of the dangerousness standard for civil commitment is illustrated by the 1493 German case of a disordered man who had committed a violent act and was ordered locked up in a tower of the city wall. When he no longer appeared violent, he was released from the tower to the custody of his family, upon condition that they would confine him themselves should he again become violent. In this event, his wife would confine him in her house or arrange to keep him elsewhere at her ex pense. If required, the council would lend her a jail. (Rosen, p. 143)

Little in terms of public attitudes changed as the Renaissance gave way to the modern era. In 1843, the London Times publishing the following ditty on its editorial page on the day after Daniel McNaughten was acquitted by reason of insanity of murdering the secretary to the prime minister:

Ye people of England exult and be glad For ye're now at the mercy of the merciless mad!

In the United States, as in Europe, the perception of a link between mental disorder and violence is as old as recorded history. The first general hospital in the American colonies to include a ward for the mentally disordered—the cellar—was founded at the urging of no less than Benjamin Franklin. After arguing in vain that the Pennsylvania colony was morally obligated to provide for the disordered, he switched tacks and petitioned the Assembly in 1751 that the Number of Persons distempered in Mind and deprived of their rational Faculties has increased greatly in this province. Some of them going at large are a terror to their Neighbours, who are daily apprehensive of the Violences they may commit. (cited in Deutsch, 1949, p. 59)

This argument hit a responsive chord, and the Pennsylvania Hospital still stands in Philadelphia.

The belief that mental disorder is conducive to violence runs deep in Western culture, but is by no means peculiar to it. Westermeyer and Kroll (1978) studied all persons known as baa, or crazy, in 27 villages in Laos, a country that at the time of the research was without a single psychiatrist, psychologist, or mental hospital. They questioned family members, neighbours, and the people seen as boa themselves about the occurrence of violence and its relationship to mental disorder. They were told that 11% of their subjects exhibited violent behaviour before they began acting in a boa manner, whereas 54% were reported to have acted violently once they became boa. At approximately the same time, Jones and Horne (1973) studied almost 1,000 people in four isolated aboriginal missions in the Australian desert.

Frequently, [they concluded,] an aggressive act by the patient causes him to present clinically, but with an explanation that was culturally appropriate—he would claim, for example, that his symptoms have been inflicted upon him by magical means and his aggression was his way of protecting himself. (p. 225)

Finally, Jane Murphy (1976), the noted anthropologist, reviewed in Science a great deal of research on responses to mental disorder among a variety of Northwestern Native American and several Central African ethnic groups. She reported great similarities among people in very different traditional societies, societies that had never had contact with one another:

There seems to be little that is distinctively cultural in the at titudes and actions directed toward the mentally ill, except in such matters as that an abandoned anthill could not be used as an asylum in the arctic or a barred igloo in the tropics .... If the behaviour indicates helplessness, help tends to be given, es pecially in food and clothes. If the behaviour appears foolish or incongruous..., laughter is the response. If the behaviour is noisy and agitated, the response may be to quiet, sometimes by herbs and sometimes by other means. If the behaviour is violent or threatening, the response is to restrain or to subdue. (p. 1025)

Of course, the anthropological fact that a popular belief has persisted since antiquity and is found in all known societies does not mean that the belief is true. Unfounded prejudices may also be enduring and shared. But if the assumption that mental disorder sometimes predisposes toward violent behaviour is a myth, it may still be worth noting that it is a myth that is both culturally universal and historically invariant.

Contemporary American Perceptions

In modern times and in modern societies, of course, we no longer have to rely on historians and anthropologists to tell us what we believe. We have survey researchers to quantify our opinions. One poll conducted by the Field Institute (1984) for the California Department of Mental Health asked 1,500 representative California adults whether they agreed with the statement, "A person who is diagnosed as schizophrenic is more likely to commit a violent crime than a normal person." Almost two thirds of the sample (61 %) said that they definitely or probably agreed. In modern as in ancient times, however, the public by no means believes that mental disorder inevitably or even frequently leads to violence. In a survey of 1,000 adults from all parts of the United States, conducted by the DYG Corporation (1990) for the Robert Wood Johnson Foundation Program on Chronic Mental Illness, 24% of the respondents agreed with the statement, "People with chronic mental illness are, by far, more dangerous than the general population," whereas twice as many (48%) agreed with the proposition, "The mentally ill are far less of a danger than most people believe."

Although ancient attitudes about the relationship between mental disorder and violence were, of necessity, based on personal observation or word-of-mouth, contemporary opinions no doubt reflect the additional impact of the image of the mentally disordered relentlessly promoted by the media (Steadman & Cocozza, 1978). One content analysis performed for the National Institute of Mental Health (Gerbner, Gross, Morgan, & Signorielli, 1981) found that 17% of all prime-time American television programs that could charitably be classified as dramas depicted a character as mentally ill. Of these mentally ill characters, 73% were portrayed as violent, compared with 40% of the "normal" characters (!), and 23% of the mentally ill characters were shown to be homicidal, compared with 10% of the normal characters. Nor are such caricatures limited to television. A content analysis of stories from the United Press International database (Strain & Phillips, 1991) found that in 86% of all print stories dealing with former mental patients, a violent crime—"usually murder or mass murder" (p. 64)—was the focus of the article.

Professional Perceptions

From reading the literature in this area, it would appear that there are only two identifiable groups in modern society who do not believe that mental disorder and violence are associated at greater than chance levels. The first group is composed of advocates for the mentally disordered, both of the traditional and ex-patient schools. The most recent pamphlet of the established National Mental Health Association (1987), for example, stated that "people with mental illnesses pose no more of a crime threat than do other members of the general population" (p. 2). Likewise, a recent volume produced by a leading ex-patient advocacy group for the California Department of Mental Health (Campbell & Schraiber, 1989) stated that "studies show that while, like all groups, some members are violent, mental health clients are no more violent than the general population" (p. 88). In making such statements, patient advocates are clearly and commendably motivated by the desire to dispel vivid homicidal maniac images pandered by the media and to counter the stigma and social distancing that are bred by public fear. Given the findings of Link et al. (1987), they surely are right to be concerned.

The second group in society that apparently believes that mental disorder is not associated with any increase in the risk of violence consists of many sociological and psychological researchers. Henry Steadman and I (Monahan & Steadman, 1983a), for example, reviewed over 200 studies on the association between crime and mental disorder for the National Institute of Justice. This was our summary:

The conclusion to which our review is drawn is that the relation between ... crime and mental disorder can be accounted for largely by demographic and historical characteristics that the two groups share. When appropriate statistical controls are applied for factors such as age, gender, race, social class, and pre vious institutionalisation, whatever relations between crime and mental disorder are reported tend to disappear. (p. 152)

I now believe that this conclusion is at least premature and may well be wrong. I say this for two reasons. First, to statistically control for factors, such as social class and previous institutionalisation, that are highly related to mental disorder is problematic. For example, if in some cases mental disorder causes people to decline in social class (perhaps because they became psychotic at work) and also to become violent, then to control for low social class is, to some unknown extent, to attenuate the relationship that will be found between mental disorder and violence. "The problem," as Bruce Dohrenwend (1990) has noted, "remains what it has always been: how to unlock the riddle that low SES can be either a cause or a consequence of psychopathology" (p. 45). If, in other cases, mental disorder causes people to be repetitively violent and therefore institutionalised, then to control for previous institutionalisation also masks, to some unknown degree, the relationship that will be found between mental disorder and violence.

The second reason that I now think the no-relationship conclusion may be wrong is that new research—by no means perfect, yet by all accounts vastly superior to what had been in the literature even a few years ago— has become available. These new studies find a consistent, albeit modest, relationship between mental disorder and violent behaviour. I will now turn to this literature, both old and new. As before (Monahan & Steadman, 1983a), I find an epidemiological framework conducive to clear thinking on this topic.

Mental Disorder and Violence: Evidence for a Relationship

There are two ways to determine whether a relationship exists between mental disorder and violent behaviour and, if it does, to estimate the strength of that relationship. If being mentally disordered raises the likelihood that a person will commit a violent act—that is, if mental disorder is a risk factor for the occurrence of violent behaviour— then the actual (or true) prevalence rate for violence should be higher among disordered than among non-disordered populations. And to the extent that mental disorder is a contributing cause to the occurrence of violence, the true prevalence rate of mental disorder should be higher among people who commit violent acts than among people who do not. These two complementary ways of estimating relationships with epidemiological methods follow.

1. True prevalence of violent behaviour among persons with mental disorder

a. Among identified mental patients

b. Among random community samples

2. True prevalence of mental disorder among persons committing violent behaviour

a. Among identified criminal offenders

b. Among random community samples

Within each generic category, two types of research exist. The first seeks to estimate the relationship between mental disorder and violence by studying people who are being treated either for mental disorder (in hospitals) or for violent behaviour (in jails and prisons). The second seeks to estimate the relationship between mental disorder and violence by studying people unselected for treatment status in the open community. Both types of studies are valuable in themselves, but both have limitations taken in isolation, as will become clear.

Violence Among the Disordered

Three types of studies provide data from hospitalised mental patients that can be used to estimate the relationship between mental disorder and violence. One type looks at the prevalence of violent acts committed by patients before they entered the hospital. A second type looks at the prevalence of violent incidents committed by mental patients during their hospital stay. A final type of study addresses the prevalence of violent behaviour among mental patients after they have been released from the hospital. (I restrict myself here to remarking on findings on violent behaviour toward others and exclude violence toward self, verbal threats of violence, and property damage. By men tal disorder, I refer, unless otherwise noted, to those major disorders of thought or affect that form a subset of Axis I of the Diagnostic and Statistical Manual of Mental Dis orders, 3rd edition, revised [DSM - III - R; American Psychiatric Association 1987]. Three excellent recent reviews (Mullen, in press; Otto, 1992; Wessely & Taylor, 1991) make my task of summarising these studies much easier.

Together, these three reviews report on 11 studies published over the past 15 years that provide data on the prevalence of violent behaviour among persons who eventually became mental patients. The time period investigated was typically the two weeks prior to hospital admission. The findings across the various studies vary considerably: Between approximately 10% and 40% of the patient samples (with a median rate of 15%) committed a physically assaultive act against another shortly before they were hospitalised; 12 studies with data on the prevalence of violence by patients on mental hospital wards are found in these reviews. The periods studied varied from a few days to a year. The findings here also range from about 10% to 40% (with a median rate of 25%; see also Davis, 1991).

There is a very large literature, going back to the 1920s, on violent behaviour by mental patients after they have been discharged from civil hospitals (Rabkin, 1979). The best recent studies are clearly those of Klassen and O'Connor (1988, 1990). They find that approximately 25%-30% of male subjects with at least one violent incident in their past—a very relevant, but highly selective sample of patients—are violent within a year of release from the hospital. The ongoing MacArthur Risk Assessment Study (Steadman et al., 1992) is finding that 27% of released male and female patients report at least one violent act within a mean of four months after discharge.

Each of these three types of research has important policy and practice implications. Studies of violence before hospitalisation supply data on the workings of civil commitment laws and the interaction between the mental health and criminal justice systems (Monahan & Steadman, 1983b). Studies of violence during hospitalisation have significance for the level of security required in mental health facilities and the need for staff "raining in managing aggressive incidents (Binder & McNiel, 1988; Roth, 1985). Studies of violence after hospitalisation provide essential base-rate information for use in the risk assessments involved in release decision making and in after-care planning (Monahan, 1988).

For the purpose of determining whether there is a fundamental relationship between mental disorder and violent behaviour, however, each of these three types of research is unavailing. Only rarely did the studies provide any comparative data on the prevalence of similarly defined violence among non hospitalised groups. Steadman and Felson (1984) is one study that did. The authors interviewed former mental patients and a random sample of the general community in Albany County, New York. The percentage of ex-patients who reported at least one dispute involving hitting during the past year was 22.3, compared with 15.1% for the community sample. For disputes in which a weapon had been used, the figures were 8.1 % for the ex-patients and 1.6% for the community sample. When demographic factors were controlled, however, these differences were not significant. Although the rates of violence by mental patients before, during, or after hospitalisation reported in the other studies certainly appear much higher than would be expected by chance, the general lack of data from non patients makes comparison speculative. But even if such data were available, several sources of systematic bias would make their use for epidemiological purposes highly suspect. Because these studies dealt with persons who were subsequently, simultaneously, or previously institutionalised as mental patients, none of them can distinguish between the par ticipation of the mentally disordered in violence—the topic of interest here—and the selection of that subset of the mentally disordered persons who are violent for treatment in the public-sector inpatient settings in which the research was carried out. (There is virtually no research on private hospitals or on outpatients.) Furthermore, studies of violence after hospitalisation suffer from the additional selection bias that only those patients clinically predicted to be non violent were released. Nor can the studies of violence during and after hospitalisation distinguish the effect of the treatment of potentially violent patients in the hospital from the existence of a prior relationship between mental disorder and violence.

For example, to use the prevalence of violence before hospitalisation as an index of the fundamental relationship between mental disorder and violence would be to thoroughly confound rates of violence with the legal criteria for hospitalisation. Given the rise of the dangerousness standard for civil commitment in the United States and throughout the world (Monahan & Shah, 1989), it would be amazing if many patients were not violent before they were hospitalised: Violent behaviour is one of the reasons that these disordered people were selected out of the total disordered population for hospitalisation. Likewise, the level of violent behaviour exhibited on the ward during hospitalisation is determined not only by the differential selection of violent people for hospitalisation (or, within the hospital, the further selection of "violence-prone" patients for placement in the locked wards that were often the sites of the research), but by the skill of ward staff in defusing potentially violent incidents and by the efficacy of treatment in mitigating disorder (or by the effect of medication in sedating patients). As Werner, Rose, and Yesavage (1983) have stated,

To the extent that hostile, excited, suspicious, and recent assaultive behaviour is viewed by ward staffing as presaging im minent violence, it is the patient manifesting such behaviour who is singled out for special treatment (e.g., additional medications, more psychotherapy); such selection may reduce the likelihood of engaging in violence. Thus, paradoxically, if the patient who "looks" imminently violent in this setting is given effective treatments that forestall violent behaviour, he will not in fact engage in violence as predicted. (p. 824)

Because the prevalence of violence after hospitalisation may be a function of (a) the type of patients selected for hospitalisation, (b) the nature and duration of the treatment administered during hospitalisation, and (c) the risk assessment cut-offs used in determining eligibility for discharge, these data, too, tell us little about whether a basic relationship between mental disorder and violence exists. Only by augmenting studies of the prevalence of violence among treated (i.e., hospitalised) samples of the mentally disordered with studies of the prevalence of violence among samples of disordered people unselected for treatment status in the community can population estimates free of selection and treatment biases be offered. Fortunately, a recent and seminal study by Swanson, Holzer, Ganju, and Jono (1990) provides this essential information. Swanson and his colleagues drew their data from the National Institute of Mental Health's Epidemiological Catchment Area (ECA) study (Robins & Regier, 1991). Representative weighted samples of adult household residents of Baltimore, Durham, and Los Angeles were pooled to form a data base of approximately 10,000 people. The Diagnostic Interview Schedule (DIS), a structured interview designed for use by trained lay persons, was used to establish mental disorder according to Diagnostic and Statistical Manual of Mental Disorders, third edition (DSM - III; American Psychiatric Association, 1980) criteria. Five items on the DIS'—four embedded among the criteria for antisocial personality disorder and one that formed part of the diagnosis of alcohol abuse/dependence—were used to indicate violent behaviour. A respondent was counted as positive for violence if he or she endorsed at least one of these items and reported that the act occurred during the year preceding the interview. This index of violent behaviour, as Swanson et al. noted, is a "blunt measure": It is based on self-report without corroboration, the questions overlap considerably, and it does not differentiate in terms of the frequency or the severity of violence. Yet there is little doubt that each of the target behaviours is indeed "violent," and I believe that the measure is a reasonable estimate of the prevalence of violent behaviour.

Confidence in the Swanson et al. ( 1990) findings is increased by their conformity to the demographic correlates of violence known from the criminological literature. As Tables I and 2 indicate (tables not available), violence in the ECA study was seven times as prevalent among the young as among the old, twice as prevalent among men as among women, and three times as prevalent among persons of the lowest social class as among persons of the highest social class.

But it is the clinical findings that are of direct interest here. Table 3 presents the prevalence of violent behaviour during the past year by DSM - III diagnosis. For these data, exclusion criteria were not used: A subject who met the criteria for more than one disorder was counted as a case of each.

