Date: 20150918

Docket: T-1350-05

Citation: 2015 FC 1093

BETWEEN:

JEFFREY G. EWERT

Plaintiff

and

HER MAJESTY THE QUEEN IN RIGHT OF CANADA (THE COMMISSIONER OF THE CORRECTIONAL SERVICE OF CANADA, THE WARDEN OF KENT INSTITUTION AND THE WARDEN OF MISSION INSTITUTION)

Defendant

REASONS FOR JUDGMENT

PHELAN J.

I.                   Nature of Claim

[1]               This is a claim, originally in damages, for the actions of the Commissioner of Correctional Services Canada [CSC] in using various psychological risk assessment tools [actuarial tests] during the Plaintiff’s incarceration. These actuarial tests were alleged to be unreliable in regard to Aboriginal prisoners. The use of these tests results in a significant adverse impact on the Plaintiff.

The claim for damages and for relief in regard to s 12 of the Canadian Charter of Rights and Freedoms [Charter] were abandoned in argument. The current claim is for relief in respect of s 7 and s 15 Charter violations.

II.                Factual Background

A.                Ewert - General

[2]               Jeffrey Ewert is a 53 year old Aboriginal offender who identifies himself as Métis. His mother was of Métis descent, and his father was a British serviceman. At approximately 6 months old, he was adopted by a Caucasian family, the Ewerts, in Surrey, British Columbia.

[3]               The evidence in this case consisted of two fact witnesses, Ewert for himself and Dr. Motiuk (Head of Research for CSC) for the Defendant, and two expert witnesses, psychologists Dr. Hart for the Plaintiff and Dr. Rice for the Defendant. It is a thin record on which to decide whether certain psychological tests administered by CSC infringe prisoner rights. The issue of the validity of these tests impacts not only Ewert but other Aboriginal prisoners and potentially all prisoners who are similarly assessed.

[4]               Ewert stands in place for himself and all Aboriginal prisoners, but he is no poster person for Aboriginal people. Ewert is serving two life sentences for second degree murder and attempted murder, and was sentenced to 15 months’ imprisonment to be served concurrently for an escape from lawful custody conviction. He has spent 30 years in various federal correctional facilities – over half of that time has been spent in maximum security institutions; the remainder of time has been spent in medium security institutions.

[5]               Ewert’s life sentence offences were brutal crimes. In the first offence, he strangled and sexually assaulted the victim, leaving her dead in the river. The second offence similarly involved strangulation and sexual assault, and the victim was left brain damaged and crippled.

[6]               Ewert’s early life was itself tragic, as it involved an alcoholic adopted father, a psychologically ill adopted mother and racism and discrimination from his adopted Caucasian siblings.

[7]               The Plaintiff has been eligible for day parole since August 1996 and for full parole since August 1999. He has never had a parole hearing, and has waived his right to each parole hearing alleging a) that he was unlikely to be granted parole because he was assessed as too great a risk of reoffending, in part due to the results of the actuarial tests; and b) that he wished to spare the families of his victims the anguish of a parole hearing.

[8]               On this latter reason for not seeking parole, there is no evidence that he waived parole for this reason. I do not accept this self-serving statement, nor many of Ewert’s other self-serving statements or attempts to blame others for his predicaments. I treat his evidence with a considerable degree of scepticism; however, there is evidence supporting his contention that the actuarial tests’ results did have an adverse effect on his incarceration conditions.

[9]               It is necessary to point out that the Defendant put in no substantive witness evidence to rebut Ewert’s narrative. Dr. Motiuk had no involvement in prison operations and no one, such as a warden, deputy warden, prison psychologist or senior “line” prison official, was called to give evidence to counter Ewert’s perspective on the factual use and impact of these tests.

[10]           The history of Ewert’s prison term demonstrates that there were times when he appears to have been productive, made progress towards his correctional plan and had a positive relationship with staff and his case management team. Over the course of 30 years, there were times when the opposite was true.

[11]           While he may not have been a model prisoner, was a bit of a jail-house lawyer and had some incidents that counted against his security classification, he had few, if any, serious valid institutional charges. His security classification since 2003 varied from high to low with “moderate” being the predominant rating.

B.                 Assessment Tools (Psychological Tests)

[12]           The assessment tools are the psychological tests (sometimes referred to in the evidence as actuarial tests). At issue are the Hare Psychopathy Checklist Revised [PCL-R] - a most important test; the Violence Risk Appraisal Guide [VRAG]; the Sex Offender Risk Appraisal Guide [SORAG]; the Static 99; and the Violence Risk Scale - Sex Offender [VRS-SO].

