Background

This guide is intended as a quick reference for those situations in which a psychologist, working in a non-forensic clinical setting, may need to assess whether his or her patient poses a risk of violence. Such situations may include, but are not limited to, assessment of risk in planning for a psychiatric inpatient discharge, release from an emergency department, or assessment of violence risk to third parties.* Also included are those situations, experienced by an estimated 6-11% of clinical psychologists at some point in their career, in which they find themselves the target of a patient who is engaging in inappropriate and threatening approach behavior such as stalking. The topic of managing patients who are actually engaging in violent acts against psychologists is beyond the scope of this guide.

The standard of practice for psychologists working in general clinical settings does not typically include obtaining and evaluating a patient’s criminal history, conducting or reviewing formal forensic evaluations, eliciting testimony from collaterals or informants, or obtaining other information typically available to the forensic examiner. However, courts, insurance companies, and the public clearly expect psychologists in general clinical practice to take “reasonable” actions to foresee and to prevent violence by their patients.

The standard of practice for clinical psychologists does include knowing the laws and resources in his/her state that may pertain to managing dangerous patients, documenting a review of the available medical and psychiatric history for every new patient, reviewing with every new patient the limits of confidentiality and documenting this conversation, and informing the patient how to access emergency or afterhours services if they feel that they may act violently against others or themselves. Psychologists should also have a thorough foreknowledge of how to invoke emergency voluntary and involuntary hospital care should that become necessary.

General strategies for managing these cases typically can be summarized with the following recommendations:

1. Documenting an informed consent discussion with every new patient about the limits of confidentiality;
2. Assessing the risks of violence by using the latest research on violence assessment to
   evaluate risk and protective factors. (Information about these factors is often available on a limited basis in general clinical practice. Good practice includes
   a review of previous medical and psychiatric records where such information may often be discovered.);
3. Working to maintain the therapeutic alliance as a protective factor in itself;
4. Obtaining expert peer and, when appropriate, legal consultation;
5. Documenting the violence risk assessment as well as a plan for mitigating the risk of violence, including actions considered but not chosen; and
6. Implementing the plan to reduce the risk of violence. Included in the plan should be follow-up to assure compliance by the patient and others who are part of the plan.

When a patient appears to pose a threat the psychologist must maintain two primary objectives: protection of the interests of the patient and safety for all parties. Because of the possible tension between these two objectives, the psychologist may feel overwhelmed with fear of the possibility of harm to the patient, harm to a third party, harm to oneself, and fear of liability and professional ignominy if it is later determined that the case was handled badly to the detriment of any of the actors.

The methodical and dispassionate approach that should characterize psychological assessment may be logistically and emotionally impractical in these situations. Fortunately, in the vast majority of such situations, the clinician will have some time to approach the situation thoughtfully. It is relatively rare for an outpatient psychologist to encounter an existing patient who has appeared without warning to pose an imminent risk of harm to a third party. However, in the unlikely event that imminent risk does appear, the psychologist ought to have plans in place to obtain emergency assistance.

To prepare for the possibility of having to evaluate and manage a threatening patient, one or more colleagues who are knowledgeable about managing violence risk should be identified before these situations occur. The opportunity to access expert clinical as well as legal consultation can help the clinician to avoid the “tunnel vision” that can produce substandard risk assessments and precipitous actions that could escalate risk, needlessly damage the patient, or place the psychologist in a position of high risk for litigation or worse.

In considering legal advice, it is important to be clear about who the attorney considers to be his or her client. The attorney may owe a duty to the clinician’s employer but not to the clinician. Further, the role of the attorney is to offer advice only. Any decisions should be made by the clinician with input from clinical supervisors. In keeping with that advice, documentation in the clinical record should not generally refer to advice from legal counsel.

The Nature of Violence Risk and Protective Factors

Clinical prediction of violence remains challenging for all mental health professionals. Results of recent research have underscored the importance of identifying factors and base rates that are associated with the risk of violence in various populations and in various settings. Risk factors may be static, such as gender, age, a history of violence against others or a history of having been victimized. Or risk factors may be dynamic, such as the presence of an acute mood disorder. In general, taking a complete history with the voluntary outpatient will reveal the presence of any static violence risk factors. As the therapeutic alliance develops, dynamic risk factors may also be revealed. Dynamic risk factors, unlike static risk factors, by definition, require repeated measurement, as these factors will be the ones most promising for intervention.