Three findings are immediately evident: (a) The prevalence of violence is more than five times higher among people who meet criteria for a DSM - III Axis I diagnosis than among people who are not diagnosable. (b) The prevalence of violence among persons who meet criteria for a diagnosis of schizophrenia, major depression, or mania/bi-polar disorder are remarkably similar. (c) The prevalence of violence among persons who meet criteria for a diagnosis of alcoholism is 12 times that of persons who receive no diagnosis, and the prevalence of violence among persons who meet criteria for being diagnosed as abusing drugs is 16 times that of persons who receive no diagnosis. When both demographic and clinical factors were combined in a regression equation to predict the occurrence of violence, several significant predictors emerged. Violence was most likely to occur among young, lower class men, among those with a substance abuse diagnosis, and among those with a diagnosis of major mental disorder (see Swanson & Holzer, 1991).

One final and equally notable study not only confirms the ECA data but takes them a large step further. Link, Cullen, and Andrews (in press) analysed data from a larger study conducted by Bruce Dohrenwend and his colleagues (Shrout et al., 1988), using the Psychiatric Epidemiology Research Interview (PERI) to measure symptoms and life events. Link et al. (in press) compared rates of arrest and of self-reported violence (including hitting, fighting, weapon use, and "hurting someone badly") in a sample of approximately 400 adults from the Washington Heights area of New York City who had never been in a mental hospital or sought help from a mental health professional with rates of arrest and self reported violence in several samples of former mental patients from the same area. To eliminate alternative explanations of their data, the researchers controlled, in various analyses, for an extraordinary number of factors: age, gender, educational level, ethnicity (Black, White, and Hispanic), socioeconomic status, family

' The items were, (a) Did you ever hit or throw things at your wife/ husband/partner? [If so] were you ever the one who threw things first, regardless of who started the argument? Did you hit or throw things first on more than one occasion? (b) Have you ever spanked or hit a child (yours or anyone else’s) hard enough so that he or she had bruises or had to stay in bed or see a doctor? (c) since age 18, have you been in more than one fight that came to swapping blows, other than fights with your husband/wipe/partner? (d) Have you ever used a weapon like a stick, knife, or gun in a fight since you were 18? (e) Have you ever gotten into physical fights while drinking?

composition (e.g., married with children), homicide rate of the census tract in which a subject lived, and the subject's "need for approval." This last variable was measured by the Crowne-Marlowe (1960) Social Desirability scale and was included to control for the possibility that patients might be more willing to report socially undesirable behaviour (such as violence) than were non-patients.

The study found that the patient groups were almost always more violent than the never-treated community sample, often two to three times as violent. As in the ECA study (Swanson et al., 1990), demographic factors clearly related to violence (e.g., men, the less educated, and those from high-crime neighbourhoods were more likely to be violent). But even when all the demographic and personal factors, such as social desirability, were taken into account, significant differences between the patients and the never-treated community residents remained. The association between mental patient status and violent behaviour, as the authors noted, was "remarkably robust" to attempts to explain it away as artefact.

Most important, Link et al. (in press) then controlled for "current symptomatology." They did this by using the False Beliefs and Perceptions scale of the PERI, which measures core psychotic symptoms via questions such as "How often have you felt that thoughts were put into your head that were not your own?", "How often have you thought you were possessed by a spirit or devil?", and "How often have you felt that your mind was dominated by forces beyond your control?" Remarkably, not a single difference in rates of recent violent behaviour be tween patients and never - treated community residents re mained significant when current psychotic symptoms were controlled. The Psychotic Symptomatology scale, on the other hand, was significantly and strongly related to most indices of recent violent behaviour, even when additional factors, such as alcohol and drug use, were taken into account. Thus, almost all of the difference in rates of violence between patients and non-patients could be accounted for by the level of active psychotic symptoms that the patients were experiencing. In other words, when mental patients were actively experiencing psychotic symptoms like delusions and hallucinations, their risk of violence was significantly elevated, compared with that of non-patients, and when patients were not actively experiencing psychotic symptoms, their risk of violence was not appreciably higher than demographically similar members of their home community who had never been treated. Finally, Link et al. (in press) also found that the Psychotic Symptomatology scale significantly predicted violent behaviour among the never-treated community residents. Even among people who had never been formally treated for mental disorder, actively experiencing psychotic symptoms was associated with the commission of violent acts.

The data independently reported by Swanson et al. ( 1990) and Link et al. (in press) are remarkable and provide the crucial missing element that begins to fill out the epidemiological picture of mental disorder and violence. Together, these two studies suggest that the currently mentally disordered—those actively experiencing serious psychotic symptoms—are involved in violent behaviour at rates several times those of non-disordered members of the general population, and that this difference persists even when a wide array of demographic and social factors are taken into consideration. Because the studies were conducted using representative samples of the open community, selection biases are not a plausible alternative explanation for their findings.

Disorder Among the Violent

Recall that there is a second empirical tack that might be taken to determine whether a fundamental relationship between mental disorder and violence exists and to estimate what the magnitude of that relationship might be. If mental disorder is in fact a contributing cause to the occurrence of violence, then the prevalence of mental disorder should be higher among people who commit violent acts than among people who do not. As before, there are two ways to ascertain the existence of such a relationship: by studying treated cases—in this instance, people "treated" for violence by being institutionalised in local jails and state prisons—and determining their rates of mental disorder, and by studying untreated cases—people in the open community who are violent but not institutionalised for it—and determining their rates of mental disorder.

A large number of studies exist that estimate the prevalence of mental disorder among jail and prison inmates. Of course, not all jail and prison inmates have been convicted of a violent crime. Yet 66% of state prisoners have a current or past conviction for violence (Bureau of Justice Statistics, 1991), and there is no evidence that the rates of disorder of jail inmates charged with violent offences differ from those of jail inmates charged with non-violent offences. So I believe that data on the prevalence of disorder among inmates in general also apply reasonably well to violent inmates in particular.

Teplin ( 1990) reviewed 18 studies of mental disorder among jail samples performed in the past 15 years. Most of the studies were conducted on inmates referred for a mental health evaluation, and thus present obviously inflated rates of disorder. Among those few studies that randomly sampled jail inmates, rates of mental disorder varied widely, from 5% to 16% psychotic. Roth (1980), in reviewing the literature on the prevalence of mental disorder among prison inmates, concluded that the rate of psychosis was "on the order of 5 percent or less of the total prison population" (p. 688), and the rate of any form of disorder was in the 15%-20% range. More recent studies have reported somewhat higher rates of serious mental disorder. Steadman, Fabisiak, Dvoskin, and Holohean ( 1987), in a level-of-care survey of more than 3,000 prisoners in New York State, concluded that 8% had "severe mental disabilities" and another 16% had "significant mental disabilities" (see also Taylor & Gunn, 1984).

Although the rates of mental disorder among jail and prison inmates appear very high, comparison data for similarly defined mental disorder among the general non-institutionalised population were typically not available. As well, the methods of diagnosing mental disorder in the jail and prison studies often consisted of unstandardized clinical interviews or the use of proxy variables, such as prior mental hospitalisation (see, e.g., Steadman, Monahan, Duffee, Hartstone, & Robbins, 1984).

Recently, however, four studies, one with jail inmates and three with prisoners, have become available that use the DIS as their diagnostic instrument. This not only allows for a standardised method of assessing disorder independent of previous hospitalisation, it permits comparison across the studies and between these institutionalised populations and the random community samples of the ECA research.

In the first study, Teplin (1990) administered the DIS to a stratified random sample—one half misdemeanants and one half felons—of 728 men from the Cook County (Chicago) jail. In the most comparable of the prison studies, the California Department of Corrections (1989) commissioned a consortium of research organisations to administer the DIS to a stratified random sample of 362 male inmates in California prisons (see also Collins & Schlesinger, 1983; Hodgins & Cote, 1990; Neighbours et al., 1987). Comparative data from the ECA study for male respondents were provided by Teplin (1990). The findings for current disorder are summarised in Table 4 (not available).

It can be seen that the prevalence of schizophrenia is approximately 3 times higher in the jail and prison samples than in the general population samples, the prevalence of major depression 3-4 times higher, the prevalence of mania or bi-polar disorder 7-14 times higher, and overall, the prevalence of any severe disorder (i.e., any of the above diagnoses) 3-4 times higher. Although there were no controls for demographic factors in the prison study, Teplin (1990) controlled for race and age in the jail study, and the jail-general population differences persisted. Although these studies all relied on male inmates, even more dramatic data for female prisoners have been reported in one study (Daniel, Robins, Reid, & Wilfley, 1988).

These findings on the comparatively high prevalence of mental disorder among jail and prison inmates have enormous policy implications for mental health screening of admissions to these facilities and for the need for mental health treatment in correctional institutions (Steadman, McCarty, & Morrissey, 1989). But given the systematic bias inherent in the use of identified criminal offenders, they cannot fully address the issue of whether there is a fundamental relationship between mental disorder and violence. Mentally disordered offenders may be more or less likely to be arrested and imprisoned than are non-disordered offenders. On the one hand, Robertson (1988) found that offenders who were schizophrenic were much more likely than were non-disordered offenders to be arrested at the scene of the crime or to give themselves up to the police. Teplin (1985), in the only actual field study in this area, found the police more likely to arrest disordered than non-disordered suspects. On the other hand, Klasisen and O'Connor (1988) found that released mental patients whose violence in the community evoked an official response were twice as likely to be re-hospitalised— and thereby avoid going to jail—than they were to be arrested. An individual's status as a jail or prison inmate, in short, is not independent of the presence of mental disorder.

As before, complementary data on the prevalence of mental disorder among unselected samples of people in the open community who commit violent acts is

necessary to fully address this issue. And as before, the analysis of the ECA data by Swanson et al. (1990) provides the required information, which is summarised in Table 5 (not available).

The prevalence of schizophrenia among respondents who endorsed at least one of the five questions indicating violent behaviour in the past year was approximately four times higher than among respondents who did not report violence, the prevalence of affective disorder was three times higher, the prevalence of substance abuse (either alcohol or other drugs) was eight times higher, and overall, the prevalence of any measured DIS diagnosis—which here included anxiety disorders—was almost three times higher.

Implications for Research and Policy

The data that have recently become available, fairly read, suggest the one conclusion I did not want to reach: Whether the measure is the prevalence of violence among the disordered or the prevalence of disorder among the violent, whether the sample is people who are selected for treatment as inmates or patients in institutions or people randomly chosen from the open community, and no matter how many social and demographic factors are statistically taken into account, there appears to be a relationship between mental disorder and violent behaviour. Mental disorder may be a robust and significant risk factor for the occurrence of violence, as an increasing number of clinical researchers in recent years have averred (Bloom, 1989; Krakowski, Volavka, & Brizer, 1986; Mullen, in press; Wessely & Taylor, 1991).

Should further research solidify this conclusion, would it mean—to return to the points we began with— that laws that restrict the freedom of mentally disordered people for long periods of time or the pervasive social rejection of former mental patients are justified, or that the media is correct in its portrayal of people with mental disorder as threats to the social order? No, it would not and for two reasons.

First, as the Link et al. (in press) study makes clear, it is only people currently experiencing psychotic symptoms who may be at increased risk of violence. Being a former patient in a mental hospital—that is, having experienced psychotic symptoms in the past—bears no direct relationship to violence, and bears an indirect relationship to violence only in the attenuated sense that previous disorder may raise the risk of current disorder.

Second and more important, demonstrating the existence of a statistically significant relationship between mental disorder and violence is one thing; demonstrating the social and policy significance of the magnitude of that relationship is another. By all indications, the great majority of people who are currently disordered—approximately 90% from the ECA study—are not violent. None of the data give any support to the sensationalised caricature of the mentally disordered served up by the media, the shunning of former patients by employers and neighbours in the community, or regressive "lock 'em all up" laws proposed by politicians pandering to public fears. The policy implications of mental disorder as a risk factor for violent behaviour can be understood only in relative terms. Compared with the magnitude of risk associated with the combination of male gender, young age, and lower socioeconomic status, for example, the risk of violence presented by mental disorder is modest. Compared with the magnitude of risk associated with alcoholism and other drug abuse, the risk associated with major mental disorders such as schizophrenia and affective disorder is modest indeed. Clearly, mental health status makes at best a trivial contribution to the overall level of violence in society. (But see "Developments in the Law", 1974, on the legal justification—"because [the mentally disordered] are . . . unable to make autonomous decisions" (p. 1233)—for preventively intervening in the lives of disordered people in situations in which we do not intervene with non-disordered people, even when the non-disordered people present a higher risk of violence.)

What, then, are the implications of the conclusion that mental disorder may be a significant, albeit modest, risk factor for the occurrence of violent behaviour? I see four principle ones. First, the empirical question of the relationship between mental disorder and violent behaviour has only begun to be addressed. That major mental dis order as a generic category relates to violence would be important to know, but it is by no means all that clinicians and policy makers need to know. They need to know the specific features of mental disorder that carry the increased risk. Do disordered perceptions (e.g., hallucinations), disordered assumptions (e.g., delusions), or disordered processes of reasoning or affect relate most closely to the occurrence of violent behaviour? It is unclear whether mental disorder should be unpacked by diagnosis, by course, by symptom pattern, or by specific types of offender-victim interactions for the purpose of answering these crucial questions. Indeed, the victim's manner of reacting or overreacting to "fear-inducing" aspects of the disordered person's behaviour may itself be a mediating factor in the occurrence of violence (Link et al., in press). Violence itself may be only a by-product of a more generic tendency to "norm violation" that accompanies some forms of mental disorder. Epidemiological methods have yielded considerable insights in this general research area to date. "It is questionable, however, whether this group comparison approach can shed a great deal of light on more refined questions that may be posed at this point regarding the relationship between mental illness and criminality" (Mulvey, Blumstein, & Cohen, 1986, p. 60). The use of more longitudinal "career" methods at the individual level of analysis may have much to offer in this regard. Such studies could investigate, for example, how a person's likelihood of violence changes as his or her symptoms and life circumstances change.

Second, the data suggest that public education programs by advocates for the mentally disordered along the lines of "people with mental illness are no more violent than the rest of us" may be doomed to failure, as indeed research shows they have always failed (summing & Cumming, 1957). And they should fail: The claim, it turns out, may well be untrue. It will no doubt be difficult for mental health advocates to convey more accurate but more complex information about the relationship between mental disorder and violence in the sound bites and bumper stickers that have come to frame our public discourse. But the flat denial that any relationship exists between disorder and violence can no longer credibly be prefaced by "research shows" (Steadman, 1981). As Swanson et al. ( 1990), in commenting on their ECA data, stated, public fear of violence committed by the mentally disordered in the community is "largely unwarranted, though not totally groundless" (p. 769). 1 agree with Bloom (1989): "Few are interested in either heightening the stigmatisation of the mentally ill or impeding the progress of the mentally ill in the community. Yet this progress is bound to be critically slowed without a realistic look at dangerousness" (p. 253).

Third, the antipathy toward dangerous to others as one criterion for involuntary hospitalisation frequently expressed by mental health professionals and professional organisations may be unwarranted. A concern with violence to others may not be a responsibility arbitrarily foisted on the mental health professions by an ignorant public that would better be left exclusively to the police. A somewhat heightened risk of violence may inhere in the disorders that it is the business of psychologists and psychiatrists to treat. It is not unreasonable of society to ask us to attend to this risk, within the limits of our ability to assess it (Grisso & Appelbaum, in press; Monahan, in press-b).

Finally, the data underscore the need for readily available mental health services in the community and in correctional institutions. If the experience of psychotic symptoms elevates the risk of violence and if psychotic symptoms can usually be controlled with treatment (Krakowski, Jaeger, & Volavka, 1988), then the provision of treatment to people in need of it can be justified as a small contribution to community safety, as well as a telling reflection on our common humanity.


American Psychiatric Association. (1980). Diagnostic and statistical manual of mental disorders (3rd ed.). Washington, DC: Author.

American Psychiatric Association. (1987). Diagnostic and statistical manual of mental disorders (3rd ea., rev.). Washington, DC: Author.

Appelbaum, R (1988). The new preventive detention: Psychiatry's problematic responsibility for the control of violence. American Journal of Psychiatry 145, 779-785.

Binder, R., & McNiel, D. (1988). Effects of diagnosis and context on dangerousness. American Journal of Psychiatry, 145, 728-732.

Bloom, J. (1989). The character of danger in psychiatric practice: Are the mentally ill dangerous? Bulletin of the American Academy of Psy chiatry and the Law, 17, 241-254.