It was common practice at CSC to use the VRAG, SORAG, Static 99 and VRS-SO to assess risk of violence in Aboriginal inmates and other prisoners, and to use the PCL-R to assess psychopathic personality disorder.

(1)               PCL-R

[13]           Psychopathy is a clinical construct with particularly important criminal justice implications, since a psychopath or psychopathic disorder poses considerable problems relating to risk and rehabilitation.

[14]           The PCL-R is a commonly used measure of this clinical construct, developed by Dr. Robert Hare of UBC in the 1970s. Dr. Hare was not called by either party; however, the expert evidence is that the test was developed with the intention of assessing the presence of psychopathy, but it has been used to predict recidivism risk. The Plaintiff does not challenge this secondary use per se, nor does he assert that it is not reliable for non-Aboriginal prisoners. He does challenge the validity and accuracy of the use of these tools in regard to adult Aboriginal inmates.

[15]           The PCL-R is a 20-item inventory of personality traits or behaviour. The items are divided into two groups: Factor 1 and Factor 2.

[16]           Factor 1 contains eight factors, which are further sub-divided into two “facets” described as “interpersonal” and “affective” facets (e.g. “glibness and superficial charm” or “shallow effect”). Individuals are scored 0, 1 or 2 on the absence, potential presence or actual presence of each of the listed personality features.

[17]           Factor 2 contains nine personality features and scores the individual on a value of 0-2 based on the presence or absence of various past types of behaviours. Factor 2 also has two facets described as “Lifestyle” and “Antisocial” facets (e.g. “early behavioural problems” or “juvenile delinquency”).

[18]           Factors 1 and 2 variables are weighed equally. PCL-R generates a total score by separately summing the items under each of Factor 1 and Factor 2. Percentages are assigned to each Factor and then an average of the percentage is taken. In addition, a total numeric score is generated by summing Factor 1 and Factor 2 item scores and those items that are not included in either Factor 1 or Factor 2.

[19]           The evidence is that in using the PCL-R, one must use both Factor 1 and Factor 2. Over time, it has become clearer that Factor 1 has substantial reliability issues and has been called “junk” in respect of its use for Aboriginal prisoners. Factor 2 does not suffer from that criticism.

(2)               VRAG/SORAG

[20]           VRAG and SORAG are actuarial tools designed for the prediction of violent recidivism. VRAG and SORAG assign a probability from 0% to 100% that an offender will commit a new violent offence or sex offence within a specific period of community access, comparing the risk posed by the offender to the risk posed by other offenders.

[21]           VRAG and SORAG both incorporate PCL-R scores as the most heavily weighted factors. I accept Dr. Hart’s evidence (for reasons to follow later) that the PCL-R score was responsible for most of the predictive power of the VRAG and SORAG.

(3)               Static 99

[22]           The Static 99 is an actuarial tool designed to estimate the probability of sexual and violent recidivism among adult males who have been convicted of at least one sexual offence. The Static 99 is intended to measure long-term risk potential by assessing a number of unchanging factors. The Static 99 test is intended to be administered in an interview setting by probation and parole officers, case managers or mental health professionals.

(4)               VRS-SO

[23]           This test is a rating scale designed to assess risk and predict sexual recidivism, to measure and link treatment changes to sexual recidivism and to inform the delivery of sexual offender treatment. The VRS-SO comprises static and dynamic factors and generates both qualitative and quantitative assessments of inmates. The VRS-SO is used following sex offender treatment to assess the success of that treatment.

III.             Expert Evidence

[24]           The parties submitted opposing experts’ reports. For reasons which will become obvious, the Court generally accepted the evidence of Dr. Hart on behalf of the Plaintiff and put little weight on the evidence of Dr. Rice called for the Defendant. Dr. Rice’s evidence was so infirmed, so inconsistent with the role of CSC and so infused with a singular narrow view that it was not helpful to the Court or even to the Defendant.

A.                Dr. Stephen Hart

[25]           Hart is a professor of psychology at Simon Fraser University in Burnaby, British Columbia; a visiting professor of psychology at the University of Bergen, Norway; and a threat assessment specialist at ProActive Re Solutions Inc in Vancouver. He was qualified to give opinion evidence in the area of the development, design, application, evaluation, validity and reliability of actuarial and psychological instruments used by CSC, including the PCL-R, VRAG, SORAG, VRS-SO, Static 99 and their variants.

[26]           His evidence was balanced, objective, and despite the Defendant’s criticism that he cited no studies, credible. A summary of his evidence follows.

[27]           Actuarial tests, like the ones in issue, are susceptible to four types of cross-cultural bias – conceptual variance, structural variance, metric variance and predictive variance.

