Handbook of Violence Risk Assessment

Randy Otto, of the Department of Health Law and Policy of the University of South Florida, and Kevin Douglas, of the Department of Psychology at Simon Fraser University, have edited one of those books that makes one wonder why no one thought of it before. The recent explosive growth in structured approaches to violence risk assessment, whether by actuarial methods or structured professional judgment, has not, before now, generated a book that brings together the most widely used instruments and allows the authors of those instruments to describe how and why they should be used.

Handbook of Violence Risk Assessment fulfills this need, although the PCL-R, one of the most widely used instruments, is described, not by its creator Robert Hare, but by Dematteo, Edens, and Hart. The consequent description is authoritative and, perhaps inevitably, more detached and critical than the descriptions provided by those chapter authors who are describing their own intellectual offspring. I am inclined not to quibble; the PCL-R is described in a preliminary section, and the reason that it is treated differently from the other instruments is stated in the introduction. The editors say that their book is intended for correctional personnel, attorneys, judges, psychologists, and psychiatrists. I suspect that other mental health workers and probation officers will also find it useful.

The discussion of the PCL-R is followed by 11 chapters describing the other structured approaches to violence risk assessment, divided by the age groups that the instruments are designed to assess. Chapters covering child and juvenile risk describe the Early Assessment Risk Lists (EARL) for boys and girls, the Structured Assessment of Violence Risk in Youth (SAVRY), and the Youth Level of Service/Case Management Inventory (YLS/CMI). Chapters regarding adult risk describe the Violence Risk Appraisal Guide (VRAG) family (including its sex offender and domestic violence variants, SORAG and DVRAG), the Violence Risk Scale (VRS), the Historical-Clinical-Risk Management-20 (HCR-20), the Classification of Violence Risk (COVR), and the Level of Service (LS) assessment instruments. These chapters also describe instruments that have been developed to assess risk in sex offenders (the Static 99, Sexual Violence Risk-20 (SVR-20), and Risk for Sexual Violence Protocol (RSVP)) and risk of domestic violence (the Spousal Assault Risk Assessment Guide (SARA)).

An overview chapter by Heilbrun, Yasuhara, and Shah describes the uses to which structured instruments are put, and helpfully so. Several concepts, including the Risk, Needs, Responsivity model, that recur in the risk assessment literature but that may not be familiar to some readers of The Journal, are discussed and well referenced. The chapter makes one valuable point more clearly than it has been made elsewhere. An actuarial approach does not restrict the assessor to using static risk variables, those that are not amenable to change. As the authors point out, there is no reason in principle that dynamic variables, such as assessments of anger and mood, cannot be included in an actuarial instrument. The challenge lies in operationalizing them effectively.

Some points in the overview are clearly intended to review and advance the field. One point relates to the proper outcome measure to be used in studies that examine the predictive validity of structured instruments. Heilbrun and colleagues seem to suggest that clinicians should be guided by a sensitivity hierarchy, whereby self-report is a more sensitive measure of violence than collateral information which is, in turn, more sensitive than an official record, for instance of arrest or conviction. They may be correct, but sensitivity is only one of the desirable qualities in a screening instrument. Another is specificity, and one of the difficulties facing risk (or any other form of) screening is that most ways of increasing either sensitivity or specificity decreases the other. A hierarchy of outcome measures based only on sensitivity also fails to acknowledge the value of multiple sources, a point made later in the book by the authors of the HCR-20.

I would have welcomed more discussion of the suggestion that a move toward measuring situational influences on risk would increase predictive accuracy: my reading of the literature is that the predictive accuracy of structured instruments has been stubbornly resistant to improvement beyond an area under the curve of between .65 and .75, whatever variables are used. Finally, there is a strong claim that anamnestic approaches, whose roots in applied behavioral analysis are helpfully explained, are “not well suited” to risk assessment (p 8). In clinical practice, some of the first questions that arise once a history of violence has been identified concern the circumstances in which that violence occurred. As I understand it, clarifying the circumstances in which violence has occurred is central to an anamnestic approach. The confusion may stem from the chapter's rather hard distinction between risk assessment, used to mean generating a probability or odds ratio, and other clinical approaches to risk. Anamnestic assessment certainly seems not to be well suited to the provision of a probability or odds ratio, but an odds ratio is not what my colleagues usually seem to be seeking when they ask for advice on risk.

