Director of Public Prosecutions (WA) v GTR

JURISDICTION : SUPREME COURT OF WESTERN AUSTRALIA
IN CRIMINAL

CITATION: DIRECTOR OF PUBLIC PROSECUTIONS (WA) -v- GTR [2007] WASC 318

CORAM: McKECHNIE J

HEARD: 22 & 23 NOVEMBER 2007

DELIVERED : 20 DECEMBER 2007

FILE NO/S: MCS 17 of 2007

BETWEEN: DIRECTOR OF PUBLIC PROSECUTIONS (WA)
Applicant
AND
GTR
Respondent

Catchwords:

Dangerous sex offender application - Admissibility of Children's Court conviction - Risk assessment tools - Validity of opinions based on tools

Legislation:

Dangerous Sexual Offenders Act 2006 (WA)
Spent Convictions Act 1988 (WA)
Young Offenders Act 1994 (WA)

Result:

Application refused

Category: A

Representation:

Counsel:

Applicant: Mr D Dempster
Respondent: Mr L C Carter & Ms M R Barone

Solicitors:

Applicant: Director of Public Prosecutions (WA)
Respondent: Aboriginal Legal Service (WA)

Case(s) referred to in judgment(s):

Buck v Comcare (1996) 137 ALR 335
Director of Public Prosecutions (WA) v Mangolamara [2007] WASC 71; 169 A Crim R 379
P (a Child) (1997) 94 A Crim R 593
R v Mackay (Unreported, WASCA, Library No 970689, 10 December 1997)

McKECHNIE J:

Introduction

In the 2006 Census 1,959,088 persons were counted in Western Australia, of whom 49.8% were males. Of the total population in Western Australia, 3% were indigenous persons, including 27,873 Aboriginal males.
An extract from the DPP database of all DSO applications filed at the Supreme Court, up to and including 28 November 2007, indicates that of 15 applications so far filed, seven relate to indigenous respondents. Five of those indigenous respondents, including the respondent, are from remote indigenous communities. In other words, although the indigenous population of the state is 3%, the indigenous population represented in DSO applications is 46.6% with 33.3% being remote indigenous Australians.
The Director of Public Prosecutions (DPP) applies for a final order pursuant to the Dangerous Sexual Offenders Act 2006 (WA) (DSO Act) s 17 that the respondent, when not in custody, be subject to a supervision order. In deciding whether to make an order the paramount consideration is the need to ensure adequate protection of the community.
The jurisdiction to make an order arises if the court finds that the respondent is a serious danger to the community. The DPP has the onus of satisfying the court, by acceptable and cogent evidence, to a high degree of probability, that there is an unacceptable risk that if the respondent is not subject to a continuing detention or supervision order the respondent would commit a serious sexual offence.
For reasons that will appear, the respondent has a Children's Court conviction. It will be necessary to refer to this in detail. In order to preserve anonymity I have changed references to the respondent in documents to GTR.

Background of the respondent

The respondent's birth date is variously recorded on official documents. The Children's Court arrest complaint and Community Justice Services court report dated 10 December 2002 show a date of birth 20 May 1985. The criminal history shows a date of birth of 1 January 1986.
The respondent was born in the Blackstone Aboriginal Community in Western Australia situated between the Western and Great Victorian Deserts. The home of the Papulankutja Aboriginal Community it is approximately 200 kms north‑east of Warburton and only 100 kms from the South Australian border. The respondent is from a traditional background and speaks Ngaanyatjarra with English as a second language. When growing up he moved around the Ngaanyatjarra communities. When he was about 12 his father died and his mother moved back to the family's home land of Warracoona where she still lives. The Warracoona community is nestled in the Rawlinson Ranges near the Western Australian and Northern Territory border approximately 330 kms west of Uluru with a population of approximately 250 people.
In short, the respondent is an Aboriginal person from a remote community.

The significance of a Children's Court conviction

An important question arises at the outset in respect of the criminal history of the respondent. The question arises because of offences allegedly committed on 12 May 2002. In due course, after trial, the respondent was convicted by the President of the Children's Court and was sentenced to a term of detention. Each psychiatrist, appointed by the court to provide an assessment of the respondent, and a number of report writers from Corrective Services Department, have made extensive reference to this conviction. The conviction and, to an extent the circumstances of the offending, underpin their conclusions. The issues which arise are the extent to which the court can have regard to
(a)the conviction, or
(b)the behaviour that gave rise to the conviction,
in making an assessment whether the respondent is a serious danger to the community.
The question arises because of the provisions of the Young Offenders Act 1994 (WA). Regrettably, neither counsel had turned their minds to the issue and so I was not assisted in the resolution of the question to the extent which should be expected.
Amongst other matters, the DSO Act s 7(3) requires me to have regard to whether or not there is any pattern of offending behaviour and the respondent's antecedents and criminal record. It is necessary for their evaluation to know the extent to which the Children's Court record or the corresponding behaviour can be taken into account. The Children's Court conviction also underpins each psychiatric assessment.

(a) Is evidence of the conviction admissible?

The objectives and principles of the Young Offender Act as set out in s 6 include enhancing and reinforcing the role of responsible adults, families and communities in rehabilitating young offenders who have committed offences towards the goal of their becoming responsible citizens and integrating young persons who have committed offences into the community.
By the Young Offenders Act s 189 certain offenders are to be regarded as not convicted. Section 189(2) provides:
If a young person is convicted of an offence and a period of 2 years has expired since -
(a)the discharge of any sentence imposed as a result of the conviction, or every sentence if more than one sentence was imposed; or
(b)the date of conviction, if no sentence that required to be discharged was imposed as a result of the conviction,
the conviction is not to be regarded as a conviction for any purpose, except as provided in this section.
Section 189(7):
(b)any subsequent proceedings that may be taken against the offender under this Act or on indictment in relation to the offence to which this section applies or for a subsequent offence; or
Section 189:
(9)Part 3 of the Spent Convictions Act 1988 has effect in relation to a conviction that, under this section, is not to be regarded as a conviction as if it were a spent conviction under that Act.
Section 190:
(1)If section 189 provides that a conviction of an offence is not to be regarded as a conviction, evidence of that conviction is not admissible in any proceedings, other than proceedings under this Act or on indictment for the offence or for a subsequent offence.
(2)A person, not being the person convicted, who discloses the fact of a conviction that, under section 189, is not to be regarded as a conviction commits an offence except if the disclosure is for the purposes of this Act or of any court of law.
Proceedings under the DSO Act are not proceedings for a subsequent offence within the meaning of the Young Offenders Act s 189(7) or s 190(1).
Under the Young Offenders Act s 189(9) attention is directed to pt 3 of the Spent Convictions Act 1988 (WA) which is entitled 'Effect of a conviction being spent'. There follow a number of exceptions. Division 2 s 14 provides:

Proceedings in courts not affected by Division 4
(1)Nothing in Division 4 affects -
(a)the procedure of, or evidence admissible in, proceedings of a court or tribunal that applies the laws of evidence or proceedings under section 6; or
(b)the Commissioner of Police acting under section 7.

(2)Without limiting subsection (1) -
(a)sections 25(2), 26(1) and 27 do not apply in proceedings of a court or tribunal referred to in subsection (1)(a) or proceedings under section 6;
(b)section 25(1) does not apply in a court or tribunal for the purpose of -
(i)the determination of the guilt or innocence of a person charged with an offence where a conviction is relevant to that determination; or
(ii)a determination of the appropriate punishment to be imposed by that court or tribunal for an offence.

(3)A court, tribunal or judge that receives evidence of a spent conviction shall take such steps as are reasonably available to avoid or minimise publication of that evidence.

A conviction subject to s 189 of the Young Offenders Act is 'automatically' spent so provisions in relation to serious convictions and lesser convictions under s 7 of the Spent Convictions Act do not apply.
Under the Spent Convictions Act s 14(1) a court deciding a dangerous sex offender's application is a court that applies the law of evidence as defined in s 14. The DPP points to this section as a source of power for the admissibility of the Children's Court conviction. However, s 14 does not make admissible that which the Young Offenders Act s 189(2) and the effluxion of time have rendered inadmissible. The Spent Convictions Act s 14 removes the shield of a spent conviction in limited circumstances, none of which apply in the present case.
The words of the Young Offenders Act s 190 are clear and unambiguous in their prohibition of evidence of the conviction in proceedings under the DSO Act because these are not proceedings on indictment.
In P (a Child) (1997) 94 A Crim R 593, the court considered that proceedings for a special order under the Young Offenders Act s 126 came within the expression 'other than proceedings under this Act …' in s 190(1). Section 126 is found within div 9 which allows for an indeterminate sentence to be passed.
The court also held at 598 that the Criminal CodeAmendment Act (No 2)1996 (WA) s 4(3) which provided for the avoidance of doubt that s 400(3) of the Code applies in respect offences committed at any time and to convictions recorded at any time cannot, in the absence of clear words, be taken to override the clear command of s 189 of the Young Offenders Act that a conviction falling within its provisions 'is not to be regarded as a conviction for any purpose'. The court concluded that the Amendment Act can and should be read together with s 189(2) of the Young Offenders Act with the effect that 'convictions recorded at any time' referred to in the former Act do not encompass those excluded by the operation of s 189 of the latter Act.
In R v Mackay (Unreported, WASCA, Library No 970689, 10 December 1997), Steytler J (Kennedy and Franklyn JJ agreeing), described the Young Offenders Act s 190(1) as an evidentiary provision that does not affect the operation of s 189. He said:
[T]he mere fact that evidence of prior convictions falling within s 189(2) might be given in proceedings under the Young Offenders Act or on an indictment for the offence or for a subsequent offence cannot mean (whatever s 190(1) may mean) that the conviction in respect of which evidence is to be given is no longer to fall within the provisions of s 189(2). That section provides, in terms, that the convictions there referred to are not to be regarded as convictions for any purpose, except as provided in s 189 itself. (8)
The inevitable result of the Young Offenders Act is that the respondent's conviction in 2004 is not to be regarded as a conviction for any purpose under the DSO Act.