Research on violence risk demonstrates that any violence risk factor will depend upon the context in which that factor appears. For example, expressed threats against a government official or a celebrity may not significantly elevate the evaluation of risk posed by a patient (although the threats must certainly be explored), while expressed threats against a current or former intimate partner may dramatically elevate the risk of violence. Clinicians tend to overvalue explicit verbal threats or their absence in judging risk. However, it is not who makes a threat but who poses a threat. Again, correctly identifying the context of threatening behavior as well as the base rates of violence in that context are essential prerequisites for measuring risk.

Violence risk factors are often interactive. For example, a history of Schizophrenia has generally not been found to be a positive stand alone risk factor for future violence. But many studies have found that, when paired with alcohol or substance use disorder, and especially when accompanied by homelessness or other major situational stress, a diagnosis of Schizophrenia does contribute to an increased risk of violence.

The psychologist who is seeking to measure the risk of violence cannot simply tally all the positive checks on one’s list of violence risk factors and render a unitary snapshot risk prediction for his or her patient. The presence of only one or two highly potent risk factors may be so compelling that one must render a high risk assessment e.g., the patient calls to tell you that he has been drinking, is armed and is driving now to his wife’s place of employment to confront her and her boyfriend. Conversely, the presence of a number of positive static risk factors may be balanced by the presence of a single very powerful protective factor, e.g., the patient who abhors the thought of disappointing or embarrassing her family by acting out violently. The point is that violence risk assessment, like assessment of suicidal risk, is a clinical calculation, not merely a mathematical one. While a psychologist may not be able to determine a precise point of prediction on some numerical “violence” scale, he or she ought to be able to use knowledge of violence risk factors and protective factors to determine whether a patient poses a low, medium, or high level of threat. In this way, violence risk assessment may seem similar to the multifactorial prediction of suicide – or the weather.

Listed below are some of the more common violence risk factors and mitigating factors validated by numerous research studies on violence risk. It should be emphasized once again: The most important risk and mitigating factors in a given situation will depend upon the context in which the potential violence risk appears.

Selected Individual/Psychological Correlates of Violence Risk

• male
• age 15-24
• history of violence: frequency, recency, severity, and age at first occurrence
• history of childhood abuse and neglect
• paranoid ideation
• command hallucinations with violent content
• below average intelligence
• anger-fear problems: frequency, intensity, control
• Psychopathy and other attachment pathology
• history of poor treatment compliance and response

Selected Social/Situational Correlates

• family teaches violence as a mode of communication
• peer group endorses violence, especially adolescent
• job instability, threatened loss, loss
• lower SES
• weapons interest, possession, facility, and recent use
• victim availability, victim pool
• alcohol and/or stimulant use
• pop culture influences, such as TV, media, games

Selected Biological Correlates

• history of CNS trauma (particularly frontal lobe damage/impairment)
• subjective CNS symptoms, e.g., amnesia
• objective CNS signs
• acute Psychotic disorder

Selected Violence Mitigating Factors

• home
• health
• family
• alternatives
• career
• belief systems
• resources
• self-esteem
• reputation
• dignity

Approaches to Violence Risk Assessment

The superiority of an actuarial approach to violence risk assessment over a purely clinical approach has been demonstrated in numerous studies over the past thirty years. However, recent studies have shown that an approach that combines clinical judgment with actuarial projections is superior to the actuarial approach alone. This approach is known as “structured professional (or clinical) judgment.” There has been a number of structured professional judgment risk assessment tools developed in the past 20 years for different settings and for different demographic groups. Below are a few of the ones more likely to be of value to the non-forensic psychologist.

The “Iterative Classification Tree (ICT),” developed by Monahan et al., is useful principally as a checklist against which an outpatient psychologist can consider violence associated risk factors. Completion of the ICT does not render a “score.”