Bureau of Justice Statistics. (1991). Violent crime in the United States (Report No. NCJ-127855). Washington, DC: Author.

California Department of Corrections, Office of Health Care Services. (1989). Current description, evaluation, and recommendations for treatment of mentally disordered criminal/ offenders. Sacramento: Author.

Campbell, J., & Schraiber, R. (1989). In pursuit of wellness: The Well Being Project (Vol. 6). Sacramento: California Department of Mental Health.

Collins, l., & Schlesinger, W. (1983, November). The prevalence of psy chiatric disorder among admissions to prison. Paper presented at the meeting of the American Society of Criminology, Denver, CO.

Crowne, D., & Marlowe, D. (1960). A new scale of social desirability independent of psychopathology. Journal of Consulting Psychology 24, 349-354.

Cumming, E., & Cumming, J. (1957). Closed ranks: An experiment in mental health. Cambridge, MA: Harvard University Press.

Daniel, A., Robins, A., Reid, J., & Wilfley, D. (1988). Lifetime and six-month prevalence of psychiatric disorders among sentenced female offenders. Bulletin of the American Academy of Psychiatry and the Law, 16, 333-342.

Davis, S. (1991). Violence by psychiatric inpatients: A review. Hospital and Community Psychiatry, 42, 585-590.

Deutsch, A. (1949). The mentally ill in America: A history of their care and treatment from colonial times (2nd ed.). New York: Columbia University Press.

Developments in the law: Civil commitment of the mentally ill. (1974). Harvard Law Review, 87, 1190-1406.

Dohrenwend, B. (1990). Socioeconomic status (SES) and psychiatric disorders: Are the issues still compelling? Social Psychiatry and Psy chiatric Epidemiology, 25, 41-47.

DYG Corporation. (1990). Public attitudes toward people with chronic mental illness. Elmsford, NY: Author.

Field Institute. (1984). In pursuit of wellness: A survey of California adults (Vol. 4). Sacramento: California Department of Mental Health.

Gerbner, G., Gross, L., Morgan, M., & Signorielli N. (1981). Health and medicine on television. The New England journal of Medicine, 305. 901-904.

Gove, W. (1980). Labelling and mental illness: A critique. In W. Gove (Ed.), The labelling of deviance: Evaluating a perspective (2nd ea., pp. 264-270). Beverly Hills: Sage.

Grisso, T. (1991). Clinical assessments for legal decision making Research recommendations. In S. Shah & B. Sales (Eds.), Law and mental health: Major developments and research needs. (pp. 49-80). Washington, DC: U.S. Department of Health and Human Services.

Grisso, T., & Appelbaum, R (in press). Is it unethical to offer predictions of future violence? Law and Human Behaviour

Hodgins, S., & Cote, G. (1990, September). Prevalence of mental disorders among penitentiary inmates in Quebec. Canada's Mental Health, 1-4.

Jones, 1., & Home, D. (1973). Psychiatric disorders among aborigines of the Australian desert: Further data and discussion. Social Science and Medicine, 1, 219-228.

Klassen, D., & O'Connor, W. (1988). Crime, inpatient admissions, and violence among male mental patients. International Journal of Law and Psychiatry, 11, 305-312.

Klassen, D., & O'Connor, W. (1990). Assessing the risk of violence in released mental patients: A cross-validation study. Psychological As sessment: A Journal of Consulting and Clinical Psychology, 1, 75-81.

Krakowski, M., Jaeger, J., & Volavka, J. (1988). Violence and psychopathology: A longitudinal study. Comprehensive Psychiatry, 29, 174181.

Krakowski, M., Volavka, J., & Brizer, D. (1986). Psychopathology and violence: A review of literature. Comprehensive Psychiatry, 27, 131148

Link, B., Cullen, E, & Andrews, H. (in press). Violent and illegal behaviour of current and former mental patients compared to community controls. American Sociological Review.

Link, B., Cullen, F., Frank, J., & Wozniak, J. (1987). The social rejection of former mental patients: Understanding why labels matter. American Journal of Sociology, 92, 1461-1500.

Monahan, J. (1981). The clinical prediction of violent behaviour. Washington, DC: U.S. Government Printing Office.

Monahan, J. (1988). Risk assessment of violence among the mentally disordered: Generating useful knowledge. International Journal of Law and Psychiatry, I I, 249-257.

Monahan J. (in press-a). "A terror to their neighbours": Mental disorder and violence in historical and cultural context. Bulletin of the American Academy of Psychiatry and the Law.

Monahan, J. (in press-b). Limiting therapist exposure to Tarasoff liability: Guidelines for risk containment. American Psychologist.

Monahan, J., & Shah, S. (1989). Dangerousness and commitment of the mentally disordered in the United States. Schizophrenia Bulletin, 15, 541-553.

Monahan, J., & Steadman, H. (1983a). Crime and mental disorder: An epidemiological approach. In M. Tonry & N. Morris (Eds.), Crime and justice: An annual review of research, (Vol. 4, pp. 145- 189). Chicago: University of Chicago Press.

Monahan, J., & Steadman, H. (Eds.). (1983b). Mentally disordered of fenders. Perspectives from law & social science. New York: Plenum Press.

Mulvey, E., Blumstein, A., & Cohen, J. (1986). Reframing the research question of mental patient criminality. International Journal of Law and Psychiatry, 9, 57-65.

Mullen, P. (in press). Criminality, dangerousness and schizophrenia. In D. Kavanaugh (Ed.). Schizophrenia: An overview and practical hand book. London: Chapman & Hall.

Murphy. J. (1976). Psychiatric labelling in cross-cultural perspective. Sci ence, 191, 1019-1028.

National Mental Health Association. (1987). Stigma: A lack of awareness and understanding Alexandria, VA: Author.

Neighbours, H., Williams, D., Gunnings, T. Lipscomb, W., Broman, C., & Lepkowski, J. (1987). The prevalence of mental disorder in Michigan prisons. Lansing: Michigan Department of Corrections.

Otto, R. (1992). The prediction of dangerous behaviour. A review and analysis of "second generation" research. Forensic Reports. 5, 103133.

Rabkin, J. (1979). Criminal behaviour of discharged mental patients: A critical appraisal of the research. Psychological Bulletin, 86, 1-27.

Robertson, G. (1988). Arrest patterns among mentally disordered offenders. British Journal of Psychiatry 153, 313-316.

Robins, L., & Regier, D. (Eds.). (1991). Psychiatric disorders in America: The Epidemiological Catchment Area study New York: Free Press.

Rosen, G. (1968). Madness in society: Chapters in the historical sociology of mental illness. Chicago: University of Chicago Press.

Roth, L. (1980). Correctional psychiatry. In W. Curran, A. McGarry, & C. Peny (Eds.), Modern legal medicine, psychiatry and forensic science. Philadelphia: Davis.

Roth, L. (Ed.). (1985). Clinical treatment of the violent person. Washington, DC: U.S. Government Printing Office.

Shain, R., & Phillips, J. (1991). The stigma of mental illness: Labelling and stereotyping in the news. In L. Wilkins & R Patterson (Eds.), Risky business: Communicating issues of science, risk, and public policy (pp. 61 -74). Westport, CN: Greenwood Press.

Shrout, R, Lyons, M., Dohrenwend, B., Skodol, A., Solomon, M., & Kass, E (1988). Changing time frames on symptom inventories: Effects on the Psychiatric Epidemiology Research Interview. Journal of Con sulting and Clinical Psychology, 56, 567-272.

Steadman, H. (1981). Critically reassessing the accuracy of public perceptions of the dangerousness of the mentally ill. Journal of Health and Social Behaviour, 22, 310-316.

Steadman, H., & Cocozza, J. (1978). Selective reporting and the public's misconceptions of the criminally insane. The Public Opinion Quarterly 41, 523-533.

Steadman, H., Fabisiak, S., Dvoskin, J., & Holohean, E. (1987). A survey of mental disability among state prison inmates. Hospital and Com munity Psychiatry, 38, 1086-1090.

Steadman, H., & Felson, R. (1984). Self-reports of violence: Ex-mental patients, ex-offenders, and the general population. Criminology, 22, 321-342.

Steadman, H., McCarty, D., & Morrissey, J. (1989). The mentally ill in jail: Planning for essential services. New York: Guilford Press.

Steadman, H., Monahan, J., Robbins, P., Appelbaum, R, Grisso, T., Klassen, D., Mulvey, E., & Roth, L. (1992). From dangerousness to risk assessment: Implications for appropriate research strategies. Unpublished manuscript.

Steadman, H., Monahan, J., Duffee, B., Hartstone, E., & Robbins, P. (1984). The impact of state mental hospital deinstitutionalization on United States prison populations, 1968- 1978. The Journal of Criminal Law and Criminology, 75, 474-490.

Swanson, J., & Holzer, C. (1991). Violence and the ECA data. Hospital and Community Psychiatry, 42, 79-80.

Swanson, J., Holzer, C., Ganju, V., & Jono, R. (1990). Violence and psychiatric disorder in the community: Evidence from the Epidemiologic Catchment Area surveys. Hospital and Community Psychiatry, 41, 761-770.

Taylor, R, & Gunn, J. (1984). Violence and psychosis: I. Risk of violence among psychotic men. British Medical Journal, 28S, 1945-1949.

Teplin, L. (1985). The criminality of the mentally ill: A dangerous misconception. American Journal of Psychiatry, 142, 676-677.

Teplin, L. (1990). The prevalence of severe mental disorder among male urban jail detainees: Comparison with the Epidemiologic Catchment Area program. American Journal of Public Health, 80, 663-669.

Werner, R, Rose, T., & Yesavage, J. (1983). Reliability, accuracy, and decision-making strategy in clinical predictions of imminent dangerousness. Journal of Consulting and Clinical Psychology, 51, 815-825.

Wessely, S., & Taylor, R (1991). Madness and crime: Criminology versus psychiatry Criminal Behaviour and Mental Health, 1, 193-228.

Westermeyer, J., & Kroll, J. (1978). Violence and mentally illness in a peasant society: Characteristics of violent behaviours and 'folk' use of restraints. British Journal of Psychiatry 117 529 - 541

VERNON QUINSEY (Queens University)

Assessment, Management, and Treatment of Sex Offenders

(executive summary)

The history of the management and treatment of sex offenders has been driven by spectacular incidents of sexual recidivism within particular jurisdictions. Such incidents create strong political pressures on everyone assessing, treating, and making release decisions concerning sex offenders to either become more conservative or to make foolproof decisions. When faced with great uncertainty and strong pressure not to make errors, people are prone to look for certainty in technology or to defer to experts with special knowledge. Sex offenders are, however, much like other offenders and the issues of risk pertaining to them are identical. Although there is a technology of assessment and treatment that is specific to sex offenders and a substantial proportion of them are undoubtedly paraphiliacs or sexual deviants, the technology of assessment and treatment that exists specifically for sex offenders is fallible and will not bear the weight of unrealistic expectations.

The goal of the approach recommended in this article is the improvement of the balance between the civil liberties of offenders and public safety by more accurately appraising risk. This risk can then be linked to dispositional decisions, including the provision of interventions designed to reduce it.

The literature on decision making suggests that the initial step in appraising the dangerousness of an individual is to establish the base rate or the expected likelihood that that person will commit a new violent or sex offence within a specified period of time. The initial estimate of the probability with which a sex offender will commit a new sexual or violent offence can only be made by examining the results of follow up studies of similar offenders. The initial estimate of the likelihood of recidivism is determined primarily by static or historical variables. Although variables such as offence history cannot change with time, they are vital in anchoring clinical judgement in actuarial reality. One of the reasons this anchoring is so important is that unaided human judgement is remarkably insensitive to dramatic differences in base rate in a prediction context. The final appraisal of dangerousness is made by adjusting the initial estimate upward or downward according to dynamic variables such as progress in treatment and type and quality of supervision.

Among sex offenders, a variety of historical factors, such as the number of previous sex offences, predict the likelihood of sexual and violent recidivism. Among child molesters, the sex of the victim and the relationship of the victim to the offender are also important predictors. Intrafamilial (father-daughter incest) offenders have quite low recidivism rates. Among extrafamilial offenders, those with boy victims have double the recidivism rate of heterosexual offenders.

In a recent study of sex offenders who had been assessed at a maximum security psychiatric facility, it was found that rapists were more likely to recidivate than child molesters. Psychopathy, measures of previous criminal history, and phallometric indices of deviant sexual interests were found to be useful predictors of sexual recidivism. A predictor scale developed by scaling a number of univariate predictors performed well in identifying men who committed new sexual or violent offences. It is of interest that another actuarial instrument, the Statistical Risk Appraisal Guide, developed using the same method on a larger sample of offenders (about 10% of whom were sex offenders) predicts sexual and violent recidivism on cross validation a bit better than the instrument described above which was derived exclusively from a sex offender sample. This observation is particularly useful because the general Statistical Risk Appraisal Guide does not require phallometric assessment. Using theoretically relevant and empirically tested predictors, therefore, predictive accuracy can realistically be expected to be in the 80% range.

An actuarial estimate of risk may be adjusted in a conservative manner based upon the idiosyncratic aspects of the particular case. Such adjustment might be indicated by therapeutic outcome, changed opportunities of offending, adequacy of supervision, current compliance with medication or supervision, and so on. In the best circumstance, the clinicians considering an individual case have a theory which identifies the antecedents of violent or sexual offending for that offender. Such individualised theories are often cast in terms of relapse prevention. These antecedents are the clinical issues to be considered in adjusting actuarially determined risk.

Both the actuarial model and clinical appraisals rest upon a detailed and corroborated history. Assessments of dangerousness should not be based solely on offender self report.

Broadly speaking, sex offender treatment programs employ three approaches: (a) pharmacological, in which the goal is to reduce sexual arousability and the frequency of deviant sexual fantasies through the use of anti-androgens; (b) psychotherapeutic or evocative in which the goals include increasing offender empathy for the victims of sexual assault together with their sense of responsibility for their sexual crimes, and (c) cognitive-behavioural where the object is to remedy skill deficits, alter cognitions that are believed to be related to sexual offending, and alter deviant patterns of sexual arousal or preference. Many programs also employ a cognitive behavioural relapse prevention orientation borrowed from the substance abuse area in which the focus is on eliminating idiosyncratically defined precursors of relapse and teaching the offender more effective ways of coping with these precursors in an extensive period of follow up supervision.

The evidence for the efficacy of these treatment programs is mixed and there are many methodological problems in this outcome literature. Perhaps the strongest conclusion that one can draw is that the aspects of treatment, client population, supervision, and setting characteristics related to successful outcome are at present unknown. The major implication of this observation is that progress in treatment must be interpreted very cautiously when assessing risk.

To say that treatments have not been convincingly evaluated, however, is neither to say that they do not work nor to assert that different approaches to treatment are of equivalent efficacy. The best option in these circumstances of relative ignorance is to adopt treatments that: (a) fit with what is known about the treatment of offenders in general, (b) have a convincing theoretical rationale in that they are motivated by what we know about the characteristics of sex offenders, (c) have been demonstrated to produce proximal changes in theoretically relevant measures, (d) are feasible in terms of acceptability to offenders and clinicians, cost, and ethical standards, (e) are described in sufficient detail that program integrity can be measured, and (f) can be integrated into existing supervisory procedures.

Because sex offenders are by definition criminal offenders, it is reasonable to expect that principles of treatment that apply to offenders in general also apply to sex offenders. The support for a cognitive-behavioural approach to offender treatment in general is based on a much stronger and more extensive literature than for cognitive-behavioural approaches to sex offender treatment. This more general literature, however, indirectly supports the cognitive-behavioural treatment of sex offenders.

The principles of offender treatment have perhaps been best conceptualised in terms of risk, need, and responsivity. Treatment is most effective when targeted at the criminogenic needs of high risk cases. Based upon the correctional treatment literature, characteristics of programs that have some hope of success in reducing recidivism include: a skill based training approach; the modelling of prosocial behaviours and attitudes; a directive but non-punitive orientation; a focus on modifying antecedents to criminal behaviour; a supervised community component in order to assess and teach the offender relevant skills; and a high risk clientele.

Dynamic predictors can also be monitored post-release to good advantage. Instability of living conditions, non-compliance with medication or supervision, increased drinking, negative affect, and procriminal attitudes are all variables that can be related to recidivism or relapse. These post-release predictors can be used to titrate the amount of supervision an offender receives.