(b) Is evidence of the offending behaviour admissible?

The issue whether the respondent's behaviour which led to the conviction is admissible is harder. Clearly the intention of the Young Offenders Act is to relieve a young person of the consequences of their offending, after a period of two years, including serious sexual offences. There is little point in doing so if everything but the formal record of the court is admissible in subsequent proceedings not on indictment.
It does not seem right that a judge's findings could be admissible but the resulting conviction is not. On the other hand, the provisions of the DSO Act s 7 clearly intend a judge to have regard to a respondent's history in making a prediction as to future risk. However, the history to be taken into account must be tendered by admissible evidence: DSO Act s 40, s 42.
Counsel for the respondent argued that the Young Offenders Act s 189 confers a right on the respondent. He submitted that where there is a right of significance to the individual, and where there may be a doubt as to Parliament's intention, the court should favour an interpretation that safeguards the individual: Buck v Comcare (1996) 137 ALR 335 per Finn J at 340.
While I can easily accept this statement of principle, it provides little assistance in the resolution of this case because the statute that arguably impinges on the right (the DSO Act) has as its dominant purpose the protection of the public and so a clash with a private right is precisely what Parliament intended.
I have searched in vain through the explanatory memorandum, and the second reading speech in Parliament for a reference as to whether Parliament had considered the behaviour of a respondent when a child and whether that could or should be taken into account.
There is no easy way to reconcile the implicit but very clear intention of the Young Offenders Act to rehabilitate young offenders and the explicit requirement in the DSO Act to have regard to the criminal history and antecedents of the respondent. Even within Young Offenders Act s 189, the prohibition on reference to convictions is not absolute. Past convictions may be referred to in further proceedings on indictment. Moreover, the nature of a special order, as already referred to, is another example where a conviction may be taken into account.
What is evidence of the conviction? The formal record of conviction; the proceedings; or, that in the judge's remarks at the time of sentence? Although he did not read the transcript, Dr Tanney accepted the need to see what facts the judge found about the offence in making an evaluation. Clearly, the factual context of past offending is important information in the prediction of future risk. It may be that some facts are disputed by an accused and a judge must make findings. It is these findings that are then authorative, not any party's contentions about the facts.
Necessarily, to take into account a conviction is to take into account the accompanying behaviour. Discerning as best I can, the presumed Parliamentary intention underlying the Young Offenders Act and the DSO Act, I conclude that except in the limited circumstances provided under the Young Offenders Act s 189, if more than 2 years has elapsed, no evidence of surrounding circumstances - ie the behaviour leading to the conviction - can be given in a court. To hold otherwise is to make effectively meaningless the rehabilitative policy behind Young Offenders Act s 189 and s 190. To do otherwise would involve a court in a highly artificial process. In one sense, a conviction is simply the formal record of the court. A conviction is now defined: Criminal Procedural Act s 2(a), s 147; Children's Court Act s 37(2)(a). In 2002 the law was to similar effect. However, this is a narrow view of a conviction. In a practical sense, a conviction includes the charge upon which the conviction is based, together with all the evidence led in support of that conviction, the reasoning supporting the verdict (where, as here, there is no jury trial), together with the verdict upon which the judgment of conviction is based.
As a result, the following exhibits are inadmissible: Exhibit 1, page 2; Exhibits 3, 4, 5 (coupled with Exhibit 35), 6, 7, 8, 9, 10, 11 and 12. Also inadmissible are references to the Children's Court conviction in Exhibit 21, 22, 26, 27, 28, 29 and 30.
In addition, references to the Children's Court conviction in Dr Brett's report (Exhibit 31) and Dr Tanney's report (Exhibit 32) are inadmissible.

Conclusion on application without reference to the Children's Court conviction

I hold that the 2002 Children's Court conviction, together with all evidence of the offence is inadmissible in these proceedings. The result is that the application must fail. The impact of the offending behaviour in the formation of the opinions of each psychiatrist, and of other report writers, is such that their opinions are fatally compromised. Moreover, there is no 'pattern of offending'. Taking into account all matters under the DSO Act s 7(3) I am not satisfied to a high degree of probability that the respondent is a serious danger to the community. On that basis the application is refused.

Application considered on assumption that evidence of Children's Court conviction is admissible

However, I am acutely conscious that the relationship between Young Offenders Act and the DSO Act has not previously been considered by a court and, as I have said, I have not had much assistance from counsel in deciding the point.
Counsel for the DPP concedes that the conviction is a spent conviction and not admissible, although he contends that evidence of the underlying behaviour is. In these circumstances, I think it best to determine the application on that basis so that if my conclusion about the Young Offenders Act turns out to be wrong, the parties will have the benefit of my consideration of the whole of the evidence.

The evidence before the court

Following the practice of the court, at the time of the preliminary hearing, the judge ordered reports from two psychiatrists whose names were available from a panel. They were Dr Brian Tanney and Dr Adam Brett. Each gave oral evidence and was cross‑examined at the hearing of the application. So also was Ms Ostaszewskj, a senior Community Corrections Officer with Community Justice Services based at Kalgoorlie. She has not met the respondent but has undertaken enquiries about the feasibility of a supervision order and has recommended certain conditions that might be imposed in a supervision order. She did not purport to give any evidence as to the assessment of the risk of re‑offending.
At the hearing, the DPP tendered a book of evidence, and subsequently other documents were tendered. Each document within the book was marked as a separate exhibit. No attempt was made by counsel for the DPP to direct the court's attention to any particular matter within the material or to discharge the high onus resting on the DPP by reference to any particular exhibit. At the conclusion of the hearing, I was left in the very unsatisfactory situation of having the oral evidence of two psychiatrists, supplemented by their reports, together with a quantity of other material without the benefit of any analysis by the parties. Apparently the DPP thinks it is sufficient to discharge its high onus by serving up to the court, in undigested form, a mass of material and leaving the court to sort it out as best it may. As a result, I am left entirely unassisted.
There are questions legitimately raised by the respondent as to the voluntariness of the respondent's records of interview. He is a juvenile tribal indigenous person. At his Children's Court trial, the prosecution did not tender the record of interview as proof of guilt. Now the DPP seeks to rely on it. Moreover, there are real questions as to what evidence is admissible in a hearing such as this. I have had little assistance in resolving these questions.
The preparation of psychiatric reports and the provision of information to psychiatrists is dealt with under the DSO Act pt 5. While pt 6 deals with evidence, s 40 provides that the proceedings are to be taken to be criminal proceedings for all purposes. The exact content of this section is unclear. For example, the criminal burden of proof does not apply to the formation of an opinion under s 7. Section 7 provides a lesser standard; namely high degree of probability. However, the high degree of probability must be reached by acceptable and cogent evidence.

It is significant that at a preliminary hearing, s 41(2) provides that the court may decide on the documents filed whether there are reasonable grounds for believing that a court may make an order under s 7. The objective test of reasonable grounds for belief is a common enough concept in criminal law and applies extensively, for example, under the Criminal Investigation Act 2006 (WA). Formal evidence is not generally necessary for the formation of reasonable grounds. So also a decision to amend a supervision order which has already been made - that is, a judge has formed a concluded view as to serious risk - may be done on the papers.
However, a different and more formal standard of evidence applies at a final hearing. The less formal requirements of s 41 may be contrasted with the requirement for a hearing for a div 2 order under s 42 which provides:
…
(2)Before the court makes a decision or order on the hearing of an application it must, if the evidence is admissible -
(a)hear evidence called by the DPP; and
(b)hear evidence given or called by the offender or person subject to the order, if that person elects to give or call evidence.
(3)Except as modified by subsection (4), ordinary rules of evidence apply to evidence given or called under subsection (2).
(4)In making its decision, the court may receive in evidence -
(a)any document relevant to a person's antecedents or criminal record;
(b)anything relevant contained in the official transcript of any proceeding against a person for a serious sexual offence, or contained in any medical, psychiatric, psychological or other report tendered in a proceeding of that kind.
The respondent did not object to the formal tender of the documents contained in the book of evidence which were given separate exhibit numbers. I understood from this that the documents were what they purported to be and it was unnecessary to call the writers to prove them or the facts contained within them.
In these circumstances, I therefore assume that documents, such as pre‑sentence reports and psychological reports, are tendered under s 42(4)(b), while documents relating to the respondent's periods in custody, prepared for various purposes, including parole assessments and sex offender risk updates, fall within s 42(4)(a).