Likewise, other structured professional judgment instruments, such as the HCR-20 (Webster, Douglas, Eaves, and Hart), originally developed for a forensic population but now validated on a number of other populations, may also help assure that the outpatient psychologist and his/her consultant are weighing the most relevant violence risk factors as they consider a plan for reducing the risk.

Another proprietary structured professional judgment instrument that psychologist may find helpful is the WAVR-21 (White & Meloy, 2007). Developed for use by multidisciplinary teams tasked with measuring the risk of violence in places of employment, the WAVR-21 offers the advantage of tracking changes in both dynamic risk factors and protective factors as the case unfolds.

The SAVRY (Borum, Bartel, & Forth, 2006) is modeled after existing structured professional judgment instruments like the HCR-20, but it is designed specifically for assessing violence risk in adolescents (ages 12 – 18). The authors of the SAVRY have included items for rating both risk factors and protective factors. They have designed the protocol with the understanding that youth are actively developing and changing. These characteristics and behaviors can vary in presentation at different stages of psychosocial and emotional development.

The SARA (Kropp, Hart, Webster, & Eaves, 2008) is a clinical checklist of risk factors for spousal assault. Relevant studies using the SARA cited risk factors that have discriminated those who were violent toward spouses and those who were not, as well as on risk factors associated with recidivistic violence among known spousal assaulters. The 20 risk items are grouped into five content areas: (1) criminal history, (2) psychosocial adjustment, (3) spousal assault history, (4) index offense (i.e., current or most recent incident of assault), and (5) a final section that allows the examiner to note risk factors not included in the SARA.

The best violence risk assessment approaches and tools include assessment of both protective and risk factors. Protective factors counterbalance violence risk factors. For example, consider an outpatient who feels rage at the boss who fired him and is having dreams and fantasies about killing the boss. The patient also has a high degree of access to and facility with firearms. However, knowing of his strong family ties, his lifelong commitment to behaving lawfully and morally, his willingness to discuss these feelings with his psychologist, and his willingness to cooperate with a brief hospitalization or medication if necessary, the psychologist may reasonably conclude that the risk of violence from this patient is low.

Treatment Planning with the Potentially Violent Patient

Of course, the determination of high, medium, or low risk of violence is not the end point for the psychologist who is faced with a potentially dangerous patient. Once a risk of serious or moderately serious violence is identified, a plan for addressing the risk must be crafted, documented, and followed-up on. This is where success or failure in pursuing the dual objectives of protecting the interests of the patient and promoting safety for all parties will be measured. Again, as with rendering the assessment, it will be very important to document consultation from a knowledgeable colleague in developing a plan.

Hospitalization is often a good solution to resolving concerns about violence and protecting the interests of the patient. Hospitalization may be voluntary or it may be involuntary depending upon the facts, the laws on involuntary commitment where the psychologist is practicing, the availability of acute psychiatric beds, and other factors. Coordinating with the inpatient staff to develop post-hospital follow-up plans will be essential as most acute psychiatric hospitalizations are rather brief.

Apart from hospitalization, there are many other actions that may resolve acute risk of violence by a patient toward others. It is beyond this guide to enumerate them all. In cases where risk is not considered to be high, it may be possible to develop an outpatient safety plan with the patient and manage risk by increasing the frequency of treatment visits, obtaining a medication consult, or arranging telephone contacts between visits. In such instances, there will need to be frequent re-assessments of the risk level with appropriate modifications in the safety plan and/or treatment plan. The most ethically defensible plans for mitigating the risk of violence will address those risk factors that are dynamic in nature and will protect the interests of the patient.

In considering management of legal risks associated with caring for a patient who poses a risk to others, one should exhaustively document assessments, consultations, intervention plans (including those considered but rejected), and follow-up actions. This documentation will constitute a psychologist’s best defense against later litigation.

Document prepared by the Section VII Working Group on Violence Risk Assessment: David Drummond, Phillip Kleespies, Marc Hillbrand, and Lisa Firestone.

December 2012

*A full discussion of the clinical, ethical, and legal dimensions of the so-called duty to warn or protect third parties (“Tarasoff”) is beyond the scope of this guide. However, there are a number of excellent books and papers on this topic.