The probability and type of recidivism is strongly affected by victim age, sex, and relationship to the offender, the seriousness and nature of the sex offence, and the number of previous sex offences. Because sex offenders are heterogeneous even within categories defined by offence history, sex offender treatment programs should be organised so as to take account of these differences.

Perhaps the most important of these differences and the most relevant for the design of individual treatment programs, is the nature of the offender's sexual preferences. Some offenders have marked paraphilic (sexually deviant) interests in children, sadistic sexual assault, and so on. These sexual preferences can be measured with varying degrees of adequacy by offender self report, offence history, or phallometric assessment. The measurement of these interests is important because it provides clues as to the motivation underlying the offence, an idea as to the nature of possible future acts of sexual aggression, and a focus for a treatment intervention.

Other offender characteristics are related to the probability of treatment success and the design of individual programs. Alcohol abuse is a common problem among sex offenders, as it is among offenders more generally. Such problems can, if not effectively addressed, undermine treatment effectiveness by reducing offender compliance and self control.

The limitations of current treatment technology also interact with individual differences among offenders. Variations in the seriousness of offence history are among the most important of these. In view of the limitations of current treatment technology, it cannot be expected that very serious sexual offenders, such as serial murderers, will or should be viewed as less of a risk as a result of progress in a treatment program.

This brief overview suggests an individualised treatment planning process that uses the results of a variety of standardised assessments to formulate a theory of offender motivation and choose a combination of specific interventions designed to prevent recidivism. Thus, for an individual offender, any or all of a variety of interventions, such as treatments designed to reduce sexual arousability, modify inappropriate sexual preferences, control drinking, improve assertive or heterosocial skills, secure employment, prevent depression, etc., might be appropriate.


Assessment, Management, and Treatment of Sex Offenders

Vernon L. Quinsey, Ph.D.

1. Actuarial instruments for the prediction of violent or sexual re-offending among rapists and child molesters have recently been developed.

2. These instruments can help determine the appropriate level of supervision for individual offenders and in selecting offenders for programs designed to reduce the risk that they present to the community.

3. The sex offender treatment outcome literature does not permit strong conclusions regarding efficacy.

4. What is known about the characteristics of sex offenders and about offender treatment in general suggests that interventions to reduce risk should focus on the criminogenic needs of individual offenders and are best conceptualised in a relapse prevention model.

JANICE K. MARQUES (California State Department of Mental Health)

Sex Offender Treatment And Evaluation Project

Progress Report, 11/94

The Sex Offender Treatment and Evaluation Project (SOTEP) is a legislatively-mandated, longitudinal study of the effectiveness of an intensive treatment program for incarcerated rapists and child molesters. The goals of SOTEP are: (a) to prevent reoffense among rapists and child molesters who are released to the community, and (b) to provide California's Governor and Legislature with a sound basis for determining future public policy regarding this dangerous population. The project is specifically designed to determine whether a state-of-the-art sex offender treatment program can significantly enhance public safety by reducing the number of women and children who are sexually assaulted.

Legislative Background

In 1981, SB 278 repealed existing statutes providing for the commitment of Mentally Disordered Sex Offenders (MDSO's) to state hospitals, and required that convicted sex offenders be sent to the Department of Corrections after sentencing. While this legislation eliminated the direct commitment of these offenders to state hospitals, it also added Section 1364 to the Penal Code, which allowed for the voluntary transfer of certain sex offenders to the Department of Mental Health for treatment during the last 2 years of their prison terms. Section 1365 of the Penal Code required that this new state hospital program be "established according to a valid experimental design in order that the most effective, newest and promising methods of treatment of sex offenders may be rigorously tested."

Project Description

The key features of SOTEP are: (a) a rigorous design that includes random assignment of volunteers to either treatment (state hospital) or control (prison) conditions; (b) an intensive, highly structured treatment program designed specifically to prevent relapse among sex offenders; (c) a one-year aftercare program designed to maintain treatment gains and monitor released offenders; and (d) a comprehensive evaluation of both immediate and long-term treatment effects (including a follow-up period of 5-14 years in which recidivism rates for treated and untreated subjects are measured).

SOTEP is funded by the California Departments of Mental Health and Corrections, and by the National Institute of Mental Health. Treatment services are provided on a 33-bed unit at Atascadero State Hospital. The treatment program opened in 1985, and is currently scheduled to end in 1995.

During its nine years of operation, SOTEP has gained international recognition for both its innovative Relapse Prevention program and its rigorous evaluation of treatment outcome. Support for the project has been expressed by a broad range of professional and community groups representing law enforcement, prosecutors, judges, and crime victims, as well as local, state and national child abuse prevention and treatment organisations. SOTEP's treatment and evaluation methods are being applied in numerous other state programs, and its Relapse Prevention model has recently been adopted by HM Prison Service (England) and by the Correctional Service of Canada. As a panel of experts recently concluded at the New York Academy of Sciences' conference on sexual aggression, "many states are watching California's program as a possible model."

Since 1989, additional research activities using the project's database have been funded by a multi-year enhancement grant from the National Institute of Mental Health. The major aims of the grant are to improve the methods used to measure sex offender recidivism, to identify significant predictors of rape and child molestation, and to study re-offenders to learn more about risk factors for recidivism.

Preliminary Findings

As of September 1993, 195 offenders have been admitted to SOTEP's treatment program, 159 of whom have been released. Of those discharged, 123 completed the program and went into the aftercare program, while 36 were returned to the Department of Corrections before finishing the program. Preliminary findings indicate that those completing the program have made significant progress toward several important goals, including: (a) an increased sense of personal responsibility and decreased attempts to justify sexual deviance; (b) a decrease in deviant sexual interests; and (c) better skills in the areas of avoiding and coping with high-risk situations.

At this point in SOTEP's longitudinal study, only preliminary and incomplete data are available on the effects of treatment on recidivism, and valid statistical tests cannot be conducted. Early results have been mixed: The 1993 panel of data showed that 11% of the treated group had committed new sex or other violent crimes, compared with 17% and 15% of the matched offenders in the two untreated control groups. This year's data show that 25% of the treated offenders have reoffended, compared with 24% and 27% of those in the control groups. Given the preliminary nature and volatility of these data, conclusions about the treatment program's effectiveness cannot be made until the end of the study in about 6 years. In addition to information on the overall effects of treatment, the study will provide data on the importance of various risk factors for reoffense, and will evaluate the relative effectiveness of treatment for different types of sex offenders. This information will provide a basis for future policy decisions regarding how treatment resources can be used to have the greatest impact on public safety.

Recent SOTEP Publications

Marques, J. K., Nelson, C., West, M. A., & Day, D. M. (1994). The relationship between treatment goals and recidivism among child molesters. Behaviour Research and Therapy, 32, 577-588.

Marques, J. K., Day, D. M., Nelson, C., & West, M. A. (1994). Effects of cognitive-behavioural treatment on sex offender recidivism: Preliminary results of a longitudinal study. Criminal Justice and Behaviour, 21, 28-54.

Marques, J. K., Day, D. M., Nelson, C., & West, M. A. (1993). Findings and recommendations from California's experimental treatment program. In G. C. N. Hall, R. Hirschman, J. R. Graham, & M. S. Zaragoza (Eds.) Sexual aggression: Issues in etiologv, assessment, and treatment (pp. 197-214). Washington, DC: Taylor & Francis.

HOWARD BARBAREE (Clarke Institute of Psychiatry and Warkworth Sexual Behaviour Clinic)

Treating Sex Offenders In A Society Intolerant Of Sexual Reoffences

I’m going to present my views from the perspective of a treatment provider and professional clinician. I believe that what I’ve learned will be of interest to other sex offender treatment providers. I believe also however, that my remarks will be relevant to case management officers and decision makers such as wardens and members of the parole board, etc. Some of my remarks today will pertain to the law. The law suits and the events leading to it have been widely covered in the press and are general and public knowledge. Out of respect for the victim and the privacy of the family, that have not been named in the press and I will not do so here. The offender’s name, however, has been widely used in the press, so I will use it in this presentation.

Phillippe Clément is a 37 year old sex offender serving a life sentence for second degree murder. In Montreal, at the age of 17, he killed a woman after a sexual assault, stabbing her 80 times. After serving several years of his criminal sentence in Quebec, Mr. Clément was treated at Pinel Institute for a period of about five and a half years. In 1989, Mr. Clément was transferred to the Warkworth Institute in Ontario where he participated in the treatment program at the Warkworth Sexual Behaviour Clinic for a further one and a half years. In early 1991 we submitted a written post treatment report containing our assessment of the progress we felt he had made in therapy, our post treatment assessment of Mr. Clément’s risk for reoffence and our recommendation that Mr. Clément would be appropriate for a transfer to a minimum security institution. With a similar recommendation from case management, Mr. Clément was transferred from a medium security Warkworth Institution, to a minimum security Beaver Creek Institution in August of 1991. Beaver Creek is set in a rural area near Braisebridge in the Muskoka area north of Toronto. Mr. Clément spent 13 months at Beaver Creek, participated in temporary absence programs, was seen regularly by a contract psychologist, and was described there as a model prisoner.

Shortly after 6:00am on the morning of September 15, 1992, Mr. Clément walked away from Beaver Creek just 2 hours before his parole hearing was due to start. At about 10:00am, wielding a knife, he entered a nearby home. There he found a middle aged mother of 5 alone, after sending her children to school. After demanding money, car keys and liquor, he brutally assaulted her both physically and sexually. He later left the home and was apprehended 2 days later.

Early in 1993, the CBC’s Fifth Estate aired a program on Mr. Clément and his reoffence. In interview with the Fifth Estate Mr. Clément avoided responsibility for his reoffence blaming the Correctional Service of Canada, various individual correctional staff, and treatment professionals for the reoffence. He claimed that his treatment had been inadequate, and that the treatment professionals had not been properly trained or prepared to deal with his problems. He claimed further that he had affairs with various female correctional and treatment staff over his years of incarceration. Most recently at Beaver Creek, previously at the Warkworth Sexual Behaviour Clinic and earlier at Pinel. Clément claimed that his recent violent reoffence was triggered by the female correctional officer’s rejection of him. He also claimed that the correctional authorities should have seen the precursors of his offence and that they should have appreciated the recurrent pattern of his violent behaviour. He claimed that he still harboured anger at a female therapist at the WSBC for her rejection of him. Taking Mr. Clément’s lead, the Fifth Estate painted a sorted picture of sexual inpropriety, potential professional misconduct and incompetence among treatment and correctional staff. The program criticised the authorities, including me, for not providing adequate treatment, adequate supervision of female staff, for assigning female staff to Mr. Clément who is a known woman hater, for not responding appropriately to the repeated pattern of affairs and rejection followed by violent or near violent behaviour, and for housing such a dangerous offender at a minimum security institution. In addition, I was accused of not heeding warnings that had been expressed to me by one of my staff.

In the spring of 1993, Mr. Clément’s victim and her family filed suit against Mr. Clément, the Correctional Service of Canada, various individuals in the CSC, and me. The plaintiff’s claimed 3.75 million dollars in damages. The suit claimed negligence on our part and the nature of their allegations followed closely the Fifth Estate program. Depositions were taken over that summer and later fall. Various attempts were made to settle the suit to no avail. Court proceedings began on August 16, 1994 and continued until the end of November with a total of 42 days of testimony. I personally had the unhappy distinction of spending more hours than anyone else in the witness box giving a total of 5 days of testimony.

Madame Justice Jay Lange released her judgement on the case on February 9, 1995. In one sense we lost the suit. Justice Lange found liability on the part of the CSC on the issue of the timeliness of the reporting to the OPP that Mr. Clément had walked away. However, on other issues, we defended ourselves successfully. From the judgement statement summary on liability Madame Justice Lange writes, "the individual defendants and Dr. Barbaree acted properly in all circumstances and are not liable for the harm suffered by the plaintiff. CSC was not negligent either in transferring Mr. Clément to Beaver Creek or in monitoring him while he was at Beaver Creek."

In her findings, contrary to the approach taken by the Fifth Estate, Justice Lange does not place any weight on the testimony of Mr. Clément. She writes, "Mr. Clément is suffering from a disorder that prevents him from distinguishing between fantasy and reality. Although, Clément was eager to blame others for his actions, he was unwilling to accept the appropriate responsibility himself.

Justice Lange found that the affairs that Mr. Clément had claimed at Beaver Creek and Warkworth were fictions and had no basis in fact. Again from the judgement statement summary on liability, Madame Justice Lange writes, "As well, Phillippe Clément is liable for his assault on the plaintiff. His claim that he should escape liability because he was incompetently managed and treated by the CSC staff is absurd. Mr. Clément must be responsible for his own conduct and viciously assaulting the plaintiff."

Importantly for all of us I think, Justice Lange provides strong support for the treatment that had been provided to Mr. Clément for the assessment of risk that we had done and the recommendations for the transfer to minimum security. Further, in the judgement, Justice Lange gives support to the appropriate use of professional judgement. She writes, "with the clarity of hindsight it would be easy to look back to the 3 days preceding the walkaway and see that CSC should have realised that Clément had become a security risk. However, to do so would be wrong. Rather I must assess CSC’s conduct from the perspective of what was known at that time. Conscientious and dedicated staff exercised their best judgement based on years of experience and armed with adequate knowledge of Clément’s history and circumstances. The court would err if it tried to now second guess those decisions."

I think that the judgement of Judge Lange will help in making those judgements and being able to defend them in the future.

I thought it would be helpful for me now, for treatment providers particularly, to review in general terms, the way in which plaintiffs and their lawyers attacked the Warkworth Sexual Behaviour Clinic and its staff and the assessment and treatment that Mr. Clément received there. Lawyers launching future cases of this kind will study their case carefully and the plaintiff’s strategy in our case may provide a kind of blueprint for their future actions. Future lawyers may learn from what they perceived as mistakes by the current legal teams so we may predict what the future legal teams might do. My experience here may assist in preparing you for this kind of action in the future.

First I’ll go through the various ways in which the plaintiff’s legal team made criticism of the work that we had done. I’m putting this in a kind of general term so that you can, as we’re going through these, imagine how you would defend yourself in an action like this.

First of all, plaintiff’s will claim the treatment provided was ineffective, that it was incompetently done, that it might have been inappropriate for this particular offender. We defended ourselves in this criticism by comparing our treatment program, in terms of its content and methodology we used, with the treatment programs made available to offenders in other institutions. We compared our program with those that are described in the literature and we showed that our program was state of the art in many ways and that it contained interventions targeting those problems that are generally recognised as problems in sex offenders.

After this mornings presentation we’ll try to keep future legal teams from any contact with Janice Marques. She would be a bad expert witness against us in the sense of the ineffectiveness of the program anyway.

It would be very helpful here with respect to the effectiveness of treatment if their were standards of care that were generally recognised. I understand that such standards are being worked on and yesterday afternoon some of that work was presented to you. It would make it much easier to present yourself in court if you were able to point to a document which describes the standards of care and then be able to show simply that your program meets those standards.

The plaintiff will claim that the assessment of risk was inaccurate, incompetently done or inappropriate. They’ll do that in terms of your assessment of risk as the man comes into the program and goes out of the program. We defended our assessment by describing the method of risk assessment that we use in detail, showing how it is similar to the methods of risk assessment that are currently being recommended by authorities in the field. Again, if we had standards of care in this regard and we could simply just point to these standards and say that the program met the standards, then we would be able to defend ourselves a lot more easily.

One of the issues we faced in this particular case was some contention on the basis of the plaintiff’s argument and also with the Fifth Estate that assigning a female therapist to Mr. Clément was inappropriate given his history. We argued that it would be inappropriate to withhold contact with females from a man who was slated for some kind of gradual release in the future and that if you’re going to have difficulty with females in a medium security setting, there would be no rationale for having them released or transferred to lower security from there. We also argued that it was a bit chauvinistic to say that females would make less effective therapists than males for any of the offenders that we see. In the end, in the judgement, Judge Lange agreed with our argument.

We also defended ourselves on this count by describing in detail the training and preparation that our therapists have had. In particular we described the university courses and the onsite training that we provided to the therapist that was in question here. Again, you can defend that successfully. One of the problems I think we have as a group in defending ourselves on this issue, however, is that there are no programs of certification or training for people working in this field. The amount of and experience training that people have varies widely from totally inadequate to expert. I think what we need to do in the long run to help defend ourselves here is to develop programs and training that are nationally recognised that provide for some kind of certification and more properly prepare people to work with these offenders. You have to be able to show that the people who are working for you are properly supervised and that means that the advice that you give them is documented, that they meet with you on a regular basis and seek advice when it is appropriate. You get yourself into difficulty here when you’re dealing with an employee who is disgruntled. In our case we had an employee who claimed, in the stand, that she had not been properly supervised. Luckily, I had documents that indicated we had supervised her quite closely and there was no evidence indicated by her that she had not received appropriate supervision. So, it’s really important here that, as part of your practice, you document the supervision that you provide and hope that employees don’t leave disgruntled.