Is the respondent a serious danger to the community?: DSO Act s 7

In deciding whether to find that the respondent is a serious danger to the community, I must have regard to 10 enumerated matters set out in the DSO Act s 7(3). I will deal with each in turn.

Section 7(3)(a)

any report that a psychiatrist prepares as required by section 37 for the hearing of the application and the extent to which the person cooperated when the psychiatrist examined the person;

Counsel for the respondent challenged the expertise of Dr Tanney and Dr Brett to predict risk of sexual re‑offending and cross‑examined each doctor to support the submission. However, the answers that each gave satisfy me that each has long experience in the general fields of risk assessment and risk management and that each has particular experience in assessment of sex offenders, including offenders who might fall within the purview of the DSO Act.

Dr Brian Tanney

Dr Tanney provided a report dated 6 November 2007 (Exhibit 32). In preparation of the report Dr Tanney noted:
The critical aspect of this report involves a determination of risk of recurrence of a serious sexual offence. The ability to predict unacceptable human behaviour with any accuracy and reliability has been the subject of considerable academic and clinical interest over the past several decades. In particular, the expertise of psychiatrists and other behavioural scientists in prediction of violence has been significantly questioned. In response, a brief of worthy research has been accumulating, largely directed towards the prediction of violence and including sexual violence. From this work, several processes for enabling risk prediction have been identified: unstructured clinical assessment, guided clinical risk appraisal, and actuarially derived, standardised instruments. Each of these processes has inherent procedural defects requiring consideration when applying and interpreting results and conclusions drawn by their application.
Notwithstanding the above, the preparation of such reports should embrace the variety of procedures available. In the following, I report the results of actuarially derived risk ratings for [GTR] utilizing the Risk for Sexual Violence Protocol (RSVP) 2003 and the Static‑99 measures. As indicated, I also conducted a highly‑focused risk appraisal interview that was guided by my understanding of relevant and important constructs that have been identified in enabling risk prediction and management. The final opinion is derived from my integration of the data provided by these various informing processes.
This process of synthesis is not without its own idiosyncratic defect. Synthesis and amalgamation of data must always be created in the context of the person undertaking this work process. The process can never be entirely objective. Even for the actuarially derived instruments that strive for objectivity, efforts towards standardisation and reliability of ratings cannot be guaranteed, although clearly desired. In particular, the actuarially derived instrument RSVP does not provide any rule or process for summating a low, medium or high category of prediction risk from the twenty‑two, highly specified variables. Accepting that this data synthesis involves clinical expertise and experience, I believe it important to declare at outset the subjective framework or attitudinal perspective of any person preparing such reports.
He continued further:
I can state that this report was prepared and an opinion formulated with the perspectives that information from interview should be valued equally with that of actuarially derived instruments, and that a relatively low standard of certainty for the likelihood of recurrent serious sexual offending would be implicit in my considerations.
Dr Tanney administered the Static‑99 assessment. The respondent scored 7. Scores of 6 and over derive the same 'high' risk rating for recidivism of sexual or violent offending. With this instrument, a high re‑offending likelihood is still associated with a less than 50% probability of sexual re‑offending, as concluded from the original data set.
Dr Tanney noted:
This instrument has only moderate predictive accuracy and does not include the factors that might be considered in a dynamic risk assessment. The applicability of Static‑99 to populations dissimilar to those from whom the instrument was created has been challenged, especially for Aboriginal persons. The issue is unresolved, but there is interpretable data applying it to Western Australian Aboriginal offenders.
The interpretable data referred to by Dr Tanney was not tendered in evidence.
Dr Tanney agreed that a further limitation of Static‑99 is that it is directed to sexual re‑offending of any kind rather than to the commission of serious sexual offences. Dr Tanney acknowledged that actuarial processes have an inherent flaw, which is using group data to predict an individual's behaviour. He did not concede that Static‑99 was of no utility for an indigenous population but accepted that it was of less utility. He said that Static‑99 tells nothing about positive changes exhibited by an offender over time.

Risk for Sexual Violence Protocol 2003 (RSVP)

This actuarially derived but clinically interpreted instrument makes an effort to estimate the likelihood of actual attempted threatened sexual contact with a person who is non‑consenting or unable to give consent. It is an updated version of the SVR‑20 and is recommended and used in other Australian states. Twenty‑two items are available for scoring and a further category of other/idiosyncratic considerations is available. The instrument affords some dynamic representation of risk in two ways. It requires comment respecting the future relevance of each risk factor for the development of the risk management plans.
It is notable that the purpose of RSVP as defined by Dr Tanney is:
To elaborate a comprehensive risk management plan that addresses every item indicating risk, this allows as many items as possible to be included in the planning process. There is no weighting of risk quantity based on a specific number of coded yes items.
There is no coding guide or translation matrix for a summary score to be created based on particular numbers or patterns of risk factors. It is designed to further the development of specific risk management strategies that might lessen the likelihood of sexual re‑offending.
It seems to me that one weakness of the RSVP in prediction of risks is the weakness identified in the cross‑examination of Dr Tanney:
[W]hen you read the coding manual … for the RSVP, one of the things it includes is he may have appropriate plans - he may have plans that are feasible, but can they be put forward because of other restrictions that he perhaps might have no control over; like geography or other controls. That's where you have to still give it as a 'yes', as a possible risk issue. (ts 57)
And further:
Each of those last three items in that cluster are the RSVP: 20, 21, and 22 have a coding requirement that the plans or the supervision or the treatment may have - he may have cooperated with it all, but if there is a possibility that it might not be able to be provided wherever he's going to be, one still has to put it as a potential risk factor but it doesn't speak at all to him; it speaks to the system in which he operates.
…
They all get a 'yes' or a 'possible', simply because they might not be available. Even though what he is prepared to do and what he would propose to do would be acceptable, if it's not possible to happen, the requirement of the RSVP protocol is that you indicate that it is something to be considered as a risk management issue.
This lends force to the view that RSVP may have a valid use for risk management but may be an imperfect tool for risk assessment as required under the DSO Act.

