FWCFB 4075
The attached document replaces the document previously issued with the above code on 19 August 2015.
The references to ‘s.400’ and ‘Section 400(2)’ in paragraph  have been corrected to read ‘s.604’ and ‘Section 604(2)’ respectively.
Associate to Senior Deputy President Hamberger
Dated 24 August 2015.
 FWCFB 4075
FAIR WORK COMMISSION DECISION
Fair Work Act 2009 s.604—Appeal of decision
Construction, Forestry, Mining and Energy Union-Construction and General Division v Port Kembla Coal Terminal Limited (C2015/2695)
Coal export terminals
SENIOR DEPUTY PRESIDENT HAMBERGER
DEPUTY PRESIDENT GOSTENCNIK
SYDNEY, 19 AUGUST 2015
Appeal against decision  FWC 2384 of Commissioner Cambridge at Sydney on 8 April 2015 in matter number C2014/1370 - workplace drug testing comprising urine and oral fluid random testing - whether either method of testing establishes impairment - whether implementation unjust or unreasonable - error by Commissioner established - permission to appeal granted - appeal upheld - on reconsideration of dispute implementation of preferred method of random drug testing in circumstances of the employer not unjust or unreasonable.
 The Construction, Forestry, Mining and Energy Union (CFMEU, the appellant) has sought permission to appeal a decision of Commissioner Cambridge on 8 April 2015 in the matter of Construction, Forestry, Mining and Energy Union v Port Kembla Coal Terminal Limited (C2014/1370). 1
 The Commissioner’s decision dealt with a dispute involving the introduction by the Port Kembla Coal Terminal Limited (PKCT, the respondent) of a workplace drug testing regime as part of an alcohol and other drugs policy (‘AOD Standard’). The respondent wished to introduce drug testing that involves urine sampling. This was opposed by the appellant. During the case, the respondent altered its position from initially proposing sampling by urine only, to the adoption of randomly-selected use of both urine and oral fluid sampling.
 The purpose of the AOD Standard is stated in clause 1.0 of the document:
‘The primary Value at Port Kembla Coal Terminal (PKCT) is ‘We place safety above all else.’ We are committed to achieving an injury free and healthy workplace through encouraging an environment where everyone is committed to working safely. PKCT seeks to reduce at risk individual behaviour and organisational exposure to the potentially harmful consequences of Alcohol and Other Drugs (AOD) in the workplace. Workers impaired by AOD are a safety risk to themselves and all others present at the workplace.’ 2
 The respondent proposes that testing under the AOD Standard take place in accordance with AS 4760-2006 the Australian Standard governing procedures for specimen collection and the detection and quantification of drugs in oral fluid and AS/NZS 4308:2008 the corresponding Australian Standard for urine specimen collection and testing. The threshold concentration limits to be used would be those prescribed by the applicable standard.
 Prior to the hearing of the appeal, both parties were granted permission for legal representation. The applicant was represented by Ms L Doust of counsel. The respondent was represented by Mr J Kirk SC and Mr B Rauf of counsel.
Previous decisions of the Fair Work Commission (the Commission) concerning drug testing
 There have been a number of previous single-Member and Full Bench decisions of the Commission and its predecessor that have touched on the issue of urine versus oral fluid testing.
 On 26 March 2012, Senior Deputy President Hamberger issued a decision that dealt with a range of matters relating to the introduction by Endeavour Energy of a new policy and procedure dealing with drug and alcohol testing in the workplace (Endeavour Energy). 3 The type of testing to be used was one of the issues between the parties. In particular, the employer proposed urine testing and the unions proposed oral fluid testing. The Senior Deputy President considered the evidence and submissions presented and found that the introduction of urine testing by Endeavour Energy would be unjust and unreasonable. He concluded that the appropriate method of drug testing should be through oral fluid and that the testing was to be done in accordance with AS 4760-2006.
 The Senior Deputy President set out his conclusions in the following paragraphs of his decision:
‘ It is clear from all the evidence presented during the hearings that neither oral fluid nor urine testing devices are perfect. Seen from one perspective, urine testing can be seen as more ‘accurate’ in that it is more likely to pick up whether an employee has at some stage taken certain substances. However, that is not necessarily the goal of a workplace drug testing regime. I repeat what I said in Shell Refining (Australia) Pty Ltd v CFMEU ( AIRC 510):
‘ Neither party in this dispute sought to argue that random testing for drugs (or alcohol) was unjust or unreasonable. However both parties also recognise that random testing is an intrusion on the privacy of the individual which can only be justified on health and safety grounds. The employer has a legitimate right (and indeed obligation) to try and eliminate the risk that employees might come to work impaired by drugs or alcohol such that they could pose a risk to health or safety. Beyond that the employer has no right to dictate what drugs or alcohol its employees take in their own time. Indeed, it would be unjust and unreasonable to do so.’
 Based on the evidence presented to me in this case I draw the following conclusions.
 Both methods are susceptible to cheating. For example, cleaning one’s mouth thoroughly after smoking cannabis would minimise the risk of being caught by an oral fluids test. Urine can also be adulterated. There is some evidence that saliva/oral fluid screening is less susceptible to specimen adulteration or substitution compared to urinalysis. In practice however, the likelihood of someone being in a position to cheat effectively when a test is conducted at random and with no prior warning is in my opinion relatively low.
 Australian standards exist governing both methods; and there are laboratories accredited for the analysis of both oral fluid and urine samples. Systems are in place to verify on-site testing devices for both oral fluids and urine.
