Structured professional judgment (SPJ) is a formal method of evaluating the risk of future violence. Such evaluations are required in numerous contexts, such as parole decisions, release from forensic hospitals, civil commitment, when violence threatens the workplace, among others. Originating in the mid-1990s, the SPJ approach consists of a family of risk assessment measures devoted to specific types of violence. This article reviews the following topics: brief description of the SPJ approach, SPJ instruments, core assumptions of SPJ, use of the SPJ model in practice, and research support for SPJ.
Description of the SPJ Approach
The SPJ approach is a comprehensive approach to violence risk assessment. It structures the decision-making process from the gathering of information necessary to complete a risk assessment, to rating risk factors, formulating risk, specifying future scenarios of violence, suggesting risk management strategies, and facilitating risk estimates of the likelihood of future violence. It is a method of decision-making that is founded upon empirically supported violence risk factors and well-supported criminological and psychological theories devoted to understanding why people decide to act violently (collectively called decision theory, specific examples of which include situational action theory and rational choice theory). It helps evaluators determine the presence of risk factors, which risk factors are most relevant or important in a given case, why a person has chosen to act violently (what motivates, disinhibits, or destabilizes the person’s decision-making), what the person might do in the future, what risk reduction strategies make most sense in a given case, and the level of risk a person poses.
The SPJ approach, though structured and comprehensive, does not use equations, formulas, or algorithms to produce numeric estimates of future risk of violence. Some approaches propose to do so, but the developers of the SPJ approach have noted shortcomings in these attempts. For instance, they have argued that actuarial estimates are often based on single samples or a small number of samples; that they often omit important risk factors due to sample dependence; and that their numeric estimates vary across samples and settings. The authors have also asserted that actuarial estimates tend to weight risk factors in a manner that assumes equal weights for all people, which could force evaluators to provide higher weight to certain factors that, in their judgment, had little to do with a person’s violence compared with factors that were essential to understanding a person’s violence. In addition, according to the authors, actuarial estimates are less geared toward risk management and capturing change over time, and they do not facilitate formulation or scenario planning.
SPJ Instruments
There are a number of SPJ instruments, each containing between 20 and 30 violence risk factors. The most general, and most widely used, instrument is called the Historical–Clinical-Risk Management-20 (HCR-20), which is applicable to all forms of violence. The HCR-20 contains 20 risk factors dealing with historical factors (personality disorder, mental illness, past violence, substance use problems, and past noncompliance), recent functioning (violent ideation and instability), and future functioning (adequacy of professional plans, personal support, and stress and coping). In a survey of over 2,000 clinicians, the HCR-20 was found to be the most commonly used violence risk assessment instrument across more than 40 countries. It has been translated into 20 languages and has had several hundred empirical evaluations conducted on it.
There are also SPJ instruments that deal with specific forms of violence, including sexual violence (Risk for Sexual Violence Protocol [RSVP] and Sexual Violence Risk-20 [SVR-20]), stalking (Guidelines for Stalking Assessment and Management [SAM]), intimate partner violence (Spousal Assault Risk Assessment [SARA]), group-based violence (Multi-Level Guidelines [MLG]), honor-based violence (the PATRIARCH), violence screening (the TRIAGE), short-term violence (Short-term Assessment of Risk and Treatability [START]), violence among children (Early Assessment Risk List-20, for Boys [EARL-20B]; Early Assessment Risk List-21, for Girls [EARL-21G]), and violence among adolescents (Structured Assessment of Violence Risk in Youth [SAVRY]).
Core Assumptions of SPJ
The SPJ model makes some basic assumptions. These assumptions are as follows: (a) professionals with appropriate training, education, and experience can make reliable decisions about risk; (b) there are changeable, or dynamic, risk factors that must be included on any risk assessment instrument for the purposes of risk reduction; (c) when dynamic risk factors change, the likelihood of future violence changes in kind; (d) dynamic risk factors must be monitored over time; (e) in general, the more risk factors present in a case, the higher the risk, but that risk is not solely determined by a linear combination of risk factors (i.e., a small number of highly salient risk factors can lead to violence).
Using SPJ in Practice
SPJ instruments share a fundamentally common approach to the violence risk assessment task. There are some slight differences across measures depending on the date of publication. For all measures, evaluators rely on interview information (when possible and not contraindicated) and file-based information (e.g., correctional or psychiatric records; employment records). Supplemental sources of information include collateral interviews (e.g., with family members, victims, or coworkers, when appropriate), observation (e.g., when a patient resides on an inpatient unit), and medical or psychological testing.
In addition to determining which risk factors are present, evaluators are guided through several assessment steps that facilitate answers to the following questions:
- Of the risk factors that are present, which are the most important to understanding why a person has been violent?