3‑Predictor model

West Australian investigators have proposed a locally developed instrument for prediction of violent re‑offending, developed with the sexual offender population of inmates. Their work also offered some tentative validation of applicability of the Static‑99 to this population. They propose three dynamic factors that were linked to likely violent and sexual re‑offending in a group of male, Western Australian indigenous sex offenders. This instrument appears to have some value for estimating the likelihood of risk for such re‑offending in the respondent.
Again, the 3‑Predictor model has presently not been validated. Therefore, the court is being asked to rely on opinions of psychiatrists when those opinions are based, in considerable part, on tools that are not validated. There is a possible error of logic to proceed on the basis, as Dr Brett suggests, that even though the tools are unvalidated, when the three give a similar response, there is some validity or relevance in the result.
The 3‑Predictor model factors are unrealistic long term goals, unfeasible release plans, and poor coping skills prior to release.
Dr Tanney notes:
[GTR] has evidenced difficulties in coping in an appropriate way with strongly emotional situations involving important others. This was in part related to immaturity and lack of experience. He has expressed considerable self‑awareness and the capacity to utilize the relapse prevention strategies he has learned as a template for avoiding exposure to situations of increased risk for him.
He has not consistently evidenced poor coping skills prior to release. This item is difficult to specify as his general coping may be adequate, but his vulnerability in specific situations of emotional stress may disable his general functional competency. The opportunity to demonstrate his application of reinforced relapse prevention strategy is to avoid these specific situations has not been fully tested.
Dr Tanney concluded:
On the 3‑Predictor model, GTR is NOT at increased likelihood of reoffending violently and sexually.
To the extent then that the 3‑Predictor model does have some usefulness, its results as interpreted by Dr Tanney weaken, to some degree, the assessment of the respondent as a serious danger to the community.
Dr Tanney's observation at interview was that there was no evidence for major mental disorder symptomology. General screening probes and observations for psychosis, anxiety disorders and major mood disorders were negative.
Dr Tanney's opinion is:
1.Without ongoing management (including monitoring and possible treatment) and the implementing of certain risk diminishing measures, this man is at some risk of further serious sexual offending, as estimated by available procedures and measures.
…
2.Both treatment efforts and his Banksia Hill learning experience suggest that he is able to understand and to intend to change his behaviour, both generally and in respect of preventing reoffending.
3.The objects of his sexual offending behaviour have been older, adult women. It can be supposed that the probable object(s) of any further sexual reoffending would be adult women.
4.Both episodes of serious sexual offending have incorporated violence with some indication of escalation. The linkage between negative emotions such as anger and sexual arousal is a dangerous pairing. It should be concluded that further sexual offending is likely to involve violence. There appears to be no general predisposition towards violence.
5.I am unable to speculate on the possible time course of further offending beyond the risk values specified above in Static 99 as 'within 5 years'. This inability to specify a time frame for recurrent offending is a major failing of most risk prediction science. In general, the categorical expression of risk likelihood as Low, Moderate and High is taken to mean an outcome of recurrent or similar offending that can be construed as indicating time until re‑offending (relapse). Using this convention, 'High' likelihood would indicate shorter time until relapse of offending behaviour, although without specific temporal anchor points.
6.For [GTR], his risk scenario for likely offending is quite specific and has been well described. Additionally, relapse prevention strategies are available for this scenario.
Finally, Dr Tanney noted in his conclusions:
1.This report has made effort to meet the purpose of section 37 of the DSO Act 2006, but the scientific underpinnings to satisfy the standard of predicting with high probability the likelihood of serious sexual re‑offending by any offender are incomplete at this time.
2.Despite this, three instruments derived from an actuarial or guided clinical risk appraisal framework have been realized by myself. Their disparate prediction of reoffending likelihood reflects the consideration of static versus dynamic (modifiable) aspects of his risk. There is support for the opinion that [GTR] is at significant, even 'High', likelihood of serious sexual re-offending against adult women with accompanying violence within the next five years or sooner, should no further intervention take place. 'High' is taken to mean re-offending rates approximating 40% at five years. [This likelihood estimate informs, but does not constitute, a direct probability of personal re-offending nor does it address the ascertaining of 'unacceptable risk'].
3.There are protective factors in his social, personal and care giving environment. [GTR's] ability and demonstrated willingness to utilize such resources represents the strongest dynamic risk element for minimizing his risk of re-offending. The risk elements that are likely to exacerbate the likelihood of further serious sexual offending are substance abuse, lack of appropriate social roles and a diminished ability to regulate his emotional state when negative affects are present. With enforcement of the proposed supervision, each of these dynamic risk elements seems capable of alteration or change in a direction that would diminish the likelihood of re‑offending.
4.Any agency charged with implementing a Community Supervision order will be severely challenged to afford adequate monitoring of unacceptable risk and to offer community‑based rehabilitation services sufficient to lessen the likelihood of re‑offending. This is largely due to geographical isolation. Relapse prevention strategies and the development of social and personal protective competencies are feasible and possible within the currently proposed release plan.
In cross‑examination Dr Tanney confirmed that he had used the various tools together with a guided clinical assessment. He noted that when a young person offends at an extremely young age they may be reflecting a process that is well‑entrenched. He accepted there is a body of evidence in that regard and there is a great margin for error; but he would submit in either direction.
Dr Tanney's approach is partially encapsulated in the following:
The legislation was - and has had teething difficulties. The requirement and the request for psychiatric input around specifically the issue of risk management and assessment of risk brought forward all of the issues that you point out, and perhaps other. I think it's remarkably important to acknowledge, in any report that one prepares, that there has to be - unless it's done 100 per cent on the numbers actuarially, there has to be an element of subjectivity in the objectivity, and that part of what I would have said I bring as a psychiatrist to that program is that that's one of the things that I'm trained to do. So when I look at a young man like [GTR], I try to take into account, sometimes in a positive way and sometimes to discount things that might influence me that I don't think really have anything to do with objectivity. But I have to appreciate that they're there. One of these is what's happened in the last year as we attempt to define what risk is and, in my context, what risk management is, because I think the two are linked; and what I've unfortunately seen in the past year is that the ability that we've been able to put forward to provide the requirements to appropriately manage risk in the community, such that it might be regarded as acceptable risk, has been something where in previous situations we've let people down, and I have to acknowledge that that's had an influence on my creation and perspective in creating reports. (ts 52)
Dr Tanney amplified this in re‑examination:
My interest here was in coming to the court and providing a measure of the likelihood of risk but couched in that is always the idea that when you assess risk you have to see it in the context of scenarios. You know, there is a certain quantum of risk, but we can diminish or change that risk under certain conditions. That's what I'm interested in, is how one manages that risk. So when I have to create my assessment of the quantum of risk, it has to be influenced by the awareness of whether that risk can be modified in any way; and there was an expectation, I think, over the past year, and inexperience, that several of us have had - and I think that's been reflected in discussions in the court - that that has not been as supportive as it could have been in helping to manage the quantum of risk that was there; and so I could not - - -
Are you saying that, in practical terms, the risk assessment has been proven not to be - - -?---No, the risk assessment, I think, hasn't changed. The ability to manage that risk has proven to be not as helpful as we would have wanted it to be in the community, and, I think, as a result of that I've shifted a bit, and I'm not sure about my colleagues, towards a more conservative perspective. (ts 61)

It would be troubling if an assessment of unacceptable risk and serious danger to the community is made on the basis that the state is unable to properly provide resources for a supervision order that would otherwise be appropriate to enable the person to live in the community. A judge approaching the decision under s 7 should do so on the basis that any supervision plan that is proposed will be delivered. To do otherwise would be to invite a finding not that the person is a serious danger to the community but that the lack of resources to provide supervision creates the danger.
Dr Tanney noted that Static‑99 tells nothing about positive change exhibited by an offender over time. He agreed that when things like psychopathy, major mental illness and sexual deviancy are absent, especially in the case of a very young person, that these are highly significant matters in favour of a positive prognosis. The respondent has no such mental illness or disorder.

Dr Adam Brett

Dr Brett provided a report dated 25 October 2007 (Exhibit 31). He noted under Psychiatric Diagnosis:
Clinical Disorders
Nil significant.
Personality Disorders and Mental Retardation
Possible cognitive impairment.
General Medical Conditions
Nil known.
Psychosocial and environmental problems
[GTR] did not appear to be too distressed by his current situation or future plans.
Global Assessment of Functioning
[GTR] would score in the severely impaired range on his global assessment of functioning.
Dr Brett used the same risk assessment tools as Dr Tanney noting:
It should be noted that the risk assessments used have not been validated in the Indigenous population. A risk assessment process has been used which has been trialled in a Western Australian Indigenous population (the 3‑predictor model). These tools make the most sense to be used and if all tools point towards the same answer they have increased weight.
I doubt that the last sentence is an expert opinion. In any event, one conclusion is that they have increased weight. Another may be that all three are, for various reasons, wrong. Each is designed to assist in the prediction of sexual or violent recidivism in different ways but when all get unvalidated they may reflect different errors albeit leading to a common result.
Dr Brett's administration of Static‑99 returned a score of 7, the same as Dr Tanney. This is an actuarial assessment and so the result is unsurprising.

The 3‑Predictor model

Dr Brett appears to have come to a different conclusion from Dr Tanney. He said:
This model was proposed from some Western Australian research in Indigenous sex offenders. The three items are all dynamic and relate to poor coping skills, unfeasible release plans and unrealistic long‑term goals. It appears that [GTR] has difficulties in making plans for himself. He has a history of poor coping skills. His release plan appears feasible but he has had limited input into that as far as I am aware. The community where he is planning to live is accepting of him, however, there are no facilities available there to address his offending behaviour. I understand that his long‑term goals are unrealistic with respect to relationships. He had few long‑term goals, which he articulated to the author.
Dr Brett sets out the headings for RSVP and comments thereafter in relation to them.
Dr Brett's ultimate conclusion:
[GTR] appears to be a significant risk of sexual offending in the future. This conclusion is reached from an individual assessment of [GTR], review of collateral information and use of risk assessment tools, the STATIC 99, the Risk for Sexual Violence Protocol and the Three‑Predictor Model. All these tools support the clinical opinion that [GTR] remains a significant risk to others. This risk can be reduced by putting in place strict behavioural controls including abstinence from marijuana and limited access to females and also with the Community being aware of his offending history. [GTR] does not appear to be a candidate for any pharmacological interventions. He has completed some Sex Offender Programs but he does not appear to be psychologically minded and the mainstay of his treatment should be behavioural.
In cross‑examination Dr Brett said that he found Static‑99 an extremely useful tool. It is the foundation of risk assessment and management. He agreed that it is not validated for use on indigenous Australians. It takes no account of reformation and gains made over the years in respect of a child offender or any young offender. He agreed that the Static‑99 is not of utility in determining whether somebody is ready for release or discharge. It would be a very bad tool for that because the score is unlikely to change. He acknowledged the inherent flaw transferring the probabilities of re‑offending within a group to an individual setting. He said:
The three tools work in three very different ways. They have different methods of evaluating them. So, because they are very different, it would be very unlikely that the same error would be in each of the tools. (ts 67)
He acknowledged that with the 3‑Predictor Model that it was not a validated tool and there has been very little research on it. He indicated a similar point to Dr Tanney that the offending at a young age may mean the earlier age of sex offending the worse the prognosis. The risk of recidivism is increased.
Dr Brett was challenged as to his assessment but firmly maintained that his report can be relied upon with confidence.

Publications about the three tools

The coding rules for Static‑99 were referred to by both psychiatrists who were trained in their application.

