By: Jennifer Koshan
Case Commented On: Stewart v Elk Valley Coal Corp., 2017 SCC 30 (CanLII)
The Supreme Court of Canada released its decision in Stewart v Elk Valley Coal Corp., 2017 SCC 30 (CanLII) on June 15, 2017. As noted in earlier ABlawg posts on the case (see here and here), the case involves a long-term employee whose job was terminated when, after a minor workplace accident, he tested positive for cocaine and admitted to having consumed the drug while off work a couple of days prior. Elk Valley Coal, the employer, had a policy providing some lenience for employees who disclosed drug or alcohol addictions and sought treatment, failing which its practice was to automatically terminate employment where an employee tested positive for drugs or alcohol following a workplace accident. Stewart did not avail himself of this policy because he did not realize he had an addiction until after the accident. He alleged that his termination amounted to discrimination on the basis of disability contrary to section 7 of the Alberta Human Rights Act, RSA 2000, c A-25.5, and that he had not been reasonably accommodated by Elk Valley. Stewart’s claim was dismissed by the Alberta Human Rights Tribunal, a decision upheld by the Court of Queen’s Bench and a majority of the Alberta Court of Appeal. A majority of the Supreme Court of Canada (per Chief Justice McLachlin) upheld as reasonable the Tribunal decision that there was no discrimination. Justices Moldaver and Wagner disagreed with this conclusion but concurred in the result, finding that the Tribunal was reasonable in concluding that Elk Valley had fulfilled its duty to accommodate. Justice Gascon dissented, characterizing the Tribunal’s decisions on both discrimination and the duty to accommodate as unreasonable.
I find Justice Gascon’s decision most persuasive and most in keeping with a broad, generous approach to interpreting human rights legislation. His remark (at para 59) that drug-dependent persons can “easily be caught in a majoritarian blind spot in the discrimination discourse” was evident in the decisions of the Tribunal and courts below, and in the reasons of the majority of the Supreme Court, as I will elaborate on in this post.
Standard of Review
All members of the Supreme Court agreed that the appropriate standard of review for the Tribunal’s decision was reasonableness. According to the majority, the case law on the test for discrimination and for the duty to accommodate — part of the test for a bona fide occupational requirement (BFOR) — are both settled and were properly cited by the Tribunal (at paras 6, 22). Its task was thus to apply these recognized legal tests to the facts of the case, a task requiring deference from reviewing courts (at para 21).
This aspect of the Court’s decision is interesting for at least two reasons. First, the appropriate standard of review was the subject of much discussion in the courts below, as noted in a post by my colleague Shaun Fluker. The Alberta Court of Queen’s Bench and Court of Appeal both reviewed the Tribunal’s decision on the basis of correctness, in spite of the arguments of the Alberta Human Rights Commission and recent Supreme Court case law to the contrary. The Supreme Court took only four brief paragraphs to discuss the standard of review, suggesting that the approach of the Alberta courts on this issue was clearly wrong.
Second, to the extent the Supreme Court’s decision on standard of review is based on the “settled” state of the case law on discrimination, I would beg to differ. There were marked differences in the articulation of the test for discrimination in the courts below (see Justice Gascon’s decision at paras 69-73), including the issue of whether the complainant must prove arbitrariness or stereotyping on the part of the respondent (see my posts on the earlier decisions here and here). The starkly different applications of the test for discrimination by the majority and dissenting justices of the Supreme Court in Stewart also belie the claim that the test is settled. Moreover, Supreme Court decisions subsequent to the cases that purportedly settled the test for discrimination (Moore and Bombardier, which I will discuss below) continue to cast doubt on the Court’s willingness to apply the test without veering into the language of arbitrariness and other BFOR / justification-stage considerations (see e.g. McCormick v Fasken Martineau DuMoulin LLP, 2014 SCC 39 and my comment on that case here; this is also an issue with the test for discrimination under the Charter (see e.g. here)).