- How can the risk factors be integrated to produce a formulation of risk (essentially, why has the person done what he or she has done)?
- What is the evaluator worried the person might do in the future?
- What can be done to prevent the violence?
- What level of risk does the person pose?
These are basic questions that any good risk assessment must address. The SPJ approach s imply provides a structure to help evaluators address these questions. It is reflective of a larger principle—supported by theory and research— that human beings make better decisions about complex matters involving many pieces of information within a structure than without structure.
To illustrate the typical SPJ decision-making process, HCR-20 Version 3 will be used as an example. To be clear, all other SPJ instruments use a similar logic. Steps 1 and 2 are simply gathering information and rating risk factors. Risk factors on SPJ instruments are rated based on operational definitions and coding notes. Risk factors are rated as absent, present, or possibly/partially present. No risk factors are deemed, a priori, as definitively more important than others in any given case. That is the work of the evaluator to determine. Nomothetic data (support for a risk factor at the sample or population level) are treated as support for hypotheses to explore at the case, or idiographic, level. In one case, substance use problems might be more important than child abuse; in a different case, the opposite might be true. Evaluators must have the discretion to determine this.
To facilitate evaluators’ judgments of the importance of risk factors within individual cases, SPJ instruments have introduced the concept of idiographic relevance. On the HCR-20 Version 3, this is Step 3. Idiographic relevance of a risk factor is different from its risk-enhancing status in the population. Rather, it refers to the risk-enhancing property at the case level. Just because a person has a risk factor that is supported at the population level does not necessarily mean that it is associated with violence equally strongly, or at all, for any given person. The evaluator’s job is to determine how important a given risk factor is for understanding a given person’s violence.
This issue of relevance is central to the next step within SPJ, represented in Step 4 on the HCR20 Version 3, formulation. Formulation is a core competency in most mental health fields. Its essence is to explain why something happened. For example, if a therapist sees a new client who is claiming to be depressed, that therapist needs to know why the person is depressed. What sorts of thoughts lead to her depression? Does she have negative self-evaluations? How does she interpret significant others’ opinions of her? Risk formulation is not different. The evaluator needs to know why a person has been violent in order to make decisions about risk and make suggestions for how to prevent future violence.
The SPJ approach allows a variety of conceptually and theoretically meaningful methods of formulation. It generally promotes the integration of risk factors into conceptually meaningful clusters with root causes. Then, it promotes ascertaining how such risk factors influence a person’s decision-making about violence. Why does a person think violence can achieve some goal? Why is he not worried about the negative consequences of violence? One of its major recommendations draws from decision theory, an approach with a broad support in criminology and psychology. Essentially, this approach recommends that evaluators consider how risk factors influence a person’s decisions to act violently. It describes motivations (i.e., factors that increase the perceived benefit of violence, such as gaining status, control, or dominance), disinhibitions (i.e., factors that decrease the perceived cost of violence, such as lack of empathy, negative attitudes, or nihilism), and destabilization (i.e., factors that impair careful decision-making, such as impaired cognition, perception, or attention). It should be noted that this approach is also highly consistent with the assumptions of empirically validated risk reduction techniques that focus on offender thinking, such as cognitive skills training or problem-solving skills training.
After formulation, evaluators turn their attention to the future. In essence, this step (Step 5 of the HCR-20 Version 3) asks evaluators to specify what forms of violence they are worried the person might commit in the future, if the person were to be violent, as well as the likelihood of those scenarios. For example, will a person rob a bank, sexually assault a child, beat someone up in a bar, or punch his mother? Scenario planning is used in numerous fields, including military planning, emergency preparedness, health-care planning, and business. It helps people plan in a context of uncertainty, which describes the risk assessment context. In crafting scenarios, evaluators are mindful of the motivations of violence, who might be future victims, and how serious the violence might be. They draw on a person’s past acts of violence, but also think about whether any twists or exacerbations of violence might occur.
Next, the evaluator specifies what might be done to reduce the risk of violence. This step (Step 6 on the HCR-20 Version 3) takes into account previous steps (why has a person acted violently and what might she do in the future?) to put plans into place or recommend to others to do so. Evaluators first specify what the warning signs for an escalation of risk would be based on the person’s history and future context (e.g., resuming drinking, dropping out of treatment, and getting into arguments with a spouse). Then, they recommend strategies and tactics for monitoring (frequency and method of observation), supervision (its strictness and method), intervention (types of rehabilitation), and victim safety planning (how prospective victims might spot warning signs and cope with stressors).