Static‑99 Coding Rules Revised - 2003

The Nature of the STATIC-99
The STATIC-99 utilizes only static (unchangeable) factors that have been seen in the literature to correlate with sexual reconviction in adult males. The estimates of sexual and violent recidivism produced by the STATIC‑99 can be thought of as a baseline of risk for violent and sexual reconviction. From this baseline of long-term risk assessment, treatment and supervision strategies can be put in place to reduce the risk of sexual recidivism.
The STATIC‑99 was developed by R Karl Hanson, Ph.D. of the Solicitor General Canada and David Thornton, Ph.D., at that time, of Her Majesty's Prison Service, England. The STATIC‑99 was created by amalgamating two risk assessment instruments. The RRASOR (Rapid Risk Assessment of Sex Offender Recidivism) developed by Dr Hanson, consists of four items: 1) having prior sex offences, 2) having a male victim; 3) having an unrelated victim, and 4) being between the ages of 18 and 25 years old. The items of the RRASOR were then combined with the items of the Structured Anchored Clinical Judgement - Minimum (SACJ‑Min), an independently created risk assessment instrument written by Dr Thornton (Grubin, 1998). The SACJ‑Min consists of nine items: 1) having a current sex offence, 2) prior sex offences, 3) a current conviction for non‑sexual violence, 4) a prior conviction for non‑sexual violence, 5) having 4 or more previous sentencing dates on the criminal record, 6) being single, 7) having non‑contact sexual offences, 8) having stranger victims, and 9) having male victims. These two instruments were merged to create the STATIC‑99, a ten‑item prediction scale.
The strengths of the STATIC‑99 are that it uses risk factors that have been empirically shown to be associated with sexual recidivism and the STATIC‑99 gives explicit rules for combining these factors into a total risk score. This instrument provides explicit probability estimates of sexual reconviction, is easily scored, and has been shown to be robustly predictive across several settings using a variety of samples. The weakness of STATIC‑99 are that it demonstrates only moderate predictive accuracy (ROC=.71) and that it does not include all the factors that might be included in a wide‑ranging risk assessment (Doren, 2002).
…
The STATIC-99 does not address all relevant risk factors for sexual offenders. Consequently a prudent evaluator will always consider other external factors that may influence risk in either director. An obvious example is where an offender states intentions to further harm or 'get' his victims (higher risk).
Or, an offender may be somewhat restricted from further offending either by health concerns or where he has structured his environment such that his victim group is either unavailable or he is always in the company of someone who will support non‑offending (lower risk). These additional risk factors should be stated in any report as 'additional factors that were taken into consideration' and not 'added' to the STATIC‑99 score. Adding additional factors to the STATIC‑99, or adding 'over‑rides' distances STATIC‑99 estimates from their empirical base and substantially reduces their predictive accuracy.
…
STATIC-99 with Juvenile Offenders
It should be noted that there were people in the original STATIC‑99 samples who had committed sexual offences as juveniles (under the age of 18 years) and who were released as adults. In some cases an assessment of STATIC‑99 risk potential may be useful on an offender of this nature. If the juvenile offences occurred when the offender was 16 or 17 and the offences appear 'adult' in nature (preferential sexual assault of a child, preferential rape type activities) - the STATIC-99 score is most likely of some utility in assessing overall risk.
Evaluations of juveniles based on the STATIC‑99 must be interpreted with caution as there is a very real theoretical question about whether juvenile sex offending is the same phenomena as adult sex offending in terms of its underlying dynamics and our ability to affect change in the individual. In addition, the younger the juvenile offender is, the more important these questions become. In general, the research literature leads us to believe that adolescent sexual offenders are not necessarily younger versions of adult sexual offenders. Developmental, family, and social factors would be expected to impact on recidivism potential. We have reason to believe that people who commit sex offences only as children/young people are a different profile than adults who commit sexual offences. In cases such as these, we recommend that STATIC-99 scores be used with caution and only as part of a more wide‑ranging assessment of sexual and criminal behaviour. A template for a standard, wide-ranging assessment can be found in the Solicitor General Canada publication, Harris, AJR (2001) High-Risk Offenders: A Handbook for Criminal Justice Professionals, Appendix 'd' (Please use the references section).

In a paper published in the British Journal of Psychiatry 2007 at 190 the authors Stephen D Hart PhD, Christine Michie BSc and David J Cooke PhD, FBPsS FRSE, examine the precision of actuarial risk assessment instruments (ARAIs). They note the ARAIs conceptualise violence risk solely in terms of probability of future violence, ignoring other facets of risk, such as the possible nature, severity, imminence, duration or frequency of future violence. They use fixed and explicit algorithms, developed on the basis of data from known groups of recidivistic and non‑recidivistic violent offenders and patients, to estimate the specific probability or absolute likelihood that a person will engage in violence in the future. They note:
There are two major types of error relevant in the case of violence predictions made using ARAIs. The first is group error. …
The second type of error is individual error. Moving the focus of analysis from groups to individuals changes the way in which risk is conceptualised.
The two risk assessments tools examined were the Violence Risk Appraisal Guide (VRAG) and Static‑99.
After examining each to determine risk estimates they concluded:
The potential implications of these findings for the practice of forensic mental health are profound. At best, they suggest that professionals should be extremely cautious when using ARAIs to estimate or draw inferences about an individual's risk of violence. This means, as Henderson & Keiding (2005) have recommended, 'avoiding use of a single quantity to characterise a probability distribution, whether a point or categorical prediction, prognostic index, relative risk, or probability of surviving a given time' (p 705). At worst, they suggest that professionals should avoid using ARAIs altogether, as the predictive accuracy of these tests may be too low to support their use when making high‑stakes decisions about individuals. Low predictive accuracy not only makes reliance on ARAIs ethically problematic, it also means that they may not meet legal standards for the admissibility of expert or scientific evidence.
The authors noted the use of professional judgment to modify or override test‑based decisions quoting Meehl (1998) The Power of Quantative Thinking:
This sounds amicable, tolerant and even‑handed, but it is actually stupid. It does not make sense to fudge the results of a statistically derived estimate on the basis of personal preference; in addition, there is simply no empirical evidence that this improves the accuracy of predictions.
They conclude:
Test users should be very careful when using ARAIs to make sure that consumers of test findings are the mental health professionals, patients, courts etc, understand that it is, at least at present, impossible to make accurate predictions about individuals using these tests; this may help to minimise their potentially prejudicial impact on decision‑making. Also, it may be wise to limit or avoid the use of ARAIs in situations where the cost of potential decision errors is high.
In a paper 'Risk and preventative detention of sex offenders in Australia and the United States', Mercado and Ogloff discuss the recent statutory schemes in the United States and Australia to keep the most dangerous sex offenders detained beyond the expiration of their prison sentence. Dr Ogloff gave training sessions in Static‑99 attended by Dr Brett and Dr Tanney.
The authors note under the heading 'The assessment of risk for violence and future sexual offending':
Historically, mental health professionals were often found to be inaccurate in their estimations of future violence. Indeed, in the oft‑cited finding Monahan (1981) suggested that clinicians were 'wrong two‑thirds of the time' in making predictions of dangerousness, with clinicians exhibiting a marked tendency toward the over‑prediction of violent behaviour.
The authors suggest that 'this spawned a flurry of research activity and scholarship related to base rates of offending, the identification of empirically validated risk factors for recidivism, and the development of a myriad of risk assessment tools'. They note:
Despite the importance of some individual risk markers, such as sexual deviancy or general criminality, no single factor has yet been identified that is sufficiently able to predict sexual recidivism (Hanson et al, 2003). As a result, increasing attention has been devoted to the development of empirically derived risk assessment tools that includes grouping of important risk factors and sometimes formulae for the combinations of such risk factors. Indeed, while some risk assessment instruments provide overall risk scores based upon the combined weightings of a set number of risk factors, such as the Violence Risk Appraisal Guide (VRAG) (Quinsey, Harris, Rice, & Cormier, 1998), Sex Offender Risk Appraisal Guide (SORAG) (Quinsey et al, 1998), Rapid Risk Assessment for Sex Offence Recidivism (RRASOR) (Hanson, 1997) and the Static‑99 (Hanson & Thornton, 2000), others, such as the Sexual Violence Risk‑20 (SVR‑20) (Boer, Hart, Kropp & Webster, 1997) or the Risk for Sexual Violence Protocol (RSVP) (Hart et al, 2003) utilize a structured professional judgment approach that provides decision‑makers with structured guidelines for considering a list of empirically validated factors but does not provide probabilistic estimates of risk based on the combination of such factors. (54)
…
Comparative analysis of the utility of specific risk instruments has been undertaken elsewhere (see eg, Barbaree, Seto, Langton & Peacock, 2001; Harris et al, 2003) and while a full review of the instruments is beyond the scope of this paper, some of these relative strengths of the instruments deserve mention. Although the Static-99, among the actuarial measures mentioned above, has been observed to have the greatest accuracy (Hanson et al, 2003), each appear to have relative strengths based upon the predictive outcome being utilized. A number of studies have found the SVR‑20 to predict sexual offending with moderate to high degrees of accuracy (Craig, Browne & Stringer, 2004; Macpherson, 2003). For example, de Vogel, de Ruiter, van Beek, and Mead (2004) recently tested the SVR‑20 and the Static-99 in a sample of sex offenders in The Netherlands. They found that the predictive validity of the SVR‑20 was very good, particularly when they considered the clinician's final risk judgment after having used the SVR‑20. They also found that the Static‑99 was a moderate predictor of sexual offending. Contrary to other findings, a study with 51 male sexual offenders in Sweden did not find that the SVR‑20 predicted sexual recidivism better than chance (Sjostedt & Langstrom, 2002). In 2003, the Risk of Sexual Violence Protocol (RSVP) (Hart et al, 2003) was developed. It is essentially builds upon the strengths of the SVR‑20, while carefully refining some of the items and adding some additional ones. While promising, at the time of writing, there has been no published data validating the RSVP.
At this point, there appears to be no single instrument in the field with a well‑accepted superior predictive capability. As mentioned, however, each of these instruments has a demonstrated reliability and predictive validity that far exceeds that of clinical judgment. Indeed, utilization of risk assessment measures allows clinicians to make a much more accurate assessment of risk than otherwise possible. The continued validation and development of these measures should foreseeably result in even more enhanced predictive capabilities.
There has been some attempt to validate predictive models in Western Australia.
In December 2003 Dawson and Allan published 'An Exploratory Study to Identify the Predictors of Sexual Reoffending by Male Sexual Offenders in Western Australia'. They concluded:
The importance of the findings of the study is that it demonstrated that it is possible to develop a risk of sexual offending instrument for the WA population that is more accurate than those currently available. However, we believe that a prospective study using the predictors identified during this study on much larger samples is what is required. This will have to be a longitudinal study to ensure a large enough sample and a long enough period at risk, as offenders have less opportunity to reoffend while in custody and may be constrained while subject to community based orders.
In the Executive Summary it was noted:
In the present study the 3‑Predictor model was cross validated and was found to be accurate for both Indigenous and non Indigenous sexual offenders when culturally appropriate definitions were used. As established in previous studies conducted by Allan and Dawson (2002), predictive accuracy declined when the model was applied to violent sexual offenders. Further investigation tentatively revealed that the most accurate way of assessing risk with this specific group was to combine the 3‑Predictor model score with the Static 99 score. This finding, while requiring more rigorous investigation was promising, as previous attempts to develop a model in WA had proved unsuccessful.
…
In summary, the present study has indicated that the 3‑Predictor model can be used with both Indigenous and non Indigenous sexual offenders if attention is given to cultural differences. Violent sexual offenders, however, appear to differ from non violent sexual offenders and a different system of risk and needs assessment is indicated. The combination of the 3‑Predictor model and the Static 99 suggests a promising alternative. The treatment needs of violent sexual offenders also require further investigation and the outcomes of the present study have highlighted potential areas of need.