[28]           Because of the significant cultural differences between Aboriginal and non-Aboriginal Canadians, the actuarial tests at issue in this action are more likely than not to be “cross culturally variant”.

[29]           Dr. Hart was not speaking specifically of Ewert. The Court is cognizant that Ewert’s cultural connection with his Aboriginal roots was also influenced by his largely Caucasian suburban Surrey upbringing – significantly different from an Aboriginal person brought up largely in an Aboriginal milieu.

[30]           Dr. Hart recognized this subjective element in his discussion of culture as an inchoate concept consisting of the sum of shared experiences, beliefs, norms, family and relationships, structures, values and so forth. While culture cannot be vigorously defined, it is an indispensable concept.

Membership in a culture or cultural subgroup is defined and assessed by the subjective participation or adherence of an individual to that culture or cultural subgroup.

[31]           Particularly important is Hart’s opinion that given the pronounced differences between Aboriginal and non-Aboriginal groups, he would not apply the scores derived from the PCL-R, VRAG, SORAG, Static 99 or VRS-SO to Aboriginal persons which includes the Plaintiff.

[32]           Hart’s evidence exhibited the type of balance that the Court found helpful. Rather than reject these actuarial tests out of hand, Hart suggested that the better approach was to have a structured clinical assessment of an Aboriginal offender which would include some consideration of the information derived from the actuarial tests in the totality of the circumstances of what is known about the offender.

[33]           A key issue in this litigation is the importance of these actuarial test scores. While Dr. Hart would contextualize and approach them with caution, Dr. Rice, on behalf of the Defendant, saw the scores as determinative. Dr. Motiuk tried to de-emphasize their importance, but Ewert’s experience shows that the scores are important and contribute to the decisions made affecting key aspects of his incarceration.

[34]           Dr. Hart testified that there were three ways to establish that an actuarial test was free from cross-cultural bias, which is important in addressing the reliability concerns of the Plaintiff and the remedy that the Court could order:

1)                  confirmatory factor analysis (which confirms whether the factors look the same in different groups);

2)                  metric analysis or Item Response Theory (which assesses whether the same things are measured in the same way); and

3)                  predictive analysis (statistical regression analysis to examine the slopes and intercepts of statistical relationships).

[35]           It is hardly practical for an individual litigant to engage in this type of analysis. Given the CSC’s legislated mandate, discussed later, it is an activity more appropriately commissioned by CSC.

[36]           None of these types of analyses have been completed on the subject of actuarial tests.

[37]           The only study of the PCL-R on Aboriginal inmates is the Olver Study, published in 2013. The Olver Study was produced by Dr. Rice, although neither she nor any other witness was involved in the study or could speak to it on a first-hand basis. While the Olver Study may fall within the hearsay exception to expert evidence, its weight in this case is questionable.

[38]           The Olver Study suggested that the PCL-R did validly predict for Aboriginal people. Dr. Hart discounted the Olver Study because of its small sample size and failure to examine predictive variance of Factor 1 which calls into question the validity of the total scores as a valid indicator of future behaviour.

[39]           The Olver Report concluded based on the statistics in the study that the predictive power of the PCL-R with respect to recidivism was primarily attributed to what is called Factor 2. Factor 1 was found to have no predictive value.

[40]           The issue raised by this conclusion is that both Factor 1 and Factor 2 are combined and used in the PCL-R scores relied upon by CSC.

[41]           The conclusion from Dr. Hart’s expert report is that the actuarial tests are not sufficiently predictably reliable for Aboriginals because of the cultural variance or bias of the tests.

Dr. Rice, the Defendant’s expert, did not overcome the essential features of Dr. Hart’s Report.

B.                 Dr. Marnie Rice

[42]           Dr. Rice is a clinical psychologist, researcher and professor of psychology and psychiatry. She has worked at the Waypoint Centre for Mental Health Care (formerly the Mental Health Centre Penetanguishene) for over 39 years as a clinical psychologist and researcher.

[43]           Dr. Rice’s opinion was that the subject actuarial tests are a reliable and valid predictor of the future risk of recidivism. She was of the view that clinical judgment was demonstrably inferior to the impugned assessment tools. She saw no basis for cultural bias in these assessment tools.

[44]           Dr. Rice’s evidence was problematic on several levels. It may have been her first experience as an expert witness; however, it was the context of her evidence that was troublesome.

[45]           The Plaintiff objected to Dr. Rice’s evidence because she failed to disclose in her Report as required by s 3(k) of the Code of Conduct for Expert Witnesses that she was one of the authors of the VRAG and SORAG manuals. She was aware of this disclosure requirement. Her only explanation for failing to disclose is that she thought her involvement was obvious because her last name is associated with those tests.