Whenever you take out the file on someone such as Mr. Clément who has a history of violent assaults against women or against children, you’ll find scattered through the file, various other psychological and psychiatric reports. Very often in cases like this, sometimes the language in those reports can be described as temperate. They say things like, "This man should never be released." In this particular case, one of the phrases used was, "Any contact with women will reawaken the anger that this man has towards women." I’m not arguing here at all that people shouldn’t say what they feel in their professional reports. I know as well as anyone that when you say these things in reports, sometimes you have to sit in the courtroom and justify them. The problem that you have with those, or opinions at least, that seem to argue against the recommendations that you made are in the record. In defending yourself, first of all, you have to be able to say that you knew those opinions were there, that you considered them carefully and that you made your recommendation in the face of those contrary opinions.

I would just like to comment that if you have concern for your fellow professionals in this area, I encourage you to use temperate language in their reports. You can get the same message using language that makes it easier for us to defend ourselves in court. One of the issues that always comes up - and this is like the Terrasoff issue - the plaintiffs will claim that you had information about the individuals heightened risk, that you kept it to yourself to support your opinion perhaps or for some other ulterior motive that you failed to properly inform the authorities. If you did properly inform the authorities and you can show that, then they’ll complain that you didn’t do it in a timely way or they may complain that you didn’t do it in a proper written format. It’s sometimes difficult to defend yourself against allegations like this. The answer to this problem is to write down everything you ever do so that you can refer to that later and use that documentation to support the decisions that you’ve taken. Whenever I’ve tried to do that I never end up doing anything else but write. I think it’s safe to say that whenever you get yourself into a situation like this, your documentation is not as good as you would like it to be and that’s a fact of life that you just have to live with. But you should document as much as you can.

The plaintiffs will claim that the progress you saw the inmate make in therapy was a fiction, it wasn’t real and the inmate fooled you into thinking that treatment had been more effective than it was and that therefore, your recommendation for transfer was based on erroneous change. In this particular issue, I guess the only thing you can do is throw yourself at the mercy of the court and say that here’s what we do to try and assess that. Here’s the information we had at the time. We recognise the possibility that the treatment gains may have been exaggerated by the offender. But we made a judgement under the circumstances based on the information that we had, that the treatment was sufficient to support the transfer that we recommended. Again this issue of decisions taken regarding the offender were not taken thoughtfully, with due care and attention to detail, with consideration of all factors that should be relevant to the decision. Again, it is a matter of documentation. You have to be able to go back and remember what it is you thought of and what you considered when you decided to do the lesson itself. If you don’t have notes or you don’t have people who took part in the decision with you that can help you remember, then sometimes it’s difficult to defend yourself. The secret again is to keep as many notes as you can.

We’ve covered now the ways in which plaintiffs can take issue with the work that you’ve done with offenders and given you a bit of an idea how you can defend yourself. Obviously, this argument needs too be flushed out in many ways. I think we have to think about how we can defend ourselves in response to these kinds of allegations. One of the things that you should be aware of apart from that is that the plaintiffs will be required to show - even if they are able to show that you were not operating at a professional standard on any of the issues that we’ve just gone through.

In order to be able to demonstrate the liability or support liability, they have to that you had a duty of care with respect to the particular victim involved. In the judgement, Judge Lange quotes from a decision by a Lord Atkin in 1932 which describes what a duty of care is. I’ll read it. It’s a bit confusing but it says it as well as I can. "The rule that you are to love your neighbour becomes in law - you must not to injure your neighbour and the lawyers question, "Who is my neighbour?" receives a restricted reply. You must take reasonable care to avoid acts or emissions which you can reasonably foresee would be likely to injure your neighbour. Who then in law is my neighbour. The answer seems to be persons who are so closely and directly effected by my act that I ought reasonably to have them in contemplation as being so effected while I am directing my mind to the acts and omissions which are called in question."

The example of a case that Judge Lange quotes is a case where a number of juvenile delinquent boys were on an outing in England on an island and they escaped from the care of the correctional people who were watching them, stole a boat from the Dorsette Yacht Company, and in making their escape, damaged the boat. In this particular case, the court found that the correctional authorities had a duty of care with respect to the yacht company because they ought to have known these boys would be motivated to escape and, on an island, they would be likely to use a boat to make good their escape. There are two issues here that contribute to this duty of care. One is the forseeability of harm. The other is the proximity of relationship. In regards to proximity, there is an American case where a man escaped from a county jail, went to another city and a couple of days later raped and murdered a 9 year old girl. The sheriff in the case was sued. But the court found that this particular sheriff didn’t have a duty of care because the proximity was not established. The offender was far enough away from the offender and the sheriff could not be expected to prevent the offence having occurred so far away.

I think what we have to develop for ourselves here is a level of comfort in making these decisions. I don’t think that we’re responsible for the reoffences that occur by the men who have been in treatment programs with us when they are released and out of our care and they re-offend against a member of the general public. In the case of Mr. Clément and his reoffence, the duty of care was established by the plaintiffs because he escaped a known woman hater and attacked a woman in the vicinity within hours of walking away from the institution. So the plaintiffs were able to establish this forseeability of harm and the proximity of relationship.

It’s important for us to begin a discussion of how we might develop our practices so that they’re more defensive, so that we’re better able to defend ourselves in response to allegations such as these. Thank you very much.


(Forensic Behavioural Management Clinic, Native Clan Organisation)

Moving Towards Holistic Healing For Aboriginal Sex Offenders

In the fall of 1994, the Native Clan Organisation’s Forensic Behavioural Management Clinic commenced a new sex offender group, something we had done several times since the Clinic began providing community based and institutional sex offender programming in 1987. However, this group began in a much different manner than those which had proceeded it. Rather than meeting in our office and beginning with ice breaking and rapport building exercises to initiate the 18 month process of a cognitive-behavioural, relapse prevention sex offender treatment program, the 12 offenders, aboriginal and non-aboriginal alike, the clinical treatment team and our Elder gathered at a ceremonial grounds, outside the city limits, far removed from our downtown office to commence this program with a Sweat Lodge ceremony. The circle of chairs in the group room was replaced by a teaching circle out of doors at the site of the Sweat Lodge. The Elder gathered everyone into a circle, had us join hands and discussed the process of healing, the importance of ceremony and traditional ways and explained the Sweat Lodge ceremony we were about to take part in. With this we disrobed to our shorts, walked around the fire which had been heating the rocks for the Sweat Lodge for several hours, made an offering of Tobacco and entered the Lodge. So, there I was this little Jewish guy from the south end of Winnipeg sitting in my Calvin Klein boxer shorts in a Sweat Lodge with a group of half naked sex offenders. Nobody from the Department of Psychology at the University of Manitoba said that there would be days like these!

This was an important day for us as clinicians, for the treatment program and, most importantly, for the offenders. Prior to this, we were aware of the special needs that some aboriginal offenders had presented and attempted to address these in a variety of ways. We adjusted individual and group process to compliment the style of presentation we had observed in a number of the aboriginal offenders in our program, we targeted topic areas that were particularly relevant to the aboriginal men (ie. prejudice, displacement, identity issues) and we encouraged men to become familiar with and involve themselves in their culture, tradition and to access Elders and participate in traditional ceremonies. However, until the day of our first ceremony, traditional healing had never been a part of the actual treatment program.

A decision to look beyond what was considered to be 'state of the art' sex offender treatment programming and approach native Elders to see if there was a way to incorporate traditional healing into the program was made after reviewing treatment outcome data from our sex offender program. What we found was that while aboriginal and non-aboriginal offenders who had completed the treatment program appeared to have benefited equally (in terms of recidivism rates), only half the number of aboriginal offenders completed the program compared to non-aboriginal offenders. The aboriginal offenders in the program were more likely to be suspended during treatment for breaching conditions of their conditional releases (ie. breaching abstain orders, failing to return to community release facilities), more often dropped out of treatment at the completion of their sentence after which there was no longer a legal mandate for them to attend, and, were more likely to re-offend sexually. We interpreted this information to be telling us that the aboriginal offenders in our program did not appear to become as invested or engaged in the treatment process as compared to non-aboriginal offenders and that our emphasis on a cognitive-behavioural, relapse prevention approach did not appear as relevant or meaningful to them.

There are a number of hypotheses that could be put forward which may account for the identified differences in responsiveness to treatment between aboriginal and non-aboriginal offenders. The aboriginal offenders in our program tended to have more chronic histories of offending than did the non-aboriginal offenders. They had sexually offended against a greater number of victims, had longer histories of inappropriate sexual behaviour and tended to report more extensive histories of non-sexualized violence. One would expect that offenders who evidence histories of chronic aggressive and sexually abusive/aggressive behaviours, regardless of race, would be at greater risk and, that treatment prognosis would be guarded. Treatment differences between aboriginal and non-aboriginal offenders may also be attributed to the range and degree of deficit areas experienced by the aboriginal offenders in our program. While most of the offenders we see present with multiple problem areas, this has been particularly apparent among the aboriginal offenders we have assessed and treated. Issues such as substance abuse, histories of abandonment, dislocation, victimisation, and identity issues are prominent and likely contribute to the observed differences in treatment outcome. Finally, it is at times the case that aboriginal offenders who originate from rural communities are released from correctional institutions to an urban centre for parole supervision and treatment programming. It is probable that a lack of experience with the environment of a city, limited or no contact with their home community and the absence of adequate prosocial community supports while away from the home community contribute to a difficult and all to often unsuccessful transition to the community.

While these various postulations may assist us to understand, explain and account for the treatment outcome differences between aboriginal and non-aboriginal offenders they also present us with a challenge. How do we as clinicians re-think our treatment models and attempt to develop and implement programming that will engage aboriginal offenders in a meaningful way while addressing issues related to the offenders level of risk and need. For us this meant looking at implementing both contemporary as well as traditional treatment/healing as part of our sex offender treatment program.


The first step in developing meaningful programming for aboriginal offenders must involve an awareness, appreciation and acceptance of the importance of healers and traditional healing. While it is not necessary to be versed with a detailed knowledge of the various traditional practices, there needs to be a recognition of the level of respect many aboriginal offenders have for native Elders, the degree of influence Elders possess, the customary role and ability of Elders to provide teachings and guidance, and of the powerful nature of traditional ceremony. Once the significance and potential benefits of traditional healing and healers is recognised, there is an enhanced ability to see, understand and accept the role healers and ceremony can play in addressing criminogenic factors and managing risk.

This is truly significant as, within corrections, credibility for rehabilitation is granted to psychologists, psychiatrists, program providers and case management/ parole officers and an emphasis is placed on taking programs to address substance abuse, criminal thinking, anger management and sex offender treatment. However, historically participation in traditional healing practices and counselling with Elders has not received the same level of recognition and respect as being a legitimate means of addressing risk/need. If an understanding of the potential relevancy, influence and benefits of traditional healing as `programming' develops, there is a greater chance that there will be a shift to include traditional healing practices and ceremonies as appropriate programming for aboriginal offenders.


In attempting to move towards providing culturally appropriate programming for aboriginal offenders, partnerships must be developed between healers, Elders and others providing services to native offenders. Just as case management officers and parole officers might seek consultation from or refer an offender for treatment to a mental health professional or a program, both case managers/supervisors and mental health professionals need to become open to seeking consultation and referring offenders to healers and Elders. The work of these healers also needs to hold a similar status to that of others who provide consultation or programming to offenders.

In attempting to provide more appropriate programming to the offenders in our program we needed to form a partnership with the Elders who were involved in our agency. This has included approaching Elders to assist us in developing and implementing traditional healing as a part of our program, including recommendations in assessments related to the potential benefits of healing in conjunction or in place of other forms of treatment or programming, seeking consultation in specific cases from Elders/healers, referring offenders to Elders/healers for counselling, sharing information with Elders/healers and developing joint co-operative treatment plans for mutual clients and including Elders/healers in meetings and supervision sessions to exchange information, share ideas and to learn from each other.

While this idea of partnership seems simple enough, it can be far from easy in a practical sense. In modifying our programming to include traditional healing we have encountered both support and opposition from mental health professionals as well as from individuals within the native community, healers and Elders. Those who have objected to this process believed that their way, either contemporary treatment or traditional healing, was the only path towards addressing the offenders needs and tended to discount other treatment/healing approaches. Those who supported the process appeared to see the individual in need of healing as the focus and priority rather than the means of healing. These people tended to acknowledge that a multidisciplinary approach may prove to be the most effective means of addressing risk and need and welcomed the challenges of attempting to co-ordinate the varying ideas and approaches between the contemporary and the traditional.

The End Point

In providing treatment services to offenders regardless of their race or the type of offence they have committed, the end point is attempting to provide them with treatment/healing that will allow them to identify, understand, resolve and cope with issues in their life that have contributed to criminal behaviour, to enhance their level of self-regard and to assist them to develop the desire and ability to manage their lives in functional ways. In order to accomplish these goals, the process will vary from one offender to another. In our program the objective of the clinicians and Elders are the same, to assist the offenders to effectively manage their risk and to become more healthy individuals who are capable of giving back to their communities rather than taking from and harming the community. Although our approaches differ, together we have found that a combined approach appears to be meeting the offenders needs and offers a more meaningful and thus effective treatment/healing experience.


Enter Treatment 36% 64% 53% 47%
Terminated 4% 3% 5% 0
Drop Out 19% 8% 0 0
Suspended 16% 2% 24% 0
Recidivism* Sexual 15% 0 0 0
Recidivism* Non-Sexual 4% 3%    
Recidivism** Sexual 0 0 >N/A N/A
Recidivism** Non-Sexual 4% 2% >N/A N/A

*while in treatment

**Post Treatment 9 months - 4 years post-treatment completion

DR. RALPH SERIN (Correctional Research and Development Sector, CSC)

Clinical And Empirical Issues In The Management Of Violent Offenders


This paper reviews the existing state of the art regarding the assessment and treatment of violent non-sexual offenders. Conceptual issues relating to the identification of treatment targets, measurement of treatment gain, and risk management will be considered. Specific concerns regarding methodological limitations of published studies, offender heterogeneity, and treatment responsivity will be discussed. Recommendations for future initiatives are presented to encourage advances for an increasing segment of the offender population. Finally, similarities with the literature on sexual offenders are discussed.

Clinical and Empirical Issues in the Assessment and Treatment of Violent Offenders

Advances in risk assessment knowledge and technology (Monahan & Steadman, 1994; Leis, Motiuk, & Ogloff, 1996; Webster, Eaves, Douglas, & Wintrop, 1995; Webster, Harris, Rice, Cormier, & Quinsey, 1994), to date have failed to be reflected in the published literature regarding the treatment of persistently violent or high risk non-sexual offenders (Serin, 1994).

Assessment Issues

Some general guidelines for the assessment of treatment targets are worth reviewing, as good assessment is critical to effective treatment. Assessment, here, refers to both the identification of treatment targets and measurement of treatment gain (Serin & Kuriychuk, 1994). In the area of anger and aggression assessment, there has been an over-reliance on self-report, notably with fairly transparent items on questionnaires. Often measures have been "borrowed" from non-offender settings, and adequate offender norms are either unpublished or absent. Further, there is a lack of standardisation regarding assessment protocols or test batteries, leading to a rather fragmented perspective for clinicians seeking assistance in this area. A further complication is the failure to guide assessment of offenders’ violence by a specific theoretical model (Blackburn, 1993; Novaco, 1994). For instance, it is frequently considered, at least from a review of assessment batteries, that all violent offenders are angry. While some authors have provided clinical descriptions of violent offenders which underscore their heterogeneity (Blackburn, 1993), the causal role of anger in violent behaviour remains a common view. This myth, however, has now been challenged for domestic abuse cases (Edleson & Tolman, 1992). Such a restrictive of anger being a sufficient antecedent to violence is likely to inhibit developments in the area of persistently violent offenders and is inconsistent with contemporary ideas about treatment responsivity (Andrews & Bonta, 1994).