Section 7(3)(b)

any other medical, psychiatric, psychological, or other assessment relating to the person;

A number of assessments and reports prepared by Corrective Services personnel have been tendered. In most cases their assessment is based upon use of Static‑99 and/or RSVP and/or the 3‑Predictor model. Their facility is therefore limited to that extent. None of them are disconsonant with the conclusions reached by Dr Tanney and Dr Brett. I have taken account of the reports compiled from time to time by Corrective Services. Much of it is harmless recitation of facts or expression of matters, such as participation in courses. Little of it is of much use in resolving this application. If any of it had particular significance to the DPP's case, I would have expected counsel for the DPP to draw my attention to it. He did not.
In a similar vein the decision of the Parole Board to refuse parole in the exercise of its statutory function has very limited relevance in making an assessment in the exercise of my statutory function.
On 7 March 2003, Mr Gary Duggan, a clinical psychologist (Registrar) attached to the Psychological Services, Banksia Hill Detention Centre, Juvenile Custodial Services, prepared a report for the Supervisory Release Review Board. Under 'Presentation and Motivation' he reported:
[GTR] was initially reluctant to engage, however after the importance of the counselling was made clear, he became a willing participant. He was open, honest, and responsible for his actions and was able to clearly recall the offence, which enabled steady progress with counselling.
In summary, Mr Duggan noted:
[GTR] was an active, open, honest and responsible participant in his sexual offending counselling. He identified the advantages, disadvantages and consequences of his behaviour for himself and his victim. He recognised that his lifestyle of unstructured activity, substance misuse, and his misuse, and his inability to communicate effectively and regulate his emotions led to his offence. He displayed empathy for his victim and his family. He was able to identify more effective methods of communication and strategies to prevent escalation of his feelings into aggressive behaviour.
The report writer concluded that:
[GTR] would benefit from further sex offender counselling and substance misuse counselling. Further counselling may assist in the transition of his learning from the custodial setting to the community environment where he will face most of his challenges.
There is no evidence that further counselling was ever provided.
A further psychological report was obtained for the purposes of sentencing after the offences in 2004. This was prepared by Kirstin Bouse, Forensic Psychologist on 2 August 2004. She concluded:
[GTR] is a 19-year-old man who was convicted by his own admission of 1 count of attempted aggravated sexual penetration without consent, 1 count of deprivation of liberty and 1 count of unlawful wounding. These offences involved [GTR] assaulting a 65-year-old nurse who works and resides at the Aboriginal Community where [GTR] himself lives. Further, the offences involve considerable planning and what is considered high‑level cognitive processes, to ensure that he could carry out the offence. [GTR] has prior offences of a similar nature. Throughout interview, [GTR] acknowledged his actions, including the elements of planning. However, he lacked any insight into the factors that may have contributed to his offending and failed to demonstrate any remorse in relation to his behaviour or concern for his victims. [GTR] also reported a history of cannabis abuse and stated that he is 'always wanting sex when stoned'. Given the sexually violent nature of his current offences, the fact that he has prior offences of a similar nature in addition to his history of problematic substance use, [GTR's] treatment needs are considerable. Although he is considered suitable for a sex offender treatment program, the writer stresses the importance of [GTR] participating in a culturally appropriate treatment process. To ensure this occurs, [GTR] would require imprisonment within the Greenough prison. As present, sex offender treatment programs for Aboriginal offenders are not facilitated in the remaining prisons or the community. Actuarial factors place [GTR] as being a high risk of reoffending.
On 8 September 2005 two psychologists (Ms M Dorazio and Ms J Davies) completed a Treatment Completion Report (Prison). They noted:
[GTR] completed the Sex Offender Indigenous Medium Intensity Program at Greenough Regional Prison between 07-06-05 and 09-09-05. The program was compiled and developed to meet the specific needs of Indigenous people who participate. Facilitation of the program is group therapy format of 180 hours duration. Both sex offences were examined in the above treatment.
It noted from the outset that GTR has accepted responsibility for his offending behaviour. There is no evidence of cognitive distortion or justification. It noted issues in relation to victim empathy, relationships and substance abuse reporting:
[GTR] has a history of poly substance abuse including petrol sniffing, alcohol and cannabis use. Around the age of 15 or 16 [GTR] ceased sniffing petrol as a result of exposure to anti petrol sniffing campaigns.
Under 'Relapse Prevention':
[GTR's] high risk factors include, limited access to age appropriate sexual partners, feeling lonely and socially isolated, unresolved issues associated with grief over his fathers death, alcohol and cannabis use and loss of cultural direction.
[GTR] is aware of these factors and acknowledges that they could be ongoing risk factors should he neglect to put in place appropriate coping strategies when released to his community. [GTR] states that he has an interest in forming social relationships through music, which is accessible at Wanarn Community. Further, he states that when he has completed his parole period he will reside in Alice Springs where he hopes to establish a relationship with his 'promise'. Longer-term goals include further cultural training/ceremony. He is of the view that his Step Father will be a strong positive influence in his life.
Should [GTR] choose to implement his relapse prevention plan and bring about lifestyle changes, the combination of the two should augur well for reducing his risk for reoffending in a sexual manner.
The risk assessment conducted by the psychologists used Static‑99. In summary the report noted:
Victim selection appears to have been opportunistic, based on vulnerability and isolation suggestive of predatory type behaviour. [GTR] does not appear to have developed a sexual preference for mature aged Caucasian females. Sexual preference is stated to be for age and culturally appropriate females.
…
… On a positive note, [GTR] claims to have some direction and goal whereas at the time of his previous release he claimed not [to] have developed either direction or goals. A further strength for [GTR] is the support and guidance that he claims his Mother and Stepfather have indicated they are willing to provide, whereas his previous release address was to his Grandmother in Warakurna where there was a lack of a strong positive male role model.
Facilitators are of the view that [GTR] would continue to gain self‑confidence should he be released to parole under a mentor ship program if possible.
There are a number of assessments included in the book of evidence including a Parole Assessment Report, a Sex Offender Risk Update Report, and Community Supervision Assessment. These reports are of very limited assistance on the issue of risk, being concerned in the main with the management of future risk and conditions that might be imposed either on parole or as a result of a supervision order, rather than formal assessments.

Conclusion on psychiatric opinions and reports under s 7(3)(a), (b)