With respect, having reviewed all the documents in this case, this Court was amazed by that suggestion.

[46]           Despite the Plaintiff’s objection and its reliance on the Supreme Court’s decision in White Burgess Langille Inman v Abbott and Haliburton Co., 2015 SCC 23, the Court admitted the evidence with a caution that her interest and belief in these tests would go to the weight to be given her testimony. Had her evidence been struck, the Defendant would have had no expert evidence before this Court.

[47]           In the end, Dr. Rice’s evidence was of little assistance, particularly to the Defendant. It is unnecessary to quote some of Dr. Rice’s more controversial statements about the political reasons behind the use of scientific tests. However, it was her view that the test scores are reliable and immutable. She eschewed the various rehabilitation programs run by CSC as distractions or something akin to giving prisoners something to do while in prison. In that regard, her evidence and central thesis runs contrary to the statutory purpose and the operational goals of CSC.

[48]           Dr. Rice’s absolutist view of the merits of the test scores is difficult to accept. The Court prefers the more balanced opinion of Dr. Hart.

[49]           Dr. Rice relied heavily on the 2002 Dempsey Report to support her view that the VRAG and SORAG were reliable in respect to Aboriginal offenders. That Report was an unpublished master’s thesis from the University of Leicester, which had a very small number of Aboriginal data points and very low predictive power. In fact, Dr. Rice disavowed the narrative and analysis in the paper, but accepted the numerical scores. Such selective reliance undercuts the force of her opinion.

[50]           Dr. Rice also relied on the Olver Study, discussed earlier, for the proposition that the PCL-R was equally reliable for Aboriginal offenders. However, the Olver Study shows that the PCL-R Factor 1 scores are not a reliable predictor of recidivism in Aboriginal offenders – the notion of PCL-R Factor 1 as “junk”. In fact, Factor 2 scored as less predictive for Aboriginal inmates than for others.

[51]           In this regard, Dr. Rice’s evidence more closely aligned with that of Dr. Hart with respect to Factor 1. Importantly, it was her view that to use the PCL-R, one had to use Factor 1 and Factor 2 together. Given the weaknesses of Factor 1, it is difficult to see how combining one unreliable element with one lesser reliable element is cause to have faith in the predictive value. Dr. Rice did not explain this inconsistency.

[52]           It is the Court’s conclusion that Dr. Rice’s evidence, on this case, cannot be accepted except where it is consistent with Dr. Hart’s. Little weight can be given to this testimony.

[53]           In summary, Dr. Hart’s evidence that the actuarial tests are not good predictors of recidivism in Aboriginals- that they suffer from cultural bias- is preferred and accepted.

[54]           The importance of those tests to Ewert’s situation is evident. As Factor 1 of the PCL-R is not reliable, and Ewert tested 100% and 98% on that Factor, his Factor 1 PCL-R scores are unreliable.

Significantly, given that Ewert’s Factor 2 scores were 51% and 61%, on that Factor, Ewert is a “normal or average” inmate.

[55]           The result is that if Factor 1 alone is considered, Ewert scores at the highest psychopathic level. If Factors 1 and 2 are averaged, Ewert scores at moderate to high risk and is above the cut-off grade for psychopathic designation.

[56]           The expert evidence establishes that the test scores, in and of themselves, ought not to be relied upon. The next question is what reliance and impact those actuarial tests had on Ewert’s situation.

C.                 Reliance

[57]           As outlined earlier, Ewert had high PCL-R Factor 1 scores and average Factor 2 scores. CSC relied on these actuarial scores for over 20 years; these scores were intended for and were used by CSC decision makers. The salient psychologist reports and impacts are summarized below by name of psychologist:

a)                  O’Mahoney – August 1995: This is the initial report which scored Ewert in the 100^th percentile on PCL-R Factor 1 and the 51^st percentile on Factor 2. The Factor 1 score was at the highest level of psychopathy and showed major therapy needs.

b)                  Kim/Boer – August 1999: This testing had similar scores to O’Mahoney with the 98^th and 61^st percentiles, respectively, for PCL-R Factors 1 and 2. The explanation offered to Ewert was that these scores were not a diagnosis of psychopathy per se but a quantitative tool to gauge the risk of recidivism.

c)                  Brink – September 1999: Brink concluded from the Kim/Boer report and his own PCL-R test (the evidence of which could not be located) that based on these scores, Ewert had “significant psychopathic traits”. Contrary to Kim’s view that the results are quantitative only, Brink described the results in qualitative terms that required Ewert to accept the scores as saying something about him.

d)                 Alexander – December 2001: This report from the Clearwater sex offender program is a bit of an outlier, but shows that the scores have an impact on an offender’s life. Alexander did not perform a PCL-R test but did a VRS-SO. She concluded that Ewert posed a low to moderate risk of re-offending.