The Test for Discrimination, Applied
So, what is the “settled” test for discrimination, and how was it applied by the various justices at the Supreme Court in assessing the reasonableness of the Tribunal’s finding that there was no discrimination?
The leading cases are Moore v British Columbia (Education), 2012 SCC 61 (CanLII) and Quebec (Commission des droits de la personne et des droits de la jeunesse) v Bombardier Inc. (Bombardier Aerospace Training Center), 2015 SCC 39 (CanLII). According to these decisions, complainants seeking to make out a case of prima facie discrimination must prove three elements: “that they have a characteristic protected from discrimination under the Code; that they experienced adverse impact with respect to the service [or employment]; and that the protected characteristic was a factor in the adverse impact” (Moore at para 33, cited in Stewart at para 24). Justice Gascon helpfully refers to these three elements as “ground”, “harm”, and “contribution” (at para 69).
As noted by Chief Justice McLachlin, Moore and Bombardier also affirm that discrimination can be either direct or indirect. Indirect (or adverse effects) discrimination occurs when facially neutral rules or policies have an adverse impact on individuals or groups in a manner that relates to protected grounds. The implication of recognizing indirect discrimination is that it is not necessary for the complainant to show that the discrimination was intentional; discriminatory effects are sufficient (Stewart at para 24).
The Chief Justice indicated that she could “see no basis to alter the test for prima facie discrimination by adding a fourth requirement of a finding of stereotypical or arbitrary decision-making” (at para 45). Confusingly, she then stated that:
The goal of protecting people from arbitrary or stereotypical treatment or treatment that creates disadvantage by perpetuating prejudice is accomplished by ensuring that there is a link or connection between the protected ground and adverse treatment. (at para 45)
This sentence suggests that arbitrariness and stereotyping are relevant considerations that are encapsulated by the third element of the test for discrimination, the requirement of showing that the protected characteristic was a factor in the adverse treatment (or to use the language of Justice Gascon, that the ground contributed to the harm).
However, Chief Justice McLachlin goes on to say that:
The existence of arbitrariness or stereotyping is not a stand-alone requirement for proving prima facie discrimination. Requiring otherwise would improperly focus on “whether a discriminatory attitude exists, not a discriminatory impact”, the focus of the discrimination inquiry: Quebec (Attorney General) v. A, 2013 SCC 5 (CanLII),  1 S.C.R. 61, at para. 327 (emphasis in original). The Tribunal expressly noted that proof of arbitrariness and stereotyping was not required, at para. 117. (Stewart at para 45)
This passage suggests that a complainant need not show that their employer’s conduct was arbitrary or based on stereotyping, but the majority’s commitment to this point is left ambiguous by paragraph 45, and that ambiguity is reinforced when it actually applies the test.
The majority also clarified that the third element of the test for discrimination — the factor or contribution stage — need not be revised to require a “significant” or “material” link between the complainant’s harm and the ground of discrimination they claim (at para 46). This point is consistent with the Court’s ruling in Bombardier, where it stated that there was no requirement to prove a “causal” link at the third step of the test (Bombardier at para 49, citing Peel Law Assn. v Pieters, 2013 ONCA 396 (CanLII) at para 59). Once again, however, the majority’s adherence to this interpretation of the test is called into question by its application.
How then did the majority apply the test for discrimination to the facts — or more accurately, how did it assess the Tribunal’s application of the test?
Chief Justice McLachlin noted that the first two steps of the test were not at issue. Addiction or drug dependence is recognized as a disability under human rights legislation, so the first element of the test, requiring a protected ground, was made out (at paras 3, 25). Adverse treatment or harm was also clearly made out, because Stewart’s employment was terminated. The only issue was whether the ground was a factor in / contributed to the adverse treatment / harm he suffered.
The Tribunal found that Stewart’s addiction was not a factor is his termination — he “was fired not because he was addicted, but because he had failed to comply with the terms of the Policy, and for no other reason” (at para 26). Put another way, the termination was “due to the failure of Mr. Stewart to stop using drugs and failing to disclose his use prior to the accident” (at para 32, quoting the Tribunal decision at para 120). Furthermore, the Tribunal found that Mr. Stewart had the capacity to comply with the Policy, despite his drug dependence (at para 26). Even though he was in denial about his addiction until the accident, “he chose not to stop his drug use or disclose his drug use” (at para 32, quoting the Tribunal decision at para 122).
