The leading Supreme Court of Canada case on the issue of privacy concerns relating to random drug and alcohol testing in the workplace is the 2013 case Irving Pulp & Paper Ltd v CEP, Local 30 (“Irving”), which found that random drug and alcohol testing in the workplace is a violation of employees’ privacy rights, and is only justifiable in situations where there is:

  1. A dangerous workplace; and
  2. Enhanced safety risks, such as evidence of a general problem with substance abuse in the workplace.1

The question of what is required for an employer to demonstrate “a general problem of substance abuse in the workplace” featured prominently in the Alberta case Suncor Energy Inc v Unifor Local 707A (“Unifor”).

The drug and alcohol policy at issue in Unifor required employees at Suncor’s oil sands operations working in safety-sensitive positions to undergo random drug and alcohol testing. In attempting to justify its random drug and alcohol testing policy, Suncor presented evidence to the arbitration board highlighting issues with substance abuse at their workplace. Unifor argued that the evidence was not sufficient to justify the random drug and alcohol testing, considering the competing privacy rights of the bargaining unit members.

Suncor had approximately 10,000 workers on site at any given time, of which 3,383 employees were unionized workers represented by Unifor, 2,963 were non-unionized workers, and the rest were contract employees. The tribunal found that it could only consider the “cost/benefit analysis as it relates to this bargaining unit, and not its impact vis-à-vis the other two thirds of workers.” The tribunal wanted to avoid the situation where “a group of employees could have their privacy invaded whether or not they pose any sufficient risk to an employer themselves — in order to gain access to require random testing of another group of employees.”2

The arbitration board was not able to conclude that Suncor’s evidence related specifically to members of the bargaining unit (as opposed to the non-unionized employees), and so it excluded much of Suncor’s evidence. The tribunal ultimately held that Suncor failed to establish “that there is a significant problem, or legitimate safety risk, with respect to the use of alcohol”.3

On judicial review, the reviewing justice held that the tribunal should not have excluded Suncor’s evidence on the basis that it was not specific to the members of the bargaining unit. He found that the labour tribunal’s decision was unreasonable because the tribunal had applied the test of whether there was a “significant” problem, rather than a “general” problem, with substance abuse in the workplace.4 This decision was appealed to the Alberta Court of Appeal.

In dismissing the appeal, the Court of Appeal found that by excluding much of Suncor’s evidence of substance abuse in the workplace, the labour tribunal had improperly narrowed the scope of the test by considering only substance abuse in the bargaining unit. The key question in this case was whether there was sufficient evidence of a substance abuse problem in the workplace to justify random drug and alcohol testing.5 While in some circumstances it may be reasonable to distinguish between groups of employees for the purposes of determining whether there is a problem of substance abuse in a workplace, in this case the different employees worked together in an integrated workforce. The Court of Appeal held that by unreasonably narrowing the evidence it considered, the arbitration board effectively asked the wrong question and applied the wrong legal test.6

The Court of Appeal’s decision made it clear that the Irving test is not to be interpreted more strictly or narrowly than it was framed by the Supreme Court of Canada, and remitted the matter back for consideration by a fresh panel. This decision sends the signal to employers that despite the issues relating to employee privacy, justification of random drug and alcohol testing policies is possible in the right circumstances. Whether it is justified in this case remains to be seen.