Serin and Kuriychuk (1994) report data confirming the limitations of self-report measures for the exclusive identification of treatment targets and measurement of treatment gain. Comparing violent and non-violent offenders’ scores on a variety of self-report measures of anger, aggression, and hostility, yielded no group differences, despite high inter-correlations among measures. Even stringent definitions of history of violence (>50% total convictions for violence) yielded no differences on these measures, compared to non-violent offenders. These data certainly raise doubts regarding the construct validity of such measures for offender samples. In many respects, this situation is comparable to plethysmograph assessments in sexual offenders where evidence of deviant arousal is considered to represent identification of a treatment target. However, lack of deviant arousal phallometrically does obviate the need for treatment in sexual offenders.

There is a need for the development of a multi-method assessment strategy which is theoretically-driven and sensitive to potential malingering by offenders. Ideally, there should be convergence among the different assessment strategies, (e.g., self-report, behavioural ratings, interpersonal ratings, history, cognitions regarding violence, etc.). Improved assessment could yield a hierarchical model of intervention, whereby issues of risk and severity of treatment needs could inform differential treatment programming for violent offenders (Serin, 1994).

A related concern is the need for a careful assessment of treatment gain. Both knowledge and skill acquisition are important, and each should be evaluated pre-treatment, during the treatment program, and post-treatment. Further, each component of the treatment program should be assessed to determine where "booster" work is required. It is unclear whether offenders will be successful simply if they demonstrate gains, or whether a specific threshold is required before treatment gains generalise. This addresses the fundamental problem of whether a treatment program "worked", i.e., did the offender gain relevant knowledge and/or skills. A related issue is whether such gains generalise. In the case of violent and sexual offenders, this is most often reflected in evaluations of post-treatment recidivism rates for treated and untreated offenders. Recent studies with sexual offenders have emphasised treatment gain to be relevant to post-treatment outcome (Marques, Day, Nelson, & West, 1994), although there may be differences among the types of sexual offenders.

Offender Heterogeneity

Blackburn (1993) has appropriately noted that violent offenders are a diverse group, for whom existing diagnostic nosology may be limited, particularly regarding personality dimensions. A brief review of factors purported to relate to adult violence should serve to emphasise that violent offenders are heterogeneous. Common factors reported in the literature and considered by clinicians, for both assessment and treatment include: i) developmental history (age of onset, frequency and variety of problem behaviour); ii) history of violent behaviour (criminal, self-reported, and self-inflicted); iii) violent ideation and fantasies; iv) weapon interest, access, and proficiency; v) victim affiliation (relationship to, selection, access); vi) affective versus predatory violence; vii) motivation for violence; and, viii) role of substance abuse. It is obvious these factors are also relevant to sexually violent offenders and most have been incorporated into contemporary typology strategies (Knight, Prentky, & Cerce, 1995 ).

The extent to which offenders differ on such factors raise specific questions regarding intervention. Clearly, not all factors will be relevant for all violent offenders, yet in the development of a treatment program, some consideration of how to meet offenders’ diverse needs is required..

Theoretical Model

One strategy is to use a conceptual model of offenders’ violence to guide intervention. For instance, Serin and Kuriychuk (1994) provide a model which highlights the synergistic interaction between cognitive schema of aggressive beliefs, impulsivity, arousal level, and cognitive processing deficits. They suggest that persistently violent offenders are predisposed towards aggressive explanations of events and take pre-emptive action (Novaco, 1994), without fully considering the proximal cues (Dodge & Newman, 1981; Milich & Dodge, 1984). These sequences of events are often the result of automatic processing (Newman & Wallace, 1993), but impulsivity or poor self-regulation is an insufficient explanation of persistent violence without the presence of hostile schema. This view that the world is a hostile environment leads to egocentricity and righteous explanations of violence. Such a model, then, identifies specific intervention targets.

Treatment Issues

After excluding sexual offenders and domestic abuse, there are surprisingly few published studies on the treatment of violence. There are even fewer for offender populations (Serin, 1994). In reviewing the literature it becomes clear that intervention identified by different labels, (i.e., social problem-solving, anger control, impulse control, aggression control), are similar when considering descriptions of each

program. The majority of programs imply arousal reduction and problem-solving as central components, consistent with models which suggest violence relates to high arousal or anger, and poor problem-solving or impulse control. It remains an empirical question whether this is sufficient for persistently violent offenders whose use of violence is often characterised by a marked lack of emotion. It appears that many programs approach treatment from the perspective of skills acquisition. That is, the underlying assumption is that if offenders increase specific skills, their use of violence will diminish. For persistently violent offenders, however, treatment might be better viewed as a process, with engagement, treatment compliance and skills improvement necessary. Similarly, the notion that treatment is a cure is inconsistent with risk management strategies (Blackburn, 1993, Quinsey & Walker, 1992). For offenders, then, treatment might be best viewed as an additional strategy to assess and manage risk.

At present, there is little evidence that treatment for violent offenders is either prescriptive or hierarchical. Differentiated programming would match offenders to specific treatment modules where pre-treatment assessment has indicated a deficit, not simply put all offenders through all facets of a particular program. Also, consistent with the responsivity issue, higher risk offenders should receive more intensive programming than lower risk offenders. Again, such an approach appears absent in the published literature.

A related concern is that of subject selection for programs. Typically, once a program is offered, the rule of thumb is to require all offenders to take it. Not only does this increase offender resistance, but it is an inefficient use of limited treatment spaces. The development of a standardised assessment protocol and provision of programs of different intensity levels may help address this concern.

Treatment Approaches

The most popular treatment strategies incorporate a cognitive-behavioural approach. Generally, the treatment efforts can be conceptually divided into self-regulation and cognitive processing strategies. Self-regulation strategies emphasise arousal reduction by considering anger cues, self-statements about anger, parameters of anger (duration, intensity, frequency, and outcome), self-instructional coping, and relaxation training. Communication or social skills and assertion skills training also tend to be incorporated in self-regulation strategies. It is unclear whether relaxation training is superior to skills acquisition. Also, there is dispute whether these approaches are appropriate for domestic abuse cases (Edleson & Tolman, 1992).

Cognitive processing strategies consider social-information processing deficits in violent individuals. They emphasise the role of aggressive beliefs in eliciting and maintaining violent behaviour, with problem-solving and biased thinking skills being targeted in treatment (Slaby & Guerra, 1989). Despite demonstrated treatment gains, (i.e., increased skills in problem-solving reduces support for aggressive beliefs in juvenile offenders), such an approach has not impacted on recidivism rates (Guerra & Slaby, 1990).

There is consensus in the literature for a cognitive-behavioural approach presenting the following "typical" modules: anger definition; arousal reduction; aggressive beliefs; impulse control; problem-solving; empathy or moral reasoning; communication skills; and, assertiveness training. Relapse prevention is beginning to be applied to violent offenders (Prisgrove, 1993). This is important in that it permits a case-specific approach and underscores offender heterogeneity, yet the relevance of the offence cycle for violent non-sexual offenders is unclear.


The level of sophistication of published studies on the treatment of violent offenders is disappointingly low. For instance, random assignment and use of control groups is exceedingly rare. Also, assessment tends to be in a single domain, e.g., psychological testing. Therefore, there is an over-reliance on self-report measures in assessing treatment need and treatment gain. Further, there is a lack of standardisation for assessment strategies. Multiple outcome measures are rarely considered. Post-treatment assessment occurs, but lack of follow-up is the norm. Rarely do authors describe the theoretical model(s) underlying their intervention, so it is not always readily clear why specific components of the program were included. Lastly, there is virtually no investigation of differential treatment effects, (e.g., what kind of offenders responded best, and to what aspects of the intervention).

Treatment Responsivity

Treatment responsivity is the consideration that offenders need to be matched to particular aspects of a treatment program, including program content and therapists’ characteristics. In this way, treatment is prescriptive, not generic. There is now evidence that prescriptive intervention can enhance efficacy (Andrews & Bonta, 1994). Related to the issue of matching is ensuring that the program content is presented at a level the offender can understand, and delivered in a manner most likely to engage the offender. Therapist characteristics, then, become integral to responsivity factors. For violent offenders, the role of duress or coercion in treatment is unlikely to be avoided, yet skilled therapists are able to circumvent such difficulties (Miller & Rollnick, 1991).

Recent studies in the treatment of psychopaths (Ogloff, Wong, & Greenwood, 1990; Rice, Harris, & Cormier, 1992) highlight that high risk or persistently violent offenders are resistant towards treatment. Similar findings have been reported for sexual offenders (Marques et al, 1994) and in the area of addictions (Miller & Rollnick, 1991). Therapists working with these clients specifically target denial, minimisation and engagement issues. That this is a prerequisite to the actual treatment program, should not be ignored by those treating violent non-sexual offenders. On a broader level, this literature suggests that not all therapists are suited to working with violent offenders and unique issues will be encountered which might interfere with successful therapy (Meloy, 1995).

Future Directions

In order for this area to progress, improvements in the methodology of treatment programs is a given. This includes: the need for a standardised, multi-method assessment; application of theoretical model(s); consideration of treatment responsivity factors; investigation of differential treatment effects; the matching of offenders’ level of risk and treatment needs to a particular program; and investigation of multiple outcome measures. Also of assistance would be the completion of an inventory of existing CSC programs, preferred assessment instruments, and suggested best practices. Consolidated, multi-site efforts with a parallel research agenda would significantly enhance intervention initiatives for this population. What is now required is the evaluation of existing programs; the development of assessment technology in the identification of treatment needs and the measurement of treatment gain; and, the careful delivery and follow-up of theoretically-based treatment.

Comparison to Sex Offender Treatment

Similar to violent non-sexual offenders, sexual offenders are a heterogeneous group. This diversity has been well established because victim information yields types of sexual offenders, (e.g., rapists, extra-familial child molesters, incest offenders), yet it is rarely considered in the admission to a treatment program and only minimally in terms of program content. Surprisingly, programs for rapists are not unlike those for incest or child molesters, and there may be clinical reasons to support mixed selection criteria for program involvement. The assessment of sexual offenders tends to be multi-method by the inclusion of phallometric data, yet not all sites have access to this technology. Similar to the treatment of violent non-sexual offenders, there appears to be an over-reliance of self-report in the measurement of treatment gain. Although there have been recent attempts to match treatment intensity to risk level, sometimes this is not recognised by other parties in the criminal justice field. Despite better longitudinal outcome studies for sexual offenders, the debate regarding treatment efficacy remains (Marshall & Pithers, 1994; Quinsey, Harris, Rice, & Lalumière, 1993). Central to this debate is the degree of methodological rigor in clinical treatment studies, the impact of such restrictions in the real world of service delivery, and the kinds of conclusions reasonably inferred from less than ideal studies (Quinsey et al., 1993).

The importance of relapse prevention in the assessment and treatment of sexual offenders appears to have been readily embraced, yet empirical support for its exclusive application appears only modest. Its attraction as an overall framework, however, should not inhibit the development of alternate views, especially for those offenders for whom a crime cycle is neither apparent, nor central to offending.


The issues driving the assessment and treatment of sexual offenders and persistently violent non-sexual offenders appear more congruent than disparate. Similar assessment concerns, treatment targets, clinical difficulties, and methodological goals suggest much is to be gained by informed integration of these areas. Advances in one area, therefore, should be evaluated for potential application to the other. While each area has unique concerns and specific issues to consider, their commonalties should not be ignored.


Andrews, D.A. and Bonta, J. (1994). The psychology of criminal conduct. Cincinnati: Anderson Publishing.

Blackburn, R. (1993). The psychology of criminal conduct. Chichester, England: Wiley & Sons Ltd.

Dodge, K.A. & Newman, J.P. (1981). Biased decision making processes in aggressive boys. Journal of Abnormal Psychology, 90, 375- 379.

Guerra, N.G. & Slaby, R.G. (1990). Cognitive mediators of aggression in adolescent offenders: 2. Intervention. Developmental Psychology, 26, 269-277.

Edleson, J.L. & Tolman, R.M. (1992). Intervention for men who batter: An ecological approach, Interpersonal Violence: The Practice series, Newbury Park, California: Sage.

Knight, R.A., Prentky, R.A., Cerce, P.D. (1994). The development, reliability, and validity of an inventory for the multidimensional assessment of sexual aggression, Criminal Justice and Behaviour , 121, 72-94.

Leis, T.A., Motiuk, L.L., & Ogloff, J.R. (1996). Forensic psychology: Policy and practice in corrections. Corporate Research and Development: Correctional Service of Canada.

Marques, J.K., Day, D.M., Nelson, C., & West, M.A. (1994). Effects of cognitive-behavioural treatment on sex offender recidivism: Preliminary results of a longitudinal study. Criminal Justice and Behaviour, 21, 28-54.

Marshall, W.L. & Pithers, W.D. (1994). A reconsideration of treatment outcome with sex offenders, Criminal Justice and Behaviour, 21, 10-27.

Meloy, R. (1995). Treatment of antisocial personality disorder. In G. Gabbard (Ed.), Treatments of psychiatric disorders: The DSM-IV Edition (pp. 2273-2290). Washington, D.C. : American Psychiatric Press, Inc.

Milich, R. & Dodge, K.A. (1984). Social information processing in child psychiatric populations. Journal of Abnormal Child Psychology, 12, 471- 489.

Miller, W.R & Rollnick, S. (1991). Motivational Interviewing. New York: Guilford Press.

Monahan, J. & Steadman, H.J. (1994). Violence and mental disorder: Developments in risk assessment. University of Chicago Press: Chicago.

Ogloff, J.P.R., Wong, S., & Greenwood, A. (1990). Treating criminal psychopaths in a therapeutic community program. Behavioural Sciences and the Law, 8, 81-90.

Newman, J.P. & Wallace, J.F. (1993). Psychopathy and cognition. In K.S. Dobson and P.C. Kendall (Eds.), Psychopathology and cognition (pp. 293-349). Orlando, FL: Academic Press, Inc.

Novaco, R.W. (1994). Anger as a risk factor for violence among the mentally disordered. In J. Monahan and H.J. Steadman (Eds.), Violence and mental disorder: Developments in risk assessment. Chicago:University of Chicago Press.

Prisgrove, P. (1993). A relapse prevention approach to reducing aggressive behaviour. In S.A. Gerrull and W. Lucas (Eds.), Serious violent offenders: Sentencing , psychiatry and law reform. Canberra: Australian Institute of Criminology.

Quinsey, V.L., & Walker, W.D. (1992). Dealing with dangerousness: Community Risk Management Strategies with violent offenders. In D.V. Peters, R.J. McMahon, & V.L. Quinsey (Eds.), Aggression and violence throughout the lifespan (pp. 244-262). Newbury Park, CA: Sage.

Quinsey, V.L., Harris, G.T., Rice, M.E., & Lalumière, M.L. (1993). Assessing treatment efficacy in outcome studies of sex offenders. Journal of Interpersonal Violence, 8, 512-523.

Serin, R.C. (1994). Treating violent offenders: A review of current practices (Report No. R-38). Correctional Research and Development, Correctional Service of Canada: Ottawa.

Serin, R.C. & Kuriychuk, M. (1994). Social and Cognitive Processing Deficits in Violent Offenders: Implications for Treatment, International Journal of Law and Psychiatry, 17, 431- 441.

Slaby, R. G. and Guerra, N. G. (1988). Cognitive mediators of aggression in adolescent offenders: 1. Assessment. Developmental Psychology, 24, 580-588.

Rice, M.E., Harris, G.T., & Cormier, C. (1992). Evaluation of a maximum security therapeutic community for psychopaths and other mentally disordered offenders. Law and Human Behaviour, 16, 399-412.

Webster, C.D., Harris, G.T., Rice, M.E., Cormier, C., & Quinsey, V.L. (1994). The violence prediction scheme: Assessing dangerousness in high risk men. Centre of Criminology, University of Toronto.

Webster, C.D., Eaves, D., Douglas, K., & Wintrup, A. (1995). The HCR-20 Scheme: The assessment of dangerousness and risk. Simon Fraser University and Forensic Psychiatric Services Commission of British Columbia: Simon Fraser University Press.