At the Clearwater program, Ewert was accused of inappropriate behaviour by a co‑patient. It later turned out that the accusations were false. Ewert was not believed by CSC staff because of his PCL-R score. This had an adverse impact on his success in the program. The high PCL-R score coloured CSC staff’s view of Ewert, leading them to conclude that he was manipulative, controlling, callous and deceitful – hallmarks of a psychopath.

e)                  Nadeau – 2003: This is the most recent psychological assessment. Relying on O’Mahoney scores and Kim/Boer scores, Nadeau concluded that the PCL-R scores were a good predictor of recidivism. She concluded, based on Ewert’s high scores, that Ewert was a moderate to high risk of re-offending. This is in marked contrast to Alexander’s report.

[58]           The evidence establishes that psychologists and CSC rely on the actuarial test scores. The scores matter. The initial scores, as found by O’Mahoney, carried forward through Ewert’s term in prison. The score is like a branding – hard to overcome. This is unsurprising, since all types of institutions in society use testing scores that have the tendency to follow the test subject throughout their life in the relevant institution. An apt parallel can be found in the example of early school IQ testing in which a child is identified as “special needs” or “gifted”, and these results or classifications follow the child throughout their institutional educational experiences. In this case, marks matter.

D.                Impacts

[59]           While marks matter, they are not always determinative of a decision. However, they are an important contributing factor - though the weight varies from situation to situation - to the institutional decisions.

[60]           With respect to parole, the Parole Board will release an offender if it is satisfied that the offender will not, by re-offending, present an “undue risk to society”. In the only CSC report prepared for Ewert’s parole, the author relied on the PCL-R scores to conclude that Ewert was an undue risk to society.

Ewert explained that he waived parole hearings because his high score meant likely rejection and once rejected, subsequent parole applications would be difficult, if not impossible. That explanation makes sense in the circumstances.

[61]           An inmate’s security level is an important aspect of prison life. The security level is based on an assessment of 1) Institutional Adjustment; 2) Escape Risk; and 3) Public Safety Risk. To get to minimum security, a prisoner must have a low rating in all three areas. The potential for violent behaviour is a critical factor in this assessment.

[62]           Ewert has been rated as either a medium or maximum public safety risk. The constant in all of these ratings has been the psychological assessment which relies in part on the actuarial test scores. This is true, for example, of the 2003, 2005, 2010(2), 2012 and 2014 security classifications.

[63]           The Defendant asserts that Ewert’s Custody Rating Scale [CRS] is not reliant on these test scores and that his problems stem from the CRS, not the security classification. However, given the repeated reliance on these scores in the context of assessing public safety risk, it is hard to accept that they play no role or only a minor role in the public safety risk assessment. If the scores were immaterial, one wonders why they appear again and again in CSC decision making.

[64]           The last example of the impact of the psychological risk assessment in Ewert’s incarceration is the denial of Escorted Temporary Absences [ETAs]. Like the parole decisions, the ETAs are based on the institutional head’s consideration of the undue risk to society while the prisoner is away from the prison.

[65]           A review of Ewert’s documentary evidence in the denials of his ETA requests discloses that on at least three occasions, the actuarial scores, as well as their detailed breakdown, formed the basis of the refusal to grant the ETA.

[66]           It is the Plaintiff’s position that the actuarial scores generated by the PCL-R, VRAG, SORAG and Static 99 were contributing factors in the denial of his ETA requests.

E.                 Dr. Larry Motiuk

[67]           The Defendant’s other witness, Dr. Larry Motiuk, was the former Director General of the Research Branch of CSC and is currently the Assistant Commissioner of Policy. He had particular knowledge of CSC policy including as it relates to security classifications.

[68]           His evidence covered the general nature of CSC decision making and the effects of the assessment tools, including the PCL-R. He explained in some detail how the security classification system worked, including the CRS and Security Reclassification Scale [SRS], and testified that neither scale took into account or was influenced by an offender’s score on any particular assessment tool.

His evidence was helpful in this regard; however, the issue of the use of these assessment tools covers more than the CRS and SRS scores.

[69]           Despite the overall thrust of his evidence, Dr. Motiuk admitted that a prisoner’s actuarial test scores were a factor that was required to be taken into account by the relevant decision maker in determining the offender’s overall security rating.

