Three events occurred in 1999 that motivated me to develop a risk assessment tool—and guidelines for risk assessment—for adolescents who had offended sexually. First, as a clinician and researcher at the former Thistletown Regional Centre, Sexual Abuse: Family Education and Treatment Program (SAFE-T) in Toronto (1983-2013), I had recidivism data from a 10-year, prospective treatment follow-up study for a sample of adolescent males and females who had offended sexually. At that time, there were very few studies regarding risk factors for adolescent sexual recidivism, so this was an opportunity to look at our sample to determine which factors might be related to future risk—in addition to an examination of the impact of specialized treatment.
The second event in 1999 was the release of the Static-99 by Karl Hanson and David Thornton: a brief, actuarial scale to estimate risk of sexual recidivism for adult males. Despite the authors’ cautions regarding the use of the tool with individuals under age 18, the Thistletown Regional Centre SAFE-T Program began to receive referrals from local evaluators where 12- to 16-year-old youth were being rated as “high risk” using the Static-99. Of course, 2 of the 10 risk factors on the Static-99 (young age, never lived with a lover for at least 2 years) were always present for these youth, and this very likely inflated the risk estimates.
The third event that occurred in 1999 was related to a risk assessment that we completed 10 years earlier. In 1989, we assessed an adolescent male who had offended sexually, and we concluded that he was a high risk to reoffend—based on our unstructured clinical judgment. In 1989, there was virtually no guidance from published research regarding risk assessments for this population, and most professionals relied on unstructured clinical judgment. Ten years later—in 1999—this young man was a father of a young boy. A child protection worker came across the 10-year-old risk assessment and subsequently removed the young boy from his parents. After a battle with child protection authorities regarding the inappropriateness of this action, I thought that it was important to develop an empirically-informed approach to providing risk estimates for adolescents who had sexually offended and provide guidelines for communicating risk estimates. It was my hope that a structured risk assessment tool would address these goals.
In 2000, I developed a pilot version of the ERASOR (Estimate of Risk of Adolescent Sexual Offense Recidivism); a 23-item checklist of risk factors for adolescent (i.e., age 12-18) sexual recidivism. The original version was subsequently adapted (ver. 1.2 and 2.0; Worling & Curwen, 2001), and the result was a single-scale instrument with 25 risk factors. Although it was my hope that a risk assessment tool could provide an adequate degree of predictive accuracy to assist with forensic decisions, the research support regarding predictive accuracy has not been consistently strong. This has also been the case for most checklists designed to predict future sexual recidivism, as most tools result in only moderate predictive validity (see Hanson & Morton-Bourgon, 2009; Viljoen, Mordell, & Beneteau (2012).
With the growing recognition of the importance of strengths and protective factors, I developed a structured checklist of putative protective factors (DASH-13) in 2013 in an effort to enhance assessments for adolescents who had sexually offended by encouraging practitioners to consider protective factors, in addition to risk factors. Given the lack of guidance in research and clinical practice for how best to combine protective and risk factors, however, I felt that the next step was to replace both the ERASOR and the DASH-13 in my practice with a new measure--one that simultaneously considers both protective and risk characteristics, and one that is designed specifically to inform treatment decisions rather than to predict future risk. In 2017, this resulted in PROFESOR.
There are several reasons why I am now choosing to use the PROFESOR rather than the existing risk-prediction measures that have been widely used in the past.
1. In their meta-analysis of the research regarding the predictive validity of the J-SOAP-II, J-SORRAT-II, ERASOR, and the Static-99, Viljoen, Mordell, and Beneteau (2012) concluded that all four measures provided estimates that were above chance levels; "however, given that the effect sizes were moderate, these tools may be insufficient to make predictions that require a high degree of precision" (p. 12). It is also critical to note that Viljoen et al. found that "in many cases there were high levels of heterogeneity across studies, meaning that studies did not find uniformly positive results" (p. 11).
2. Between one third and 100% of the risk factors on the popular risk prediction tools studied by Viljoen et al. (2012) are static; adolescents are not. This is particularly problematic for follow-up assessments, as static factors will not be removed during the course of any intervention. If an adolescent has ever offended sexually against a stranger, for example, this will always be rated as "Present"--regardless of any gains that the individual makes. Furthermore, how does it inform treatment to know that someone has offended sexually against a stranger in the past or that someone had a previous placement in a special education program? For some adolescents, the number of risk factors on the these risk assessment measures actually increases as a result of treatment, as they end up disclosing new historical offenses which then impacts the coding of the static, or historical, risk factors.
3. There are some factors on these popular risk prediction measures that are not consistently supported in the literature as risk factors for youth. Of course, this significantly compromises the validity of the risk prediction measures.
4. Many of the existing risk assessment tools designed for youth can only be used with a very narrow age range (typically a span from 12-17 or 18). This makes it very difficult to follow-up older adolescents over time using a consistent tool. Given that many risk assessment tools designed for adults are not intended to be used with individuals who offended as younger teens (e.g., see Static99R), there are currently few, if any, tools that are useful for emerging adults (i.e., those aged 18-25) who offended sexually when they were young teens (Bumby, 2014).
5. The popular risk assessment tools reviewed by Viljoen et al. (2012) contain ONLY risk factors--there are no protective factors. There have been a number of recent advances in the field with respect to strengths and protective factors, and it is now widely held that assessments for individuals who have offended sexually should be focused on both protective and risk factors (e.g., Langton & Worling, 2015).
6. It is very challenging, if not impossible, to utilize many of the existing risk assessment tools for youth who have been involved in downloading and/or distributing child abuse images. Some of the most widely used measures have been designed only for youth who have engaged in contact sexual offenses, and other risk assessment tools contain risk factors that cannot be coded without reference to a specific person who has been directly victimized.
7. The final categorization of low, moderate, or high risk on some risk tools can be misused and misunderstood. For example, some adolescents are inappropriately labeled as "high risk youth" or "high risk sex offenders". It is also critical to stress that there are vast differences in how people interpret the arbitrary risk categories of high, moderate, and low risk (Lehmann, Thornton, Helmus, & Hanson, 2016).
I hope that PROFESOR will address some of these limitations and that it will be a useful tool in our work with adolescents and emerging adults who have offended sexually by helping to inform possible interventions.
Dr. James R. Worling, Ph.D., C.Psych.