Finally, based on the preceding, the evaluator can offer summary risk ratings about the likelihood of future violence, how serious it might be, and its imminence. This is Step 7 on the HCR-20 Version 3. These are narrative, categorical judgments, framed in terms of low, moderate, and high. For reasons discussed earlier, they do not depend on numeric cutoffs or algorithms. This is a principled method for risk communication that is intended to summarize previous steps. A judgment of high risk means that the evaluator has a high level of concern that the person will be violent in the coming 6–12 months, that a high degree or intensity of intervention will be required to mitigate risk, and therefore that the person is at a high priority for the receipt of services and resources.
This method has been criticized by some commentators because it does not involve algorithms or numeric cutoffs, and it does not produce numeric probability estimates of future violence (although these are routinely published in academic articles). Despite the criticism, this method has also been shown across roughly four dozen studies to be as or more accurate than algorithmic approaches to risk estimation. Indeed, as reviewed in the next section, the judgments of low, moderate, or high risk tend to add incremental validity to optimized, numeric (i.e., actuarial) combinations of risk factors.
The Nature of Research Support for the SPJ Model
There have been several hundred empirical evaluations of the SPJ approach as of 2017. The research has occurred across more than two dozen countries. Meta-analyses show that this approach is at least as accurate as the actuarial approach in terms of the predictive validity of risk factors, if not more so. Moreover, changes in risk factors over time have been shown to predict changes in violence in the future. That is, there is support for the dynamic risk of SPJ measures.
Nonalgorithmic judgments of low, moderate, or high risk have routinely been found to add incremental predictive validity to the summing of the risk factors or to other risk assessment measures, including actuarial ones. There are several possible reasons why this might be the case. First, actuarial measures treat all persons the same and presume that all risk factors should have the same weighting across people, whereas SPJ measures do not. They require the evaluator to specify which risk factors are the most important in any given case. Second, most actuarial measures presume that a simple linear combination of risk factors is the optimal method of determining risk. The SPJ agrees that, in general, the more the risk factors are, the greater the risk. However, it specifically allows for deviations from a linear model. For instance, if a given person were to have an acute psychotic break, be heavily intoxicated, and be experiencing violent ideation, it could be reasonable to conclude that he or she is at high risk of violence. Most actuarial methods would classify such a person as at low risk because there are only three risk factors present. Third, as referred to earlier, putatively precise numeric estimates produced by actuarial methods have yet to be shown to be consistently robust across samples. That is, they change across samples, likely due to the inherent uncertainty and complexity involved in risk assessment. It is difficult to anticipate future factors that affect risk—SPJ methods such as the HCR-20 Version 3 attempt to do so through the inclusion of forward-looking decision aids (e.g., the Risk Management Scale, scenario planning).
References:
- Douglas, K. S., Hart, S. D., Groscup, J. L., & Litwack, R. (2014). Assessing violence risk. In I. Weiner & R. K. Otto (Eds.), The handbook of forensic psychology (4th ed.), (pp. 385–441). Hoboken, NJ: Wiley.
- Douglas, K. S., Hart, S. D., Webster, C. D., & Belfrage, H. (2013). HCR-20 (Version 3): Assessing risk for violence, user guide. Burnaby, BC: Mental Health, Law, and Policy Institute, Simon Fraser University.
- Douglas, K. S., Hart, S. D., Webster, C. D., Belfrage, H., Guy, L. S., & Wilson, C. (2014). Historical- Clinical-Risk Management-20, Version 3 (HCR-20 V3): Development and overview. International Journal of Forensic Mental Health, 13, 93–108. Retrieved from https://doi.org/10.1080/14999013.2014.906519
- Hart, S. D., Sturmey, P., Logan, C., & McMurran, M. (2011). Forensic case formulation. International Journal of Forensic Mental Health, 10, 118–126. doi:10.1080/14999013.2011.577137
- Heilbrun, K., Douglas, K. S., & Yasuhara, K. (2009). Violence risk assessment: Core controversies. In J. L. Skeem, K. S. Douglas, & S. O. Lilienfeld (Eds.), Psychological science in the courtroom: Consensus and controversy (pp. 333–357). New York, NY: Guilford.
- Litwack, T. R., Zapf, P. A., Groscup, J. L., & Hart, S. D. (2006). Violence risk assessment: Research, legal, and clinical considerations. In I. B. Weiner & A. K. Hess (Eds.), The handbook of forensic psychology (3rd ed., pp. 487–533). Hoboken, NJ: Wiley.
- Yang, M., Wong, S. C. P., & Coid, J. (2010). The efficacy of violence prediction: A meta-analytic comparison of nine risk assessment tools. Psychological Bulletin, 136, 740–767. doi:10.1037/a0020473