[70]           While Dr. Motiuk attempted to de-emphasize the role of the actuarial tests, his viewpoint was at odds with Dr. Hart’s evidence that it was common practice within CSC to use actuarial results on Aboriginal inmates to assess risk of violence and psychopathy. I prefer Dr. Hart’s evidence because Dr. Hart is closer to the actual experience at institutions and is more objective. I say this without suggesting any criticism of Dr. Motiuk.

[71]           Dr. Motiuk also addressed the research function at CSC and the absence of any research into the specific assessment tools at issue. He shifted responsibility in regard to these tools to the psychologists retained by CSC – a shifting of responsibility not countenanced by the legislation.

He acknowledged that CSC had conducted research on CSC’s tools, including the SIR (Statistical Information for Recidivism Scale) and the CRS and SRS (referred to earlier). He indicated that research on the assessment tools was feasible but posed some difficulties. Dr. Motiuk seemed to suggest that the loss of use of these actuarial tests would not be a great problem for prison administrators.

[72]           In regards to research, Dr. Motiuk spoke of unfulfilled research plans impacted by budgetary concerns. There was no evidence that CSC had completed the research referred to by Justice Beaudry in Ewert v Canada (Attorney General), 2007 FC 13, 306 FTR 234, and anticipated by the Federal Court of Appeal in the appeal of Justice Beaudry’s decision (Ewert v Canada (Attorney General), 2008 FCA 285, 382 NR 370).

[73]           Dr. Motiuk was aware of concerns about cultural bias in assessment tools. CSC does not use its GSIR (General Statistical Information on Recidivism Scale) in respect to Aboriginal offenders for this reason.

[74]           In the context of research, since 2000, the reliability of these assessment tools has been in issue with respect to Aboriginal offenders. The prohibition against the use of the GSIR is an acknowledgement of the existence of cultural bias. Other countries, including the UK, the USA and Australia, have all conducted research to ensure that their psychological assessment tools are reliable in respect to cultural minorities.

F.                  Summary

[75]           The Plaintiff has established that the assessment tools and actuarial tests are susceptible to cultural bias and therefore are unreliable. He has further established that these tests are used in making decisions, and are a contributing factor in decisions that have had an adverse impact on his incarceration.

IV.             Analysis

A.                Issues

[76]           The Plaintiff has framed his case principally as a Charter breach - particularly sections 7 and 15 - caused by the use of these psychological assessment tools.

The Plaintiff has also framed the issue as one of fiduciary duty owed to an Aboriginal offender (or Aboriginal offenders generally), and the breach of that duty.

[77]           With respect, I see this case more simply as firstly a breach of a statutory duty. The Charter and fiduciary duty issues will be addressed for completeness.

B.                 Statutory Duty

[78]           The starting point for this analysis is the “Purpose and Principles” section of the Corrections and Conditional Release Act, SC 1992 c 20 [Act]. Section 3 provides:

[79]           The CSC is governed by certain principles laid down by s 4, of which s 4(g) is especially relevant:

[80]           These principles are given further substance in the later sections. Section 24 is critical to the issues in this case.

[81]           In relying upon questionable tests and in failing to ensure that the tests are reliable, CSC has not taken “all reasonable steps” to ensure that the information about Ewert (or potentially other Aboriginal prisoners) is accurate, up-to-date and as complete as possible.

[82]           It is not necessary, on this issue, for the Plaintiff to establish definitively that the tests are biased; it is sufficient if he raises a reasonable challenge to their reliability pursuant to the aforementioned statutory requirements. The question is whether CSC’s lack of action is sufficient to fulfil the legislated standard of all reasonable steps to ensure accuracy, currency and completeness, particularly in the face of Ewert’s challenge, the actions of other countries and CSC’s own actions in deciding not to use a similar test.

[83]           That question must be answered in the negative. The issue has been a long-standing one; it has not been addressed, and the Defendant’s evidence in this case does nothing to confirm that it has taken the required reasonable steps.

[84]           This is not an issue which CSC missed inadvertently. It has been a live issue since 2000, has been on CSC’s “radar screen”, and the subject of past court decisions where the Court contemplated that some similar type of confirmatory research was being conducted. It is time for the matter to be resolved.

[85]           Therefore, the Court concludes that CSC has failed to meet its statutory obligation under s 24(1) of the Act. The remedy shall address this breach.

C.                 Fiduciary Duty

[86]           Despite Ewert being Aboriginal, CSC does not have an overarching fiduciary duty to him. While the government may be in a fiduciary relationship with its Aboriginal population, that does not equate with a fiduciary duty. This is especially so given the several different obligations owed to others, including the safety of the public mandated by the legislation (Manitoba Metis Federation Inc v Canada (Attorney General), 2013 SCC 14 at paras 46-64, [2013] 1 S.C.R. 623).