The majority of the Supreme Court found that this conclusion — essentially, a conclusion that there was no direct discrimination, based on Elk Valley’s reason for terminating Stewart’s employment — was reasonable (at paras 28, 35). It noted that the Tribunal “unequivocally and repeatedly stated that addiction was not a factor in the decision to terminate Mr. Stewart” (at para 36). But the Tribunal’s use of language is telling:
The Policy as applied to Mr. Stewart which resulted in Mr. Stewart’s termination was not applied due to his disability. (at para 125, cited by the majority at para 36, emphasis mine)
Mr. Stewart was adversely impacted by the Policy not because of denial through drug impairment but rather because he chose not to stop his drug use or disclose his drug use. (at para 122, cited by the majority at para 36, emphasis mine)
The language “due to” and “because of” indicates that the Tribunal sought a causal link between Stewart’s drug dependence and the termination of his employment, which is too stringent an application of the third step of the test for discrimination. It is difficult to see how this was a reasonable application of the test, as the majority held.
The conclusion of Justices Moldaver, Wagner and Gascon that the Tribunal’s decision on discrimination was unreasonable is more attentive to the actual test from Moore and Bombardier. As noted by Justices Moldaver and Wagner, “Mr. Stewart is not required to show that his termination was caused solely or even primarily by his drug dependency. Rather, Mr. Stewart must only show that there is a ‘connection’ between the protected ground — his drug dependency — and the adverse effect” (at para 50). Applying this third factor, they found that “Mr. Stewart’s impaired control over his cocaine use was obviously connected to his termination for testing positive for cocaine after being involved in a workplace accident” (at para 50).
To similar effect is the judgment of Justice Gascon. He clarified that the third, contribution element of the test must focus on the relationship between the ground and the harm, not the ground and the employer’s reasons or intent (at para 85). Stated differently, the proper approach is “whether the ground was ‘a factor’ in the occurrence of the harm, not in the employer’s decision to cause that harm” (at para 115). Applying this approach, “The evidence showed that Mr. Stewart’s addiction had indeed factored into his drug use, and in turn, his violation of the Policy.” The Tribunal’s decision to the contrary was unreasonable (at para 123).
The Tribunal’s finding that Stewart had the capacity to comply with the Policy and chose not to do so is also problematic. As argued by Stewart, denial was part of his addiction and this prevented him from disclosing it prior to the accident (at para 37). According to the majority, however:
In some cases, a person with an addiction may be fully capable of complying with workplace rules. In others, the addiction may effectively deprive a person of the capacity to comply, and the breach of the rule will be inextricably connected with the addiction. … The connection between an addiction and adverse treatment cannot be assumed and must be based on evidence. (at para 39, emphasis added)
The majority thus appears to require a causal link between the ground (addiction-related disability) and harm (termination) rather than a simple contribution. They also call into question whether they understand the nature of addictions, seemingly rejecting the possibility of compromised capacity in favour of a dichotomous capacity / no capacity to choose to use drugs. This construction of addiction is also contradicted by “the Tribunal’s finding that Mr. Stewart was not wholly incapacitated by his addiction” (see Justices Moldaver and Wagner at para 49; see also Justice Gascon at paras 88-89, 102). Does the majority truly accept that addictions constitute a protected ground of discrimination at step one of the test?