In Director of Public Prosecutions (WA) v Mangolamara [2007] WASC 71; 169 A Crim R 379, Hasluck J conducted, with respect, a very comprehensive review of the DSO Act and its requirements. In particular, Hasluck J set out issues that concerned expert evidence and particularly the use of risk assessment tools before concluding:
In the end, bearing in mind that the rules of evidence reflect a form of wisdom based on logic and experience, I am of the view, for the reasons I have referred to, that little weight should be given to those parts of the reports concerning the assessment tools. In my view, the evidence in question does not conform to long-established rules concerning expert evidence. The research data and methods underlying the assessment tools are assumed to be correct but this has not been established by the evidence. It has not been made clear to me whether the context for which the categories of assessment reflected in the relevant texts or manuals were devised is that of treatment and intervention or that of sentencing. Dr Pascu acknowledged under cross-examination that the assessment tools are directed not to the commission of serious sexual offences but to sexual re-offending of any kind (t/s 60). She acknowledged also that the database used for the mathematical model upon which Static-99 was based related to untreated English and Canadian sex offenders released back into the community on an unsupervised basis (t/s 68).
Moreover, having regard to the admissions made under cross-examination that the tools were not devised for and do not necessarily take account of the social circumstances of indigenous Australians in remote communities, I harbour grave reservations as to whether a person of the respondent's background can be easily fitted within the categories of appraisal presently allowed for by the assessment tools. ([165], [166])
Similar issues arise in the present case, although of course the evidence is different. Both psychiatrists use a variety of tools to support their assessment.
The qualifications and limitations on the use of predictive models in the evidence speak for themselves. These limitations are supported by the published literature to which I have referred. For reasons similar to those expressed in Mangolamara, I cannot attribute significant weight to the expert psychiatric opinions as to risk. I accept that the use of one or more predictive models, with or without a clinical interview and appraisal, may be helpful in determining a counselling regime or a management strategy for an offender. In such cases there has already been a determination of guilt and a sentence has been imposed. Little prediction is required by the sentencing judge. Within that context there is usefulness in the models to aid the offender's rehabilitation, to customise a course of treatment or therapy, and to plan for the offender's release to the community.
However, an application under the DSO Act requires more intense scrutiny. The respondent's liberty may be removed or curtailed because of a prediction which a judge is required to make as to future offending. For that reason, the DSO Act requires acceptable and cogent evidence to a high degree of probability. While opinions based on the present predictive models may be suitable for management purposes, they lack cogency for the purposes of the DSO Act that little weight can be attributed to the results of assessments that rely on them. Accepting the view expressed that clinical interview alone is a poor predictor, it remains the case in Western Australia that as yet the tools that are being developed to increase the accuracy of predictive outcome of dangerous sexual offenders have not developed to such a stage that the evidence can be described as 'acceptable and cogent'.
Of course, the role of a judge is different to that of a psychiatrist providing a report. A judge is required to consider all the factors in the DSO Act s 7(3); the psychiatric evidence is one of those factors. It is the combination of all the factors which must be considered in the formation of a judgment whether the DPP has satisfied the court under s 7(2).

Section 7(3)(c)

information indicating whether or not the person has a propensity to commit serious sexual offences in the future;

The evidence is silent as to propensity.

Section 7(3)(d)

whether or not there is any pattern of offending behaviour on the part of the person;

The behaviour would indicate a pattern in the sense that the respondent on each occasion was intoxicated with marijuana and attacked an older woman opportunistically.

Section 7(3)(e) and (f)

(e)any efforts by the person to address the cause or causes of the person's offending behaviour, including whether the person has participated in any rehabilitation program;
(f)whether or not the person's participation in any rehabilitation program has had a positive effect on the person;

The respondent was assessed on 2 November 2004, at Hakea, as suitable for the Indigenous Sex Offender Medium Intensity Program at Greenough. The identified issues included the predatory nature of his offences and apparent lack of victim empathy and an unwillingness to recognise and address the contributing factors to his offending behaviour. The respondent completed the Indigenous Sex Offender Medium Intensity Program at Greenough between 7 June 2005 and 6 September 2005. The program was compiled and developed to meet the specific needs of indigenous people who participated. Facilitation of the program was group therapy format of 180 hours duration. Both sex offences were examined and the treatment was noted:
Responsibility Taking
[GTR] from the outset has accepted responsibility for his offending behaviour. Initially, he claimed that he could not recall the details of the first sex offence, however during the programme he disclosed further details that eventually provided insight into the precipitating events leading to the offence. [GTR] has accepted full responsibility for both offences and has provided details consistent with the victims' depositions and statements of material facts.
Cognitive Distortion, Sexual Objectification and Justification
There is no evidence of cognitive distortion or justification. [GTR] has never suggested that the victims were in any was responsible for the sex offences. He acknowledges that the victims were selected based on opportunity and easy access and that he did not necessarily find them sexually exciting. Rather, he engaged in sexual fantasies about the act of intercourse and became aroused.
Victim Empathy
It is difficult to discern if [GTR] has established victim empathy for his first victim at any level other than cognitive. This might be due to his negative affective state at the time of the offence and his state of intoxication. Nonetheless, he appears to understand that the victim has suffered and for this he states he is deeply ashamed. [GTR] appears to have established victim empathy for his second victim at an affective level. It is noted that during the sex offence the victim attempted to reason with him and drew his attention to her age and 'grandmother' status. [GTR] claims this shocked him and he found that he could not continue with the assault. It appears that [GTR] does have the capacity to establish victim empathy and that he does have respect for the status attributed to mature females in his culture when he is not under the disinhibiting effects of cannabis. It should be noted that [GTR] had difficulty in articulating feelings and thoughts in English language (his third language) and that his expression of victim empathy is expressed in terms of shame, as an individual and culturally. (249)

Section 7(3)(g)

the person's antecedents and criminal record;

The proceedings in 2002

If evidence of that material is admissible the following is relevant:
[I]n May [2002], [the complainant], was 56 years of age. At the time … she lived at the Warburton Ranges remote community. She was … employed … as a pre‑primary and primary school teacher. At about 4.30 pm on the 12th of May, she left her home for a 1 hour walk. She decided to walk out of town for about half an hour, and then back again along the Blackstone Road, which is to all intents and purposes, a gravel track.
On the way out, she noticed an Aboriginal male sitting on the ground drawing in the dust, and she thought nothing of it. On her way back, the same Aboriginal male walked out of the bushes from her left‑hand side. She noticed that he was naked, and she also noticed that his penis was erect.
As they got closer, she thought that maybe she'd walked into men's business, and she politely said good evening to the male. And it's the Crown case of course that the male is [GTR]. [GTR] said something to her in response to her saying good evening, but she couldn't hear what he said. He then moved towards her, and pressed himself up against her.
[The complainant] … grabbed [GTR's] testicles. She then tried to hit him with a water bottle that she was carrying. He then grabbed her by the wrists and dragged her across the road some considerable distance into the bushes. Once off the road, he threw [the complainant] onto her back and he laid on top of her. He began grabbing at her pants and trying to undo the buttons. [The complainant] asked what he wanted, and [GTR] said sex.
… But he continued to try and unto her pants. Not knowing what to do, and not wanting the situation to become more physical, she said to him 'Stop. I'll give you a hand job', to which [GTR] responded 'I want you to suck me off'. Upon saying that, he put both his legs on either side of [the complainant], …
Not wanting his penis in her mouth, [the complainant] said 'No, I can't like this, I'll need to sit up'. As she tried to sit up, she grabbed hold of his penis and was twisting and turning so it couldn't enter her mouth, and almost as soon as she grabbed it, his penis that is, [GTR] ejaculated on her T-shirt and her shorts. At this point, he got up and ran away, in the opposite direction to the community.
The judge was not satisfied there was any lengthy premeditation. The judge could not draw an inference that the respondent took his clothes off for the purpose of preparing himself for an attack on the woman. The judge was of opinion that the attack fell into the category of a spontaneous opportunistic attack at a time when the respondent was intoxicated by cannabis. The attack was violent though not gratuitously violent.

The proceedings in 2004

On 20 September 2004, in the District Court the respondent pleaded guilty that:
On 29 March 2004, at Warracoona Aboriginal Community, you unlawfully detained [the complainant].
And further, on the same date and at the same place, you attempted to sexually penetrate [the complainant] without her consent, you were armed with an offensive weapon, namely a knife, and that [the complainant] was over the age of 60 years.
And further, that on the same date and at the same place, you unlawfully wounded [the complainant], and [the complainant] was over the age of 60 years.
The facts as accepted were as follows:
The complainant is a 65-year-old female Caucasian nurse who works and resides in the Warracoona Aboriginal Community located approximately 830 kilometres east of Alice Springs. The complainant is single, about 158 centimetres tall and slight of build and resides by herself at the community. The complainant's front door is secured by a mesh cage to allow her to accurately assess individuals who come to visit her without endangering herself.
The offender is an 18-year-old male of similar height and build. At about 9.30 pm on Sunday, 28 March 2004 the offender went to the home residence of the complainant at house 49 Warracoona Aboriginal Community on the Great Central Road. At the time the offender had in his possession a claw hammer approximately 35 centimetres in length. The offender presented saying he was suffering from a headache and needed medical attention. The complainant assessed the offender without opening the mesh cage and gave the offender Panadol and requested that he go to the community health clinic in the morning.
At about 12.30 am on Monday, 29 March 2004 the offender again went to the complainant's home residence. The offender said that his mother was sick and she required medical assistance. He further said that his mother was waiting at the nursing clinic. The offender accompanied the complainant to the nursing clinic which was about 200 metres away from the complainant's house.
When they got there, the complainant and the offender went in via the back door and once inside the offender said that his sick mother was at the front of the clinic. The complainant opened the front door of the clinic and turned the remaining lights on. The complainant asked the offender what his mother's name was and the offender replied 'Denise Williams'. The complainant then began searching for his mother's medical file.
The offender left the room and the complainant assumed that the offender was going to use the toilet. The offender went back to the office where the complainant was and sat directly behind her approximately two metres away from her. At the time the complainant had her back to the offender. The complainant was not able to locate any details of the name given by the offender and when she turned around to question him, she saw that the offender was exposing his erect penis.
The complainant yelled at the offender to leave the premises and the complainant tried to leave the clinic but discovered that the offender had locked both the front and the back doors, making it impossible for her to escape. Near the back door the offender grabbed the complainant from behind, wrapped his arms around her chest and pinned her arms to the side of her body. The offender produced a knife which he held in his right hand against the complainant's stomach. The knife was approximately 30 centimetres in length with a broken handle and a 20-centimetre long blade. The blade tapered to a sharp point.
The complainant and the offender continued to struggle in various places in the nursing clinic. During the struggle the offender said on a number of occasions, 'Just a quick one,' requesting the complainant for sex. The struggle continued as the complainant tried to set off a duress alarm which was in the office area. Once in the office area the offender pushed the complainant against a fax machine where he pushed the knife against her stomach. In an attempt to try and get the knife away from her the complainant received a number of cuts to her hands and arms. Each cut was about two centimetres in length and went through all layers of skin, causing bleeding.
Fearing for her safety and in an attempt to settle the offender down the complainant said that she was a grandmother. This caused a break in concentration of the offender which allowed the complainant to knock the knife from his hand onto the bench between the fax machine and the wall. The complainant then took the knife and had an argument with the offender. The argument resulted in the complainant getting to the back door, unlocking it and throwing the knife onto the ground outside. The offender then left the clinic via the back door and the complainant locked the door behind him. She contacted staff from the community for help.