D.                Charter – Section 7

[87]           In decisions such as Carter v Canada (Attorney General), 2015 SCC 5, [2015] 1 S.C.R. 331 [Carter], the Supreme Court of Canada has established a two-part test in order to establish a s 7 violation:

1)                  the use [in this case, the assessment tools] deprives the Plaintiff of a s 7 protected right (life, liberty or security of the person); and

2)                  any such deprivation is not in accordance with the principles of fundamental justice.

[88]           Liberty and security of the person are engaged because the assessment tools are used to restrict or deprive the Plaintiff’s liberty. The assessment tools are used by decision makers not just as part of the security classification process by which Ewert can move between more or less internal freedom, but also in decisions regarding release outside prison.

[89]           By negatively impacting ETA requests, which allow a temporary absence from prison, and by making parole virtually impossible to obtain, the Plaintiff’s liberty interests are negatively engaged.

[90]           The Defendant contends that there is no causal connection, or insufficient causal connection, between these deprivations and the use of the assessment tools. However, as held in Canada (Attorney General) v Bedford, 2013 SCC 72, [2013] 3 S.C.R. 1101 [Bedford], the impugned action (the use of the assessment tools) need not be the dominant cause of the prejudice.

[91]           As outlined in the Background, the assessment tool scores played, at the very least, a contributing role in the adverse decisions. One cannot escape that Ewert’s own behaviour, at times, was a contributing cause; however, the assessment tools always played a significant role, either directly or indirectly.

[92]           Nor can one ignore the impact on the Plaintiff’s “security of the person” by being labelled psychopathic. Instances of this labelling and its impact are set out in the Background.

[93]           The Plaintiff argues that the deprivations of s 7 Charter rights are not in accordance with the principles of fundamental justice because the use of the assessment tools violates fundamental justice principles of a) overbreadth; b) arbitrariness; c) the principle of parity or equality; and d) statutory convention/rule of law.

[94]           The reliance on the assessment tools in decision making raises the important fundamental justice principles of overbreadth and arbitrariness.

[95]           As held in Bedford, these two concepts overlap but remain distinct. What they share is the absence of a connection between the objective of the impugned state action and the s 7 deprivation.

[96]           Dr. Motiuk accurately described that the purpose and objective of CSC’s decision making and risk assessment is to “reliably predict an offender’s risk of reoffending as accurately as possible in the interests of public safety”. The purpose and objective is buttressed by s 24(1) of the Act.

E.                 Overbreadth

[97]           In Bedford, the Chief Justice explained overbreadth in these terms:

[112]    Overbreadth deals with a law that is so broad in scope that it includes some conduct that bears no relation to its purpose. In this sense, the law is arbitrary in part.  At its core, overbreadth addresses the situation where there is no rational connection between the purposes of the law and some, but not all, of its impacts.  For instance, the law at issue in Demers required unfit accused to attend repeated review board hearings.  The law was only disconnected from its purpose insofar as it applied to permanently unfit accused; for temporarily unfit accused, the effects were related to the purpose.

[113]    Overbreadth allows courts to recognize that the law is rational in some cases, but that it overreaches in its effect in others. Despite this recognition of the scope of the law as a whole, the focus remains on the individual and whether the effect on the individual is rationally connected to the law’s purpose.  For example, where a law is drawn broadly and targets some conduct that bears no relation to its purpose in order to make enforcement more practical, there is still no connection between the purpose of the law and its effect on the specific individual. Enforcement practicality may be a justification for an overbroad law, to be analyzed under s. 1 of the Charter. [Emphasis in original]

[98]           While there is no evidence that CSC’s use of these tools is unreasonable in respect to non-Aboriginal inmates, their use in respect to Aboriginal inmates overshoots the objective because unreliable tests likely result in unreliable public safety risk assessments.

[99]           The evidence, particularly from Dr. Hart, is that there was no evidence that the scores and the conclusions flowing from the assessment tools (especially the PCL-R) as accurately or reliably predict recidivism in Aboriginal offenders as they do in non-Aboriginal offenders. Even Dr. Rice concurred that there was sufficient reason to conclude that the actuarial scores were less reliable for Aboriginal inmates. She also raised the issue of unreliable Factor 1 scores.

[100]       The Plaintiff has established that PCL-R scores for Aboriginal offenders ought not to be relied upon because:

•                     the likelihood of cross-cultural bias;

•                     the absence of evidence or research that the PCL-R scores are valid for Aboriginals; and

•                     the PCL-R test was developed based on a predominantly Caucasian population and is vulnerable to conceptual variance. Even Dr. Rice’s reliance on PCL-R was based on her assumption that because the PCL-R sampled a wide population, some Aboriginal content was inherent in the test.