The majority’s endorsement of the Tribunal’s reliance on Stewart’s so-called choices was thoroughly critiqued by Justice Gascon, who stated unequivocally that “[a] complainant’s choices are irrelevant to contribution” (at para 97). Reliance on choice is problematic for several reasons, including the burden it places on complainants to avoid discrimination instead of on employers not to discriminate (para 99); its potential to blame marginalized communities for their (bad) choices (at para 101); its focus on conduct rather than identity (at para 100); and its inconsistency with case law rejecting the argument that addicts are responsible for the harms of drug use where laws contribute to those harms (at para 100, citing Canada (Attorney General) v PHS Community Services Society, 2011 SCC 44 (CanLII) at para 106). The Court has also repudiated the notion of choice outside the context of addictions. For example, in Brooks v Canada Safeway Ltd., 1989 CanLII 96 (SCC), the Court rejected the argument that pregnancy was a voluntary condition and therefore not deserving of protection under human rights legislation. More recently, in Quebec (Attorney General) v A, 2013 SCC 5 (CanLII), a majority of the Court declined to follow the view that marital status was a choice that undermined a claim for discrimination under the Charter (at paras 334-343).
Having found that the Tribunal’s dismissal of direct discrimination was reasonable, the majority went on to consider (at para 34) whether the Tribunal’s rejection of indirect / adverse effects discrimination was reasonable. Here, the Tribunal found that “Mr. Stewart would have been fired whether or not he was an addict or a casual user” (at para 123, emphasis added) and that “[t]he Policy as applied to Mr. Stewart which resulted in Mr. Stewart’s termination was not applied due to his disability” (at para 125, emphasis added).
Again, the Tribunal used language suggesting that a causal link focusing on the employer’s reasons was required at the third step of the test for discrimination. This is especially problematic in the case of adverse effects discrimination, where the employer’s intent is typically irrelevant and the focus should be on the effects of the employer’s actions. In addition, comparing addicts to casual drug users incorporates a formal approach to equality by suggesting that the obligation of non-discrimination is satisfied where employers treat all employees the same. As noted by Justice Gascon, this approach fails to recognize that a policy such as Elk Valley’s will have a differential and adverse impact on employees who are drug dependent and whose capacity to comply is diminished (at paras 103-105).
Nevertheless, the majority found the Tribunal’s decision that there was no adverse effects discrimination to be reasonable (at para 35). Its conclusion focuses once again on the employer’s reasons for terminating Stewart (i.e. its intent) rather than the effects of the employer’s application of the Policy to a drug-dependent employee (at para 35). For the majority, “The question, at base, is whether at least one of the reasons for the adverse treatment was the employee’s addiction” (at para 43, emphasis added). Justice Gascon is critical of the majority’s decision on this basis, noting that it “has the impact of erasing indirect discrimination from the scope of human rights protections” (at para 103).
In addition to overlooking adverse effects discrimination, a focus on the employer’s reasons rather than the impact of their rules and policies also risks consideration of arbitrariness and stereotyping at the discrimination stage. Although, as noted earlier, the majority decision attempts to remove these considerations specifically from that stage, its emphasis on the employer’s reasons may allow arbitrariness and stereotyping to creep back in to the analysis of discrimination. Whether an employer acted arbitrarily is properly considered at the BFOR stage of analysis, as correctly noted by Justice Gascon (at paras 106-107).
This risk is manifest in the majority’s closing analogy:
If an employee fails to comply with a workplace policy for a reason related to addiction, the employer would be unable to sanction him in any way, without potentially violating human rights legislation. Again, to take an example given by the majority of the Court of Appeal, if a nicotine-addicted employee violates a workplace policy forbidding smoking in the workplace, no sanction would be possible without discrimination regardless of whether or not that employee had the capacity to comply with the policy. (at para 42)
Besides using floodgates reasoning, and again calling into question the majority’s acceptance that the third element of the test requires mere contribution rather than causation, this analogy implicitly makes the point that it is reasonable or rational for an employer to discipline an employee who breaches a workplace policy, regardless of addiction. But whether such discipline is reasonable or rational is to be assessed at the BFOR stage of analysis, not by importing these considerations into the test for discrimination.
Overall, then, while the majority claims that the test for discrimination is settled, its deference to the application of that test by the Tribunal, along with its own language and reasoning, indicate that several trouble spots remain: the application of the factor / contribution step and whether a causal link is required in practice; the reliance on “choice” as a means of defeating a discrimination claim, a particular concern in cases involving addictions; the focus on the employer’s intent rather than effects and corresponding erasure of adverse effects discrimination; and the endorsement of a formal equality approach to discrimination.