DOUGLAS P. BOER, CLAUDINE M. GAUTHIER, and DAVID R. WATSON (Regional Health Centre (Pacific),CSC)

JOHN DORWARD and DAVID J. C. KOLTON - Simon Fraser University

The Assessment and Treatment of Intellectually Disabled Sex Offenders: The Regional Psychiatric Centre (Pacific) "Northstar" Treatment and Relapse Prevention Program

Intellectually disabled sex offenders (IDSOs) are offenders who, in a fashion similar to mentally ill persons, often "fall between the cracks" of social services and the various health systems. They cycle through hospitals and community agencies, and become involved in the criminal justice system by virtue of their sexually inappropriate behaviour. Once incarcerated, IDSOs may be excluded from certain treatment programs as a result of being intellectually limited, illiterate, impulsive, or having inadequate social skills. If they were deemed unsuitable for regular sex offender programming, or if they were "main-streamed" in a regular sex offender program, one would assume that their likelihood of recidivism was not reduced since their unique programming needs were not met prior to release (cf., Langevin & Pope, 1993). The "Northstar" program is aimed at treating federally incarcerated IDSOs and is designed to meet a significant proportion of these needs through the use of a variety of treatment modalities ranging from psychoeducational modules, self-directed therapies, arousal re-conditioning, adjunctive therapies, and individual treatment for specific problem areas. The present paper is a description of the research basis behind the assessment and treatment procedures employed in this program. The development of effective treatment programs for IDSOs may aid in the design of future programs for mentally ill sex offenders, another special needs population within corrections that has yet to be effectively targeted.

How Are IDSOs Different From Non-IDSOs?

The majority of IDSOs seen in specialised programs do not fall in the profound or severely retarded ranges of intellectual functioning, but rather, fall within the borderline to mildly retarded ranges of intellectually functioning (e.g., Murphy, Coleman & Haynes, 1983). Despite the implication of the label, not all "IDSOs" are intellectually disabled according to intelligence tests. For example, several members of the current Northstar program have low average IQs, but have been included in the program because of significant deficits in social functioning or because of their inability to gain insight into their behavioural problems. In general, however, IDSOs are characterised by lower intellectual functioning than non-IDSOs.

As well, Haaven and his associates reported that 74% of their IDSO clients were diagnosed as having evidence of a brain injury (Haaven, Little & Petre-Miller, 1990). They contrasted this finding to non-IDSO research findings of up to (but not greater than) 50% brain injured. In addition to the different proportions of organicity, this does not reveal the fact that brain injured IDSOs are functionally more impaired than brain injured non-IDSOs because the former group may have other learning disabilities which are further complicated by brain injury. Brain injury has been reported to cause sexual disinhibition, hypersexuality, change in sexual preference, poor abstract reasoning, inability to sequence events, reduced memory recall, aggressive behaviours, explosiveness, and anxiety disorders (Haaven et al, 1990).

There are other areas which distinguish IDSOs and non-IDSOs, some of which suggest areas of increased risk to an IDSO's propensity to re-offend. For example, although IDSOs and non-IDSOs are not different in terms of offence types (e.g., child molestation or rape), IDSOs tend to be more opportunistic and impulsive in their everyday behaviour and offence behaviour (Lane, 1991). In contrast to non-IDSOs, IDSOs tend to have fewer victims, not establish close relationships with their victims (choosing victims who are acquaintances or are people encountered during daily routines, as opposed to first degree relatives), and not discriminate according to age, gender, or appearance of their victims. Thus, it is more difficult to gauge IDSO's predatory behaviour, since it is often unclear what "victim-types" are high risk for them (Haaven et al., 1990).

In addition, Haaven and his associates suggest that IDSOs tend to use instrumental violence (i.e., use of sufficient violence or the threat of violence to gain compliance of the victim), as opposed to expressive violence (i.e., inflicting injury on the victim as part of their arousal pattern), in their offences more often than non-IDSOs. This may be because IDSOs are less able to verbally manipulate their victims into compliance and instead routinely use physical force or threat. Finally, IDSOs tend to victimise individuals who are smaller, less able to verbally protest (more passive), and less able to defend themselves (Lane, 1991).

There is some evidence that because of poor social skills (and subsequent lack of intimate relationships and an inability to have their emotional needs met appropriately), IDSOs are often lonely men who spend an inordinate amount of time fantasising and masturbating in contrast to non-IDSOs. IDSOs tend to have little sense of self-worth due to a history of being ridiculed by parents and peers during childhood and adolescence (Schoen & Hoover, 1990). They perceive themselves as victims, are unable to understand the needs of others, and tend to think their only mistake was getting caught (Lane, 1991). IDSOs also tend to lack assertion skills and acquiesce with the demands of their peers (e.g., they may commit illegal acts to gain approval). Also, Lane (1991) noted that a significant proportion of these offenders have been sexually victimised themselves and the families of these offenders often minimise the severity of the offences and the offender's risk to others. These factors may compound to reinforce the offender's view of himself as a "victim" and his sentence as being excessively harsh.

From a multiple diagnoses point of view, IDSOs and non-IDSOs appear more similar than different. Non-IDSOs also suffer from a range of affective disorders including anxiety (Meyer-Williams & Finkelhor, 1990) and depression (Langevin, Handy, Day, & Russon, 1985). Alcohol and drug abuse problems are present among non-IDSOs (Meyer-Williams & Finkelhor, 1990), and the problem of sexual deviancy among non-IDSOs is well-documented (Blackburn, 1993). However, despite these similarities, IDSOs present with a broader constellation of problems and treatment needs than non-IDSOs because of their intellectual and social limitations. The complicating factor of low intellectual functioning, and how this factor interacts with the above factors, exacerbates the IDSO client's problems. In addition, when the actuarial risk factors and elevation in risk due to the interplay of the above problems are considered, IDSOs appear to be a category of offenders which fulfil the dictum of focusing treatment on high risk, high needs cases (Andrews, Bonta, & Hoge, 1990).

Assessment of IDSOs

The assessment of IDSOs is necessarily restricted by their limited cognitive abilities and low levels of literacy. However, many of the comprehensive self-report measures that are generally used for assessment are inappropriate. For example, Haaven (1994) suggested that the MMPI-2 or similar tests are not useful for the assessment of IDSOs because it is likely that such clients misunderstand a large proportion of the items. At a recent workshop, Ben-Porath (1994) stated that this concern would only be valid if the VRIN score (a validity index which can help rule out the possibility that a client's score on another validity index reflected random responding or confusion) was not elevated. However, he agreed that it was possible to have a VRIN within normal limits and the client still misunderstand some proportion of the items. These concerns suggest limited applicability of standard personality tests such as the MMPI-2 for the assessment of IDSOs.

Nonetheless, it is still possible to do a comprehensive evaluation of the IDSO (Clare, 1993). Coleman (1994), Caparulo (1991), and Murphy et al. (1983) all recommend a standard series of guidelines for the assessment of IDSOs. First, informed consent must be obtained with an explanation of the costs and benefits of assessment and treatment to the individual, and limits of confidentiality. Coleman (1994) suggested having several "consent" discussions that are documented thoroughly. Second, a clinical interview to judge competency to complete assessment instruments and mental status should be performed. Third, a comprehensive psychosexual and family history should be completed. Coleman (1994) advised beginning with the least threatening sexual questions first, moving gradually into more "in-depth" questions. Fourth, phallometric testing is recommended. Coleman (1994) advised that some caution be exercised in interpreting these results because of a lack of normative data on IDSOs, however, she believed that the pattern of results obtained can be valuable both diagnostically and for treatment.

A thorough assessment of the IDSO may involve many areas of psychometric evaluation. Coleman (1994), Caparulo (1991), and the present authors recommend the following areas for psychometric testing: social skills, sex knowledge and attitudes (e.g., the Behavioural Checklist, Assault Knowledge Inventory, Deragatis Sexual Functioning Inventory, Socio-Sexual Knowledge and Attitude Test, Adult Self-Expression Scale), cognitive distortions associated with offending (the Multiphasic Sex Inventory, the Pedophile Cognition Scale, the Abel-Becker Cognition Scale, Abel-Becker Sexual Deviance Card Sort), depression (the Beck Depression Inventory), intellectual functioning (the WAIS-R, Shipley Institute of Living Scale, Raven's test of Intellectual Functioning), educational achievement (e.g., the Wide Range Achievement Test - Revised), and socially desirable responding (e.g., Balanced Inventory of Desirable Responding, the Marlowe-Crowne). [Please note that the references for the tests listed in this paragraph are given in the reference section.] In addition, an organicity screen or complete neuropsychological testing may be completed as deemed necessary. These paper and pencil tests represent a wide range of possible testing options. It is not necessary to assess all clients with every instrument, rather, instruments should be chosen according to the referral question and the clients level of intellectual functioning. As well, one must be pragmatic in the selection of assessment instruments with this population, since various difficulties can arise given the IDSO's personal history of problems in school (related to intellectual functioning deficits), lack of self-esteem, possible lack of frustration tolerance, and impulsive behaviour. In addition, if a client is illiterate, there are some audio-taped tests available or the items can be read to the client by a staff member. Finally, actuarial measures should be administered to determine the IDSO's risk for re-offence. The PCL-R (Hare, 1991) and the Violence Prediction Scheme (Webster, Harris, Rice, Cormier, & Quinsey, 1994) would appear to be the "state of the art" risk assessment instruments for this population (e.g., borderline intellectual functioning). As noted above, IDSOs do appear to have an exacerbated risk level due to factors uniquely associated with this population. Thus, risk assessments should address how such factors could affect risk level. Similarly, dynamic risk factors, such as level of denial and minimisation, must also be addressed in the risk assessment.Treatment of IDSOs Many authors report that the most common and most effective treatment for sex offenders is based on cognitive-behavioural methods (e.g., Marshall & Barrett, 1990). This also appears to be true for IDSOs. Haaven and his associates (1990) included cognitive restructuring and arousal re-conditioning (particularly minimal arousal conditioning and covert sensitisation) as "unexpectedly effective interventions" with IDSOs. IDSO's attitudes and cognitive distortions that are supportive of offending have been reported to be similar (but more concrete) to those of non-IDSOs, consequently, a simplified version of cognitive restructuring can be useful in treating these clients (Haaven et al., 1990; Murphy et al., 1983). Murphy and his colleagues (1983) suggested that the use of covert sensitisation as originally designed by Maletzky (1980) would be hampered by IDSO's limited imagery ability (i.e., concrete thinking). Thus, covert sensitisation was modified by Haaven et al. (1990) and further refined by Petre-Miller (1994) who found this technique effective with this client population. Haaven and his colleagues (1990) reported that individual counselling was moderately useful when it was time-limited and had specific goals (e.g., dealing with a patient's own history of abuse). They also found medical interventions (e.g., Depo-Provera) to be of limited usefulness, although other findings suggest that anti-androgens do help in decreasing deviant sexual arousal in sex offenders (Bradford, 1990). Other cognitive-behavioural techniques that can be applied with IDSOs include: self-monitoring techniques, such as behaviour contracting, daily journal writing, and self-charting (refer to Watson & Tharp, 1989). IDSOs, like most sex offenders, lack a variety of positive socialisation skills, such as those needed to communicate effectively or form healthy relationships. Murphy and his associates (1983) suggested that IDSOs may be more deficient in "heterosocial" skills than non-IDSOs. Thus, social skills need to be incorporated into all components of the treatment program to increase social competency and self-esteem. Particular skill deficits are addressed with specific training, for example, relationship skills training coupled with communication skills training is offered to help offenders learn the skills needed to increase their chances of meeting their intimacy needs appropriately. Seidman, Marshall, Hudson, & Robertson (1994) indicated that sex offenders in general, and we would suggest that IDSOs in particular, are more lonely and deficient in intimacy than non-sex offenders and community controls and improvements in this area would increase their likelihood of remaining offence-free.

Murphy and his associates (1983) indicated that it is also important to provide sex education for IDSOs. Such education can be aimed at a very basic level, using concrete examples, simple terminology, and a focus on pragmatic issues (e.g., condom use). IDSOs also have problems with anger/emotional management. They often have histories of acting impulsively, low self-esteem, and a paucity of assertion skills. Learning to express emotions (including anger) appropriately helps these patients interact with others less impulsively and more assertively, resulting in increased self-esteem and social competence. Other skill areas that can be addressed include reasoning and problem-solving. These may be treated in a group format (i.e., didactic) or in experiential settings (e.g., horticultural therapy). Expressive therapies, such as art therapy, have also been recommended to "help residents communicate issues and feelings that are normally threatening for them to share (Haaven et al., 1990, p. 21)." Relapse Prevention is the most widely promoted model for use with sex offenders in preventing re-offence. Each offender uses the model to delineate their criminal behaviour pattern or crime cycle. Strategies are then employed to intervene at points of risk throughout the crime cycle, decreasing the likelihood of re-offence (for an excellent reference see Laws, 1989). Relapse Prevention has been modified for use with IDSOs. Haaven et al. (1990) suggested that Relapse Prevention programming should be delivered at periods during treatment, post-treatment, parole planning, and community entry. They believe such continuity of care to be inversely related to recidivism, as this is widely held as a reliable strategy for ensuring the maintenance of treatment gains and a non-offending lifestyle. Other treatment areas applicable to IDSOs include a stronger focus on personality disorder treatment and the area of substance abuse treatment (Haaven et al., 1990). Substance abuse, added to impulsivity, certainly appears to be a problematic combination for some IDSOs and the issue of personality disorders in IDSOs may be easily overlooked given all their other presenting problems. There are also certain treatment methods that have been found to be ineffective with IDSOs. For example, masturbatory satiation (Marshall, 1979) was found to be inappropriate because some clients appeared not to "satiate", while others appeared unable to do the required fantasy switching involved in the technique (Haaven et al., 1990). They also found fantasy and urge logs to be unsuitable because of inaccuracy and an apparent increase in urges during tracking by some patients. Finally, empathy training as practised by Haaven and his colleagues (1990) was reportedly ineffective. These authors noted that the techniques of attempting to "teach empathy training by showing videotapes of former victims describing the abuse-related difficulties they experienced as children and adults" seemed counterproductive since some of the patients became aroused by the tapes. Recent attempts at empathy training with IDSOs (e.g., Gauthier, 1994; Zella, 1994) appear more promising, but further evaluation is required before its applicability to IDSOs can be justified.The Northstar Treatment Program There are a wide variety of treatment modalities used in the Northstar program to address the treatment and criminogenic needs that IDSOs present. All program components are designed to effectively manage these needs and are based on the literature regarding this population. Treatment modalities shown by research to be ineffective with these clients are not stressed. A multidisciplinary team approach is used to enable the delivery of consistent messages from a variety of program deliverers in a variety of modalities. This has been shown to be the most effective way in which these clients learn to change their behaviour. The program contains individual sessions, behavioural therapies, medical interventions, therapy modules delivered in a group format (see Figure 1.), as well as adjunctive therapies. Currently, the program is comprised of three 3 month trimesters.