[101]       There is no question that the assessment tools were used without qualification or caution despite the long-standing concerns about their reliability.

[102]       Therefore, the continued use of the assessment tools is overbroad of the purpose and objective of the legislation and of CSC’s decision making responsibilities.

F.                  Arbitrariness

[103]       As held in Bedford, arbitrariness arises where there is no connection between the purpose of the impugned state action and the impacts on the individual.

The effect of CSC’s unqualified reliance on the risk assessment tools for Aboriginal inmates when there is no evidence of predictive validity is inconsistent with the objective of accurately predicting risk, and/or is unnecessary to achieve that objective.

[104]       Factor 1 of the PCL-R has been held to be “junk”. Dr. Rice conceded the problems with the reliability of the PCL-R scores.

[105]       In argument, the Defendant acknowledges that the continued use of the PCL-R test in view of the problematic Factor 1 would be arbitrary, should Factor 1 be found to be “junk”. Using infirmed factors upon which to make decisions that affect s 7 rights is clearly arbitrary.

G.                General

[106]       The Plaintiff asks the Court to recognize parity as substantive equality as a principle of fundamental justice under s 7. In my view, it is unnecessary to go that far to resolve this case and do justice.

The Supreme Court declined to do just that in Carter. I will follow their lead.

[107]       The Plaintiff also argues that fundamental justice is infringed because CSC’s use is contrary to s 4 (g) and 24(1) of the Act. They rely on R v Chambers, 2014 YKCA 13 at para 74 for the proposition that the contravention of an express statutory direction may constitute a breach of fundamental justice.

[108]       The Plaintiff may well be correct, especially where the statutory breach does not lead to an adequate remedy. However, that issue, while interesting, is unnecessary to resolve. The remedy which the Court can or will order is the same for statutory breach as for Charter violation.

[109]       With great respect, I find it unnecessary to engage in a s 15 analysis. The facts in this case are not sufficiently developed to usefully engage in the nuanced analysis called for in s 15.

[110]       I would note that the Defendant has not established a s 1 justification for any Charter violation. Applying the Oakes test (R v Oakes, [1986] 1 S.C.R. 103), it is difficult to see how using (at best) questionably reliable assessment tools aids in the pursuit of a pressing and substantive objective. One would think that it is the opposite.

[111]       In like manner, given the problems with the assessment tools, it is difficult to see a rational connection between using questionable evidence and the objective of public safety. The same type of criticism applies with respect to minimal impairment and deleterious effects.

[112]       The Defendant has not made out a case that using flawed data satisfies the s 1 requirements, even if a breach of fundamental justice under s 7 could be sustained under s 1 as consistent with a free and democratic society.

V.                Conclusion

[113]       The Court concludes:

a)                  that the use of these assessment tools is both inconsistent with the principles in s 4(g) of the Act by not being responsive to the special needs of Aboriginal people and further such use breaches s 24(1) of the Act; and further

b)                  violates the Plaintiff’s s 7 Charter rights without s 1 justification.

[114]       The Court intends to issue a final order enjoining the use of the assessment tools in respect of the Plaintiff and other Aboriginal inmates until, at minimum, the Defendant conducts or has conducted a study that confirms the reliability of those tools in respect to adult Aboriginal offenders.

[115]       The Court will issue a Remedies Hearing Order to address the best and fairest manner of implementing the intended final order.

[116]       In the interim, the Defendant shall be enjoined from using the results of the assessment tools in regards to the Plaintiff.

"Michael L. Phelan"

Ottawa, Ontario

September 18, 2015

FEDERAL COURT

SOLICITORS OF RECORD

DOCKET:	T-1350-05
STYLE OF CAUSE:	JEFFREY G. EWERT v HER MAJESTY THE QUEEN IN RIGHT OF CANADA (THE COMMISSIONER OF THE CORRECTIONAL SERVICE OF CANADA, THE WARDEN OF KENT INSTITUTION AND THE WARDEN OF MISSION INSTITUTION)
PLACE OF HEARING:	Vancouver, British Columbia
DATE OF HEARING:	May 25, 26, 28 and 29, 2015 JUNE 1, 2, 4 and 5, 2015
REASONS FOR JUDGMENT:	PHELAN J.
DATED:	september 18, 2015

APPEARANCES:

Jason Gratl Eric Purtzki	For The Plaintiff
Sean Stynes Ainslie Harvey Nicholas Claridge	For The Defendant

SOLICITORS OF RECORD:

Gratl & Company Barristers and Solicitors Vancouver, British Columbia	For The Plaintiff
William F. Pentney Deputy Attorney General of Canada Vancouver, British Columbia	For The Defendant