The Test for Reasonable Accommodation, Applied
In light of its finding that the Tribunal had reasonably concluded that discrimination was not established, the majority did not consider the issue of reasonable accommodation (at para 47). This issue is where the concurring and dissenting justices parted company.
All of the justices considering reasonable accommodation agreed that the appropriate test was provided by British Columbia (Public Service Employee Relations Commission) v BCGSEU, 1999 CanLII 652 (SCC) (Meiorin), later affirmed in Hydro-Québec v Syndicat des employées de techniques professionnelles et de bureau d’Hydro-Québec, section locale 2000 (SCFP-FTQ), 2008 SCC 43 (CanLII). An employer may justify discrimination by proving:
(1) that the employer adopted the standard for a purpose rationally connected to the performance of the job;
(2) that the employer adopted the particular standard in an honest and good faith belief that it was necessary to the fulfilment of that legitimate work-related purpose; and
(3) that the standard is reasonably necessary to the accomplishment of that legitimate work-related purpose. To show that the standard is reasonably necessary, it must be demonstrated that it is impossible to accommodate individual employees sharing the characteristics of the claimant without imposing undue hardship upon the employer. (Meiorin at para 54)
This test often turns on the third criterion — whether the respondent satisfied its duty to accommodate. In Stewart, Justices Moldaver and Wagner found that the Tribunal’s decision in the affirmative on this issue was reasonable. Elk Valley’s “no free accident” Policy was intended to have a deterrent effect, and a requirement to accommodate employees with less serious consequences such as a suspension instead of termination would lessen the Policy’s deterrent effect. The Tribunal’s conclusion that individual accommodation would result in undue hardship was therefore seen as reasonable by the concurring justices (at para 55). Justices Moldaver and Wagner also noted the fact that following termination, Elk Valley had given Stewart “the opportunity to apply for employment after six months, provided that he completed a rehabilitation program at a recognized facility” (at para 56).
In contrast, Justice Gascon pointed to the fact that Elk Valley had not conducted any individualized assessment of Mr. Stewart’s circumstances as required by the case law. Indeed, contrary to the terms of the Policy that disciplinary consequences for a positive drug test would be “based on all relevant circumstances”, in practice the consequence was automatic termination of employment (at para 141). In particular, the employer did not consider that the pre-incident accommodation it provided for employees who disclosed their drug dependencies was not accessible to Stewart, who was in denial until the accident (at para 138). Justice Gascon also rejected the post-incident accommodation provided to Stewart (i.e. the ability to re-apply for employment), noting that this post-employment accommodation was inaccessible to Stewart during the relevant time period (at paras 139-140) and that it failed to maintain the seniority and benefits he had received during his nine years with Elk Valley (at para 143). The Tribunal unreasonably failed to consider the deterrent effect of a suspension with pay, a significant penalty short of termination that may not have imposed undue hardship on Elk Valley.
As I indicated at the outset, I find Justice Gascon’s approach to both discrimination and BFOR / reasonable accommodation to be more compelling and more true to the “settled law” on these issues. To the extent the other justices had difficulty in recognizing the Tribunal’s unreasonable application of these tests, this calls into question whether the law is settled, especially as it relates to the test for discrimination. Until the trouble spots I identified earlier are dealt with head-on by a majority of the Supreme Court, employees with addiction-related disabilities may continue to find themselves in Justice Gascon’s “majoritarian blind spot”.
This post may be cited as: Jennifer Koshan “’Majoritarian Blind Spot?’ Drug Dependence and the Protection Against Employment Discrimination” (20 June, 2017), online: ABlawg, https://ablawg.ca/wp-content/uploads/2017/06/Blog_JK_Stewart.pdf
To subscribe to ABlawg by email or RSS feed, please go to https://ablawg.ca
Follow us on Twitter @ABlawg