The judge's sentencing remarks included the following:
The aggravating factors here are significant. There was an element of premeditation or planning in the carrying out of the offending. There were the calculated acts of working out how to get her out of her house which he would have realised was secure because of the mesh door into the clinic by lying about the presence of his mother. I understand not only wasn't his mother ill, of course - it was simply a trick - she wasn't even in the community apparently - then locking the doors to prevent the complainant leaving and, of course, being armed with the knife, using the knife to attempt to obtain her compliance and, of course, maintaining the knife pressing against her stomach and, of course, the complainant was wounded in terms of the defensive cuts when she tried to struggle and break fee.
In addition, the complainant was aged 65 so that is also a further aggravating factor. However, I do accept on the information put before me that that is not necessarily something that would have been apparent to the offender. …
Certainly it appears that when she told him of her age or the fact that she was a grandmother, that did apparently give him sufficient pause to allow her to be able to struggle free. …
…
He is aged 19. He would have been aged 18. It was just before his 19th birthday that this offence was committed. He is a young Aboriginal man who has been raised entirely in an Aboriginal community and, of course, it was in that community that this offence occurred. …
…
Having said that, there is, …, great concern because of the offences that were dealt with by the Children's Court in 2002. …
However, it means that he comes before the court as a 19-year-old young Aboriginal man with the problems of trying to cope in the community with the social problems, the dysfunction, the violence, the petrol-sniffing, effects of long‑term petrol-sniffing and long-term cannabis abuse, but at the same time - and also a young man, it would seem, with real communication difficulties and real difficulties understanding his behaviour and what led to his behaviour, but at the same time a young man which comes before the court not - so this incident is not an isolated occasion. He comes before the court having been dealt with for a remarkably similar offence.
There was also the issue which I regard as a significant aggravating factor that the circumstances of this offence indicate a degree of a planning. …
…
Clearly the only appropriate disposition in this case is a significant term of immediate imprisonment. The principles of personal and general deterrence are of significant importance when the court is dealing with such a serious offence. It is, of course, fortunate for the complainant and in terms of the disposition and length of term for the offender that the complainant was able to break free and the offence in fact - the offence, of course, was one of attempted aggravated sexual penetration.
That means, of course, that the maximum penalty open to the court is a term of 10 years, being one half of the maximum term that would have been available. The sentence will be reduced to take into account the offender's personal circumstances and his early plea of guilty. Although there is no indication of remorse or understanding of the nature of the offence in the pre-sentence report, against that has to be factored the fact that he obviously has real difficulty in communicating.
I'm advised … and accept that he has at least some commitment to undergoing appropriate sex offender treatment programs while he is in custody and he appreciates the reasons why he is in custody for a lengthy period of time and has some commitment at least on a rudimentary basis to try to avoid being put in the same situation again.
So the sentence will be reduced to take into account his plea of guilty and his personal circumstances. He is still a very young man. …
…
He is also, of course, a young man of Aboriginal descent who has grown up in an Aboriginal community. A term of imprisonment will be a very difficult one for him even if he is placed in Greenough which seems to be some suggestion because of the availability of the program. He will be out of the community for a long period of time and he will have difficulties adjusting. All of those factors need to be taken into account in assessing the appropriate length of sentence.
…
So taking all those matters into account, I consider the sentence will be reduced from what I would otherwise consider appropriate to a term of five years in relation to the offence of attempted sexual penetration. That will have to be reduced because of the amendments to the Sentencing Act from 60 months to 40 months, being a term of three years and four months.
In relation to the other two charges - the charge of deprivation of liberty I would have sentenced him to a term of two years' imprisonment and that will have to be reduced to 16 months because of the amendments to the Sentencing Act and the offence of unlawful wounding I would have sentenced him to a term of two years' imprisonment and that will have to be reduced to a term of 16 months because of the amendments to the Sentencing Act. Those two terms will be concurrent with each other and concurrent with the term imposed for the offence of attempted sexual penetration.
The sentencing judge was concerned that:
If the sentence is too long, then I consider that it would be virtually impossible for [GTR] to be able to re-engage in the community in any kind of meaningful way and that will simply mean that any attempts at rehabilitation that are made while he is in custody may well not be successful in the sense of it may not result in him refraining from similar offending in the future because if he is unable to engage back into the community, he is likely to remain at a greater risk to the community generally.
In the event parole was refused. The term of 3 years 4 months was backdated to 20 March 2004. The sentence expired at the latest 30 July 2007. However, Blaxell J made a preliminary order on 19 June 2007. As a result the respondent has remained in custody. The integration back into the community referred to by the sentencing judge is now much harder.

Section 7(3)(h)

the risk that, if the person were not subject to continuing detention order or a supervision order, the person would commit a serious sexual offence;

There is a degree of circumlocution in s 7 because the terms of s 7(3)(h) appear within s 7(1):
Before the court dealing with an application under this Act may find that a person is a serious danger to the community, the court has to be satisfied that there is an unacceptable risk that, if the person were not subject to a continuing detention order or a supervision order, the person would commit a serious sexual offence.
The DPP seeks only a supervision order. It therefore acknowledges that the risk to the community is not so great that nothing less than removal from the community by way of imprisonment is appropriate. Rather, it asserts that with close supervision of the type outlined in the minute of proposed orders, the risk can be made acceptable.

Section 7(3)(i)

the need to protect members of the community from that risk; and

Unlike s 7(1) neither s 7(3)(h) and s 7(3)(i) refer to an 'unacceptable' risk. The normal processes for the protection of the community known to law are those provided under the Sentencing Act whereby Parliament has detailed a regime of disposition designed among other reasons to provide protection for the community against the risk of offending. The DSO Act regime commences at the conclusion of a sentence. The risk must therefore be assessed on the basis that a purpose of sentencing, namely personal deterrence, has not achieved its object. This is difficult to evaluate in the present case.

Section 7(3)(j)

any other relevant matter.

The DPP did not suggest there was any other relevant matter.

Conclusion

The DSO Act s 7 is predicated on the basis that there will often be a risk to the community that a person will commit a serious sexual offence. The court must be satisfied to a high degree of probability make a finding that the risk is 'unacceptable'. In a sense any risk is unacceptable, especially from the point of view of a victim of a subsequent offence. However, the acceptability of the risk must be weighed against the extraordinary nature of the consequences of an order for detention or supervision on a citizen who has otherwise completed all terms of imprisonment as punishment for past behaviour. There is a balance between the acceptance of some risk to the community against the right of a citizen who has discharged a debt to society to be free.
The respondent is still a young person who might reasonably be expected to mature and gain greater awareness of the consequences of his conduct. Against that, each of the offences was committed at times when he was intoxicated so that rational thought was substantially inhibited.
The court has to be satisfied to a high degree of probability that there is an unacceptable risk that the person would commit a serious offence if not subject to an order.
The two concepts do not sit well together. The concept of 'unacceptable risk' is a conclusion. The concept of 'probability' is used in law to determine the likelihood of certain facts being correct. Probability is easily applied to the evaluation of evidence. It does not sit so easily with the formation of a conclusion. Nevertheless, despite the difficulties, that is what Parliament requires judges to do.
Having taken into account all the matters set out in s 7(3), I am not satisfied by acceptable and cogent evidence to a high degree of probability that there is an unacceptable risk that GTR will commit a serious sexual offence if not subject to a supervision order. I acknowledge immediately that there is a risk. However, the evidence put forward by the DPP does not persuade me, to a high degree of probability that the risk is unacceptable in the sense that his liberty should be constrained in some way. Little weight can be placed on the psychiatric reports for the reasons I have outlined. There is no suggestion of a psychiatric diagnosis of a mental disorder. Nor has there been a comprehensive psychological assessment. There is the troubling evidence of the past history to which I give significant weight. Against that, there is also positive participation in sex offender courses. Obviously, participation in the early sex offender course did not prevent the later crime. However, the respondent is still young and it must be assumed that the later period of incarceration has had some effect on his developing maturity in teaching him consequences of unlawful behaviour. There is a pattern to his past offending but no evidence of a propensity to so offend.
The application should be refused because without regard to the Children's Court conviction there is insufficient evidence of serious risk. Nevertheless, I have conducted a review of all the evidence, taking into account the Children's Court conviction, or more precisely the behaviour disclosed by it. Following that review the result remains the same.

[2007] WASC 318

20 December 2007