The Scottish Government

« Previous | Contents | Next »

Listen

SERIOUS VIOLENT AND SEXUAL OFFENDERS: THE USE OF RISK ASSESSMENT TOOLS IN SCOTLAND

CHAPTER SEVEN: CONCLUSIONS

7.1 This study has revealed wide variations in the approaches adopted to risk assessment by different professional groups. Even within agencies or professional groups different methods of risk assessment were evident. Police respondents to the audit, for example, highlighted the absence of a national approach to risk assessment, which could create difficulties when offenders moved from one part of the country to another. This issue has been recognised by ACOP (Scotland) which has established a working group to identify the most appropriate tool for use on a national basis. The introduction of a consistent approach to risk assessment should facilitate the sharing of information across forces and reduce the likelihood of differing interpretations being placed on risk assessments derived from different approaches.

7.2 The Scottish Prison Service is also in the process of introducing a standardised approach to risk assessment which draws upon a battery of risk assessment tools. Risk assessments will be undertaken by psychologists on all prisoners sentenced to four years or more to inform their management within the prison system, to facilitate the matching of prisoners to appropriate prisoner programmes and to inform decisions about their release. However the absence of a co-ordinated approach to risk assessment between prisons and community-based services will, unless addressed, limit the usefulness of prison-based assessments once a prisoner has been released.

7.3 The audit has also highlighted the limited use being made of tools for assessing risk of sexual or violent offending that have been appropriately validated. The use of validated instruments was most evident in prison and mental health settings, though the fact that tools had not been validated on Scottish psychiatric populations was thought to detract from their value with the latter. Where tools used by social workers had been validated, these had not been validated for use with Scottish offender populations and their ability to predict risk of sexual and violent offending and risk of harm had not been assessed.

7.4 With a few exceptions - usually specialist projects working with sex offenders - the tools employed by social workers had been developed as generic risk assessment instruments and were not designed specifically to assess the risk presented by serious violent or sexual offenders. The most commonly used tools - the RAGF and the LSI-R - had complementary weakness and strengths but neither was considered sufficient for the task. Whilst RAGF was better able to identity risk of harm, it had not been validated and was considered by some to over-rely upon subjective judgement. LSI-R was viewed as useful in aiding targeting of resources, informing supervision plans and facilitating the measurement of change but it was limited in its ability to assess the risk of harm and had not been validated for use in the Scottish context. More importantly, strong doubts were expressed about the value of the LSI-R with sexual offenders, young people, violent offenders and women.

CRITERIA FOR THE CHOICE OF RISK TOOLS

7.5 There was clearly some interest among various professional groups in the development of a standard approach to risk assessment. However this was accompanied by a recognition that in practice it was not something that might easily be achieved. In the interim there would be some merit in a movement towards the adoption of tools for assessing risk of serious violent or sexual offending based on their 'fitness for purpose' since at present factors other than their validity and predictive ability often appear to have informed the approaches to risk assessment employed. This report therefore concludes by summarising what existing research would suggest are essential and desirable characteristics of risk assessment tools for use with these groups of offenders.

7.6 Existing research would suggest that the following are essential features of risk assessment tools:

• Validated - at least one peer reviewed publication on validation of the tool
• Validated against a relevant population commensurate with the target group for the tool
• Actuarially based and empirically grounded in risk factors with a proven track record in the research literature
• Must be able to differentiate between high, medium and low risk with a high degree of accuracy (to ensure risk categories and subsequent risk management plans are justified and proportionate) ²⁸
• Has inter-rater/assessor reliability (all assessors will use the tool the same with the same result)
• If possible, validated against a Scottish population (or it could be validated retrospectively in a reasonable time-scale, for example against case, prison or parole records)

7.7 The following, on the other hand, can be considered desirable features of risk assessment tools:

• User-friendly
• Resource lean
• 'Easy' to train staff to use appropriately
• Process of use is transparent and accountable ²⁹

7.8 It is then possible, drawing upon the material summarised by Kemshall (2002) and the experiences reported by respondents who have used them to classify each of the main risk assessment tools employed in Scotland against these criteria (Table 7.1).

Table 7.1: Criteria for the selection of risk assessment tools

With general offender populations but not established with serious violent or sexual offenders.
^B Validated only with male prisoner population.
^C Mostly validated on patient populations.

7.9 This classification may assist practitioners in assessing the potential value of some of the range of tools that are available to assist in the assessment risk among serious violent and sexual offenders. Overall there appears to be some trade off between predictive validity on the one hand and ease of administration on the other. Those tools which appear best suited to assessing risk of offending and risk of harm among the specific groups of offenders who are the subject of this report are those that have been developed specifically for this purpose and whose administration requires both training in their use and clinical skills.

7.10 In addition, such tools are often highly specific, developed in particular settings such as psychiatric hospitals or prisons, and are targeted at particular client groups (for example the PCL-R at high risk psychopaths). It cannot therefore be assumed that they will have general transference to all high risk groups. This suggests that tool selection must also pay attention to the target group of the tool (sexual, violent, psychopathic) as well as to the other essential criteria. This indicates that risk assessment is unlikely to be carried out in the future by one stand alone tool, but is perhaps better understood as a process in which a combination of well validated tools may be required. This may, for example, consist of a user friendly, validated tool for initial screening for high risk for use by frontline criminal justice practitioners. In those few cases that warrant further in-depth assessment in order to provide well informed risk management strategies or to inform courts in the cases where an Order for Lifelong Restriction is being considered, appropriately chosen specific tools could also be used.

7.11 The main conclusion to be drawn from this study is that there is currently no consistent, co-ordinated approach to the assessment of the risk presented by serious violent and sexual offenders in Scotland. Different professional groups are utilising different methods of risk assessment. This undoubtedly detracts from the effectiveness of multi-agency risk assessments and may have adverse consequences for the effective management of offender risk, particularly for those offenders making the transition from secure penal or health settings into the community.

« Previous | Contents | Next »