Figure 1. Group Components By Trimester

Trimester 1 Components Trimester 2 Components Trimester 3 Components
  • Sex Education
  • Goal Review
  • Communications
  • Personal Concerns
  • Anger Management
  • Disclosure
  • Sexual Deviancy
  • Identifying Feelings
  • Relationship Skills
  • Goal Review
  • Communications
  • Personal Concerns
  • Anger Management
  • Crime Cycle
  • Managing Deviancy
  • Managing Feelings
  • Substance Abuse
  • Goal Review
  • Communications
  • Personal Concerns
  • Problem Solving
  • Relapse Prevention
  • Victim Empathy
  • Living Without Violence

In general, our group therapy modules are based on Social Learning theory and follow a logical, hierarchical sequence with the goal of learning new, more rewarding and adaptive behaviours. For example, Anger Management begins with an education phase regarding the nature of anger, followed by a skills acquisition phase concerned with learning new ways of dealing with anger via analysis of current situations and discussion of appropriate interventions (e.g., anger reducers, cognitive restructuring, and problem-solving). The final application phase assists clients in applying the techniques to their specific pre-incarceration experiences. The Sexual Deviance, Feelings, Victim Empathy, Relationship Skills, and Sex Education modules follow a similar sequence; basic information is first learned and then applied to important past and present aspects of their lives. The Disclosure, Crime Cycle, and Relapse Prevention groups follow each other in a logical sequence. Due to the cognitive limitations of these clients, conceptual jargon is kept to a minimum (e.g., we do not discuss "seemingly unimportant decisions", the "abstinence violation effect", or "cognitive distortions"; rather, we use "thinking mistakes", the "poor me, what the heck, I deserve it effect", and "excuses", respectively). The Disclosure group provides clients with a non-confrontative opportunity to describe what was involved in the present offence from their viewpoint. A set of standardised questions are used to ascertain differences between the client's and the "official" version. In this manner, each offender's thoughts and feelings (and minimisation’s) can be determined. Such information is invaluable in formulating the client's crime cycle. The Disclosure group is followed by the Crime Cycle group in which the risk factors and the associated cognitive-behavioural patterns that typify the offender's actions surrounding his crimes are determined in a manner understandable to the client. Finally, the Relapse Prevention group is designed to help individuals cope effectively with high risk factors and to identify (and respond to) early warning signals that indicate that high risk factors are imminent. There are a number of therapy modules that continue throughout the program. The Personal Concerns group is a forum for learning and applying problem-solving skills on the ward that will help the clients cope better in their daily lives. The Communication Skills group is a systematic, structured educational program for teaching clients how to communicate effectively with a wide range of people with whom they encounter in everyday life. The skills learned in this group include how to converse, confront, and provide feedback in an assertive manner. Finally, the Goal Review group helps clients formulate reasonable, attainable goals, in time frames that provide the opportunity for success. Goals are reviewed on a weekly basis for feedback purposes by staff and group members. In addition to therapy modules, a wide variety of other therapeutic modalities are used in the Northstar program to address treatment needs. These modalities include: substance abuse and family violence programs (delivered in a group format), individual issue-focused sessions (to reinforce information from modules), behavioural contracts (for specific deficits or problematic behaviours), self-monitoring (e.g., masturbation records), arousal re-conditioning and sex-drive reducing medication (e.g., for deviant arousal), and a number of "adjunctive therapies" including: horticulture, art, school, and recreation (for upgrading, skill-development, and increasing their repertoire of appropriate behaviours). In conclusion, the Northstar program rests on a number of premises. One is that no module or therapy is delivered to these clients that does not have a firm research foundation as a rationale for its delivery. Another is that concepts must be kept simple, taught thoroughly, practised often, and reinforced in a consistent manner. This is done through a variety of therapeutic modalities and by a variety of therapists. Also, therapeutic relationships are well-managed since these clients are dependent and demanding. Finally, community follow-up personnel need to be fully informed of the treatment needs and gains of clients to ensure continued progress. By following these guidelines, and thereby perhaps meeting more of the IDSOs treatment needs more effectively, hopefully more of these clients may be classified as treatment "successes" rather than "failures". Tracking systems for client follow-up are being implemented with all persons admitted to the Northstar program to ascertain program effectiveness.Some Closing Practical Concerns About Working with the IDSO IDSOs have marked deficits in social functioning and may behave impulsively inside or outside of treatment programs. They often do not discriminate among victims and have been noted to "act-out" sexually (e.g., masturbate in public) with little regard to the effect of their behaviour on others. Single bunking helps to give patients a sense of security/privacy from others, reducing the chance of victimisation from someone more predatory, provides a place to retreat from the confusion he may be experiencing, and gives him a chance to make choices about how to spent his own time (e.g., to do homework when he wants to). In addition, the optimal length of treatment programs needs to be examined in working with IDSOs. Common sense would indicate that the elevated treatment needs and slower rate of comprehension of IDSOs would necessitate prolonged programs with flexible program end dates. These represent some of the more salient issues to attend to in designing effective treatment programs for this patient population.


Abel, G.G., Gore, D.K., Holland, C.L., Camp, N., Becker, J.V., & Rathner, J. (1989). The measurement of the cognitive distortions of child molesters. Annals of Sex Research, 2, 135-153.

Andrews, D.A., Bonta, J., & Hoge, R.D. (1990). Classification for effective rehabilitation: Rediscovering psychology. Criminal Justice and Behaviour, 17, 19-52.

Barlow, D.H., Abel, G.G., Blanchard, E.B., Bristow, A.R., Young, L.D. (1977). A heterosocial skill checklist for males. Behaviour Therapy, 8, 229-239.

Beck, A.T., & Steer, R.A. (1987). Manual for the revised Beck Depression Inventory. San Antonio, TX: Psychological Corporation.

Ben-Porath, Y. S. (1994, October). MMPI-2. Workshop presented at the meetings of the British Columbia Psychological Association, Vancouver, B.C.

Blackburn, R. (1993) The psychology of criminal conduct: Theory, research, and practice. New York: John Wiley & Sons.

Bourque, P., & Ladouceur, R. (1978). Validation of the Adult Self-Expression Scale of Gay, Hollandsworth and Gallassi. Canadian Journal of Behavioural Science, 10(4), 351-355.

Bradford, J.M.W. (1990). The antiandrogen and hormonal treatment of sex offenders. In W.L. Marshall, D.R. Laws, & H.E. Barbaree (Eds.), Handbook of sexual assault: Issues theories, and treatment of the offender. New York: Plenum Press.

Caparulo, F. (1991). Identifying the developmentally disabled sex offenders. Sexuality and Disability, 9(4), 311-321.

Clare, I.C.H. (1993). Issues in the assessment and treatment of male sex offenders with mild learning disabilities. Sexual and Marital Therapy, 8, 167-180.

Coleman, E.M. (1994, November). Assessment and treatment of intellectually disabled sex offenders. Paper presented at the meetings of the Association for the Treatment of Sexual Abusers, San Francisco, CA.

Demetral, G.D. (1994). Diagrammatic assessment of ecological integration of sex offenders with mental retardation in community residential facilities. Mental Retardation, 32(2), 141-145.

Derogatis, L.R. & Melisaratos, N. (1979). The DSFI: A multidimensional measure of sexual functioning. Journal of Sex & Marital Therapy, 5(3), 244-281.

Gauthier, C.G. (1994, unpublished manuscript). Victim empathy training for intellectually disabled sex offenders.Griffiths, D., Hingsburger, D., & Christian, R. (1985). Treating developmentally handicapped sexual offenders: The York Management Services treatment program. Psychiatric Aspects of Mental Retardation Reviews, 4, 49-52.

Haaven, J., Little, R., & Petre-Miller, D. (1990). Treating intellectually disabled sex offenders: A model residential program. Orwell, VT: The Safer Society Press.Hare, R.D. (1991). Manual for the Hare Psychopathy Checklist - Revised. Toronto, Ontario: Multi-Health Systems.

Jastak, S., & Wilkinson, G.S. (1984). Wide Range Achievement Test - Revised Edition. Wilmington, Delaware: Jastak Assessment Systems.Kalichman, S.C., & Henderson, M.C. (1992). Psychometric properties of the Multiphasic Sex Inventory in assessing sex offenders. Criminal Justice & Behaviour, 19(4), 384-396.

Lane, S.L. (1991). Special offender populations. In G.D. Ryan & S.L. Lane (Eds.), Juvenile sexual offending: Causes and consequences (pp. 229-332). Lexington: Lexington Press.

Langevin, R., Handy, L., Day, D., & Russon, A. (1985). Are incestuous fathers pedophilic, aggressive and alcoholic? In R. Langevin (Ed.), Erotic preference, gender identity and aggression (pp. 161-180). Hillsdale, NJ: Erlbaum.

Langevin, R. & Pope, S. (1993). Working with learning disabled sex offenders. Annals of Sex Research, 6, 149-160.Laws, D.R. (Ed.). (1989). Relapse prevention with sex offenders. New York: Guildford Press.Marshall, W.L. (1979). Satiation therapy: A procedure for reducing deviant sexual arousal. Journal of Applied Behaviour Analysis, 12, 377-389.

Marshall, W.L. & Barrett, S. (1990). Criminal neglect: Why sex offenders go free. Toronto: Doubleday Books.

Meyer-Williams, L. & Finkelhor, D. (1990). The Characteristics of incestuous fathers: A review of recent studies. In W.L. Marshall, D.R. Laws, & H.E. Barbaree (Eds.), Handbook of Sexual assault: Issues theories, and treatment of the offender. New York: Plenum Press.

Murphy, W.D., Coleman, E.M., & Haynes, M.R. (1983). Treatment and evaluation issues with the mentally retarded sex offender. In J.G. Greer & I.R. Stuart (Eds.), The Sexual Aggressor: current perspectives on treatment. New York: Van Nostrand Reinhold Company.

Paulhaus, D.L. (1988). Assessing self-deception and impression management in self-reports: The Balanced Inventory of Desirable Responding. Unpublished manual, University of British Columbia, Vancouver, Canada.

Petre-Miller, D. (1994, November). Covert sensitisation with intellectually disabled clients. Paper presented at the meeting of the Association for the Treatment of Sexual Abusers, San Francisco, CA.

Raven, J.C. (1960). Guide to the Standard Progressive Matrices. London: H.K. Lewis.Raven, J.C. (1965). Guide to using the Coloured Progressive Matrices. London: H.K. Lewis.

Robinson, S. (1984). Effects of a sex education program on intellectually handicapped adults. Australia & New Zealand Journal of Developmental Disabilities, 10(1), 21-26.

Schoen, J. & Hoover, J.H. (1990). Mentally retarded sex offenders. Journal of Offender Rehabilitation, 16(1/2), 81-91.

Seidman, B.T., Marshall, W.L., Hudson, S.M., & Robertson, P.J. (1994). An examination of intimacy and loneliness in sex offenders. Journal of Interpersonal Violence, 9(4), 518-534.Shipley, W.C. (1946). Institute of Living Scale. Los Angeles: Western Psychological Services.

Simkins, L., & Ward, W. (1989). The Multiphasic Sex Inventory: Diagnosis and prediction of treatment response in child sexual abusers. Annals of Sex Research, 2(3), 205-226.

Swanson, C.K., & Garwick, G.B. (1990). Treatment for low-functioning sex offenders: Group therapy and interagency coordination. Mental Retardation, 28, 155-161.

Watson, D.L. & Tharp, R.G. (1989). Self-directed behaviour: Self-modification for personal adjustment (fifth ed.). Belmont, CA: Brooks/Cole Publishing Co.Wechsler, D. (1981). Manual for the Wechsler Adult Intelligence Scale - Revised. San Antonio, TX: Psychological Corporation.

Zella, W.F. (1994, November). Victim empathy training with developmentally disabled sex offenders. Paper presented at the meeting of the Association for the Treatment of Sexual Abusers, San Francisco, CA.

STEPHEN J. HUCKER (Division of Forensic and Correctional Psychiatry Queens University)


Although most professionals working with sex offenders at the present time are likely to base their treatment programme on cognitive-behavioural approaches, medical treatments are still used extensively by psychiatrists. There is, however, generally a clear understanding of the limitations of exclusively medical approaches; as with the treatment of schizophrenia and mood disorders the most effective approach is a combination of drug treatment with psycho-social therapies.

Table 1. lists the various therapeutic sex drive reduction techniques that have been used by physicians. Castration or orchiectomy has never become an acceptable modality in North America, largely because of the concerns around the issue of free and informed consent in individuals facing criminal sanctions. Nevertheless, the results gathered over many years from European countries provides impressive evidence that the procedure dramatically reduces sexual recidivism rates to about 2-5%. As with most studies of this kind methodological criticisms have been levelled at the published reports. For example, the subjects have on average been past the age when sex offending is at its peak and the recidivism rate may have diminished in any case. As there have been no controlled studies to address this issue, the matter is unresolved

Table 1. Therapeutic Sex Drive Reduction.

  • Orchiectomy (Castration)
  • Oestrogens
  • Progestagens
  • LHRH Agonists
  • Antiandrogens
  • Centrally acting drugs

Oestrogens were among the first sex drive reducing hormones to be used but their association with a high incidence of thrombo-embolic disorders has led to their abandonment for this purpose. The most commonly used progestagenic hormone in sex offender treatment has been medroxyprogesterone acetate(Provera, MPA) though some practitioners in Britain and Canada have favoured the anti-androgen cyproterone acetate (Androcur, CPA). Most recently, drugs which reduce the secretion of pituitary hormones that stimulate the testes to produce testosterone, the main male sex hormone, have been introduced. Finally, there are a number of drugs acting on the central nervous system that have been employed in sex offender treatment. These have included drugs used primarily for treatment of major mental illnesses such as Thioridazine(Mellaril), the mood stabiliser Lithium Carbonate, and the newer antidepressants that specifically increase the levels of serotonin in the brain.

Hormone Treatments

There are three main hormones that are used in the treatment of sex offenders at the present time - medroxyprogesterone acetate (MPA, Provera), cyproterone acetate (CPA, Androcur) and leuprolide acetate (LA, Lupron). All of these substances lower the level of testosterone circulating in the blood but by different means. MPA reduces testosterone levels by increasing the rate the hormone is metabolised by the liver; CPA competes with testosterone at receptor sites in the testes (hence the term ‘anti-androgen’) and also blocks the secretion of the pituitary hormone that stimulates the testes to secrete testosterone; LA, unlike the foregoing, which are steroids, is a protein which suppresses the hypothalamic hormones that stimulate the testes to secrete testosterone(hence they are called Luteinising Hormone Releasing Hormone or LHRH agonists).

There is general agreement that CPA and MPA will reduce the recidivism rates in sex offenders who are willing to take them and do so reliably. In a study by Fedoroff et al. forty six individuals were followed up: Only 15% of those who took the drug relapsed compared with 68% who did not. A study by Hucker et al. also showed that the drug is effective in reducing sexual thoughts, urges and behaviour but only 10% of those who were referred for their study were prepared to consider taking the drug during the three month double blind controlled trial. Table 2. lists the main reported side effects of MPA. In clinical practise the commonest complaints are of tiredness and weight gain. More serious adverse reactions such as thrombo-embolic disorders are, in fact, very rare but the potentially serious risks must always be disclosed to prospective patients in a detailed typed consent form.

Table 2. Side Effects of MPA(Provera)

  • Fatigue
  • Leg cramps
  • Weight gain
  • Insomnia
  • Hot & cold flashes
  • Mood changes<
  • Headaches
  • Increased blood pressure
  • Lethargy
  • Raised blood sugar
  • Sweating
  • Gallstones
  • Thrombo-embolic disorders

The efficacy of CPA is likewise not in serious dispute; the drug has been used extensively with sex offenders for a number of years and its side effect profile is also well known. The adverse effects are similar to those experienced on MPA though gynaecomastia (breast enlargement) occurs in about 15 - 20% of patients.

LA has been available only for a few years. It is used chiefly in the treatment of cancer of the prostate, a malignancy that is often dependant upon testosterone for its proliferation. Being a protein LA is capable of inducing anaphylactic reactions so that a small test dose is given before starting the full dose. Also, the drug causes an initial surge of testosterone production so that it is usual to include treatment with an antiandrogen such as flutamide in the first few days of treatment with LA. Few serious side effects have so far been reported with LA. Up to 60% of patients, however, report hot and cold flashes because testosterone suppression is rapidly achieved. Sweating, rashes, ankle swelling, weakness, muscle pain, lethargy and insomnia have all been reported by some patients. It has also been observed that LA reduces the density of bones and this side effect is under study at the present time. So far, it appears not to cause any symptoms in otherwise healthy men. LA is particularly useful in individuals who have been refractory to treatment with MPA or CPA.

New Types of Anti-depressant in the Treatment of Sex Offenders.

In addition to their usefulness in treating depressions that were hitherto difficult to treat effectively, new types of anti-depressant drugs have been used in the treatment of sex offenders. These drugs increase the levels of the neurotransmitter serotonin in the brain. Deficiency of this substance has been shown to affect mood, impulse control and compulsive phenomena and so its use with sex offenders is based on the supposition that they have problems in some or all of these areas. Table 3. lists the drugs in this class (called serotonin re-uptake inhibitors - SSRI’s) that have been reported so far with sex offenders.

Table 3. SSRI’s Reported in the Treatment of Sex Offenders.

  • Fluoxetine (Prozac)
  • Fluvoxamine (Luvox)
  • Sertraline (Zoloft)
  • Buspirone (Buspar)
  • Chlomipramine (Anafranil)

Although there have been a number of anecdotal reports on the efficacy of these drugs with various types of sex offenders there have been no double blind controlled trials and it is not clear which types of sex offenders may benefit and it appears likely that many do not respond at all. Particular SSRI’s appear to be selected for their apparent usefulness in treating comorbid conditions; thus, fluoxetine and sertraline have been used in sex offenders who have depressive symptoms, chlomipramine for those who show obsessive-compulsive features and so on. Of particular merit is that these drugs , unlike hormones which suppress sex drive completely, are much more acceptable to prospective patients; there are even some reports that suggest that deviant arousal may be reduced while leaving normal sexual interests intact.


Drugs are, unfortunately, not the whole solution to the problem of treating sex offenders. Even enthusiasts would have to concede that they play only a part in a comprehensive treatment programme. They are often unacceptable to potential patients, even those who might benefit, and they provide relief only while taken. On the other hand, our efforts in the treatment of sex offenders by other means are not so encouraging that judicious use of medications can be dismissed out of hand. The risk of recidivism in many sex offenders might be significantly reduced if psycho-social therapies were more often combined with medications.