Heilbrun and colleagues describe a range of contexts in which structured assessments are now conducted. One is the workplace. Violence in the workplace is the topic of Violence Assessment and Intervention, a very different book. In fact, a title that refers to the assessment of violence risk is one of the few things that it shares with Otto and Douglas' book. In place of Otto and Douglas' authoritative reviews and provocative discussion is the confident opinions of two experienced people who have a series of techniques to offer. James Cawood is the former president of Factor One, a California-based corporation specializing in “violence assessment, security consulting and investigations” (book back cover). Michael Corcoran is an ex-Secret Service agent who, the book cover notes, is president of Henley-Putnam University, “the only accredited institution to offer Bachelor's, Master's and Doctoral degrees in the areas of Intelligence Management, Counterterrorism Studies and the Management of Personal Protection.”

Cawood and Corcoran distinguish themselves from health workers. Their clients are usually employers worried about what their employees and sometimes disgruntled ex-employees may do. The book is written for practitioners who are “Confronted on a regular basis with calls from frightened people who want you to keep them safe” (p xi) and for those who do not have the luxury of perfect information and “have to act now” (p xi). The person being assessed is usually referred to as the instigator or the person of interest. The authors emphasize “the call” that triggers the assessment, clearing one's desk, and assessing the mental state of the caller whose emotional engagement is to be reduced, when possible.

The authors make spectacular generalizations, some of which seem questionable when mental illness is part of the equation, including, “All violent behavior is caused by the need to establish control” (p 6). They summarize the thesis by stating that “if threat assessment professionals understand the instigators' conception pertaining to the need for control, they can influence this perception” by establishing rapport and by managing the environment, including that part of the environment that is a consequence of what the law enforcement agencies are doing (p 7). The aim is to allow instigators themselves to develop nonviolent ways of regaining control. The assumption throughout seems to be that in many instances the threat will first appear in the workplace and will have to be managed there.

The passages that struck me as most likely to be of interest to mental health professionals were those describing how to interview someone who is dangerous or threatening. The authors have views on where interviewer and interviewee should sit (equal distance from a door to which both have unobstructed access), their angle in relation to each other (10 and 2 o'clock), their choice of clothing (the interviewee should not be made to feel outclassed), and their general demeanor (no sunglasses during the interview). The authors state that employers may have a post-Tarasoff “duty to warn under civil law precedents” (p 318). This postulation intrigued me, not least because the equivalent post-Tarasoff duty on health professionals varies from one state to another and remains to some extent unclear. I would have welcomed a discussion: presumably, the employers with a duty would sometimes be hospitals.

Raymond Flannery, an associate clinical professor in the Department of Psychiatry at Harvard, has chosen a title that may have been designed to provoke people like me, who write that dangerousness in mental health settings is more usefully treated as a quality of situations than of people. In fact, most of The Violent Person: Professional Risk Management Strategies for Safety and Care deals with situations including domestic violence, psychiatric emergencies, and youth violence and those that generate psychological trauma. The author is an advocate of training, specifically, “enhanced behavioral emergency safety training” (p 12), as a means of reducing mental health workplace violence and its consequences.

Flannery takes a very broad approach, noting that refraining from acting violently, like intervening to mitigate the consequences of violence, requires attention to three domains of good health: attachments, mastery, and meaningful purpose. He includes instructions for relaxation exercises, noting, “If a true emergency arose, your mind and body would immediately rise from the relaxation state, and you would be capable of solving the problem” (p 187). He also provides advice on how to dress to minimize violence risk (neat, professional attire, in contrast to Cawood's injunction that the interviewee not feel outclassed) and an account of biological changes induced in the brain by PTSD.

Of the books under review, Herschel Prins has written the one that deals least directly with risk assessment but that is, by some distance, the most charming. Offenders, Deviants or Patients? contains chapter titles from Shakespeare, an autobiographical introduction, and a rather unnecessary apology concerning the limits of the author's knowledge of the law. It ends with what he calls an envoi, a term that I had to look up: it comes from the old French and refers to an author's concluding words. The book is now in its fourth edition and is a readable and very personal account of Prins' experience of the U.K.'s systems of care for mentally disordered offenders. Most of us would be content simply to get to the fourth edition, but the author has done so with insight and a personal style that derives in part from his work as a probation officer and seems to embody the U.K. Probation Service's erstwhile mission to, “advise, assist and befriend.”1 Would that it were other than erstwhile.

• © 2012 American Academy of Psychiatry and the Law