 Neither method tests directly for impairment. However, a method which tests for recent consumption (only) is more likely to identify someone who is impaired. While some witnesses regard this as a weakness, it is precisely because it only detects for recent use that oral fluid testing is a better indicator of likely impairment as a result of smoking cannabis (the most widely used drug apart from alcohol) than a urine test. Indeed, urine testing may be unable to identify that someone has smoked cannabis in the previous four hours - precisely the time frame which is most relevant for identifying likely impairment.
 Not only is urine testing potentially less capable of identifying someone who is under the influence of cannabis, but it also has the disadvantage that it may show a positive result even though it is several days since the person has smoked the substance. This means that a person may be found to have breached the policy even though their actions were taken in their own time and in no way affect their capacity to do their job safely. In the circumstances where oral fluid testing - which does not have this disadvantage - is readily available, I find that the introduction of urine testing by the applicant would be unjust and unreasonable. Accordingly I find that the system of drug testing that should be used by the applicant for on-site drug testing should be that involving oral fluids. This should be done on the basis of AS4760 - 2006: the Australian Standard governing procedures for specimen collection and the detection and quantitation of drugs in oral fluid.’
 The Senior Deputy President’s decision was appealed to a Full Bench. In its decision, 4 the Full Bench made the following observations:
‘ It was agreed by the parties that this issue and the other disagreements as to the contents of the policy and procedure were to be determined having regard to the principles in the XPT Case. This is stated by the Senior Deputy President at paragraph 6 of his decision.
 The nature of the proceedings before the Senior Deputy President and the evidence and the submissions presented by the parties on the appropriate testing method to be adopted suggest that the determination that was sought from him involved a consideration of the respective merits of urine and oral fluid testing. The question before the Senior Deputy President was therefore not simply whether the introduction of urine testing as a component of the proposed policy would be ‘unjust or unreasonable’ to employees. The Senior Deputy President was called upon to consider which modality of testing ought to be implemented specifically relative to Endeavour Energy. In effect, the Senior Deputy President was asked to determine if it was unjust or unreasonable for the employer to adopt urine testing when there was an alternative suitable method of testing available.
 In arbitration proceedings brought to Fair Work Australia pursuant to s.739 of the Act, it is for the parties to identify the nature of the dispute and the basis upon which it should be arbitrated. Whilst there might in the present case have been greater clarity in this regard, we are satisfied that the approach adopted by the Senior Deputy President was consistent with what was sought by the parties in referring the dispute for arbitration and the determination of the issue as to the appropriate testing method.
 In the appeal, Endeavour Energy in effect sought to re-run the case put to the Senior Deputy President, and for the Full Bench to place different weight on the matters considered and reach a different conclusion. Endeavour Energy also sought to raise additional considerations in support of the adoption of urine testing as part of the drug and alcohol policy. However, it is not the function of an appeal bench in a case such as the present to revisit the facts and circumstances and submissions in order to reach its own conclusions on the merits, except where error has been demonstrated in the decision at first instance. In this case, we are not satisfied, having regard to the principles in House v The King, that there has been shown to be such error in the approach of the Senior Deputy President as would warrant an appeal bench overturning the decision.
 The question of which testing method was to be adopted must be considered having regard to the purpose and aims of the drug testing policy. The proposed Company Procedure on Alcohol and Other Drugs seeks to promote a safe and healthy work environment. The procedure is aimed at (inter alia)
“preventing individuals who may be adversely hindered by alcohol and other drugs from undertaking work or authorised work-related activities, the consequences of which may result in a detrimental effect on health, safety and welfare or other significant aspects of the work environment.”
 In particular the procedure provides at clause 188.8.131.52 that:
“An individual must not work in any Company workplace whilst under the influence of: ...
any illegal drug that is detectable in their system either as the parent drug or metabolite above the prescribed levels described in this procedure ...”
 It was therefore relevant for the Senior Deputy President to give weight to the issue of impairment at work as a result of drug use in considering which testing method should be adopted by Endeavour Energy….
 Given the aim of the random drug testing policy to identify persons who may be impaired, in circumstances where both urine and oral fluid methods have problems in relation to identifying a person as possibly impaired, it was a question of weighing various factors in reaching a conclusion. These included: the potential for urine testing to fail to identify a person as potentially impaired at the time of greatest impairment; the failure of oral fluid testing to identify a person who may still have some residual impairment; the incorrect identification of a person as potentially impaired by reason of a urine test which might be conducted days after the person consumed cannabis; the availability of effective on-site oral fluid testing devices which provide a quick, less offensive, effective and reasonably reliable means of determining whether an employee has used a drug recently and who may therefore not be fit for work; privacy issues, especially in relation to urine testing; and the possibility under the proposed policy that the first positive test may result in a first and final written warning being given to an employee and a second positive result may result in termination of employment.
 Having regard to such matters, it was open to the Senior Deputy President to place significance in reaching his conclusion on the evidence that oral fluid testing rather than urine testing was more likely to detect recent drug use and therefore impairment, and that a positive test result from a urine test might detect drug use at a time which in no way affected their capacity to do their jobs safely.’
 In 2013, Endeavour Energy applied to have the Senior Deputy President’s decision revoked, on the basis that the National Association of Testing Authorities, Australia (NATA) had issued a note that it was not in a position to consider accrediting entities for testing in accordance with AS 4760-2006, Section 3 (On-Site Initial Testing).
 The Senior Deputy President rejected the revocation application. In his decision 5 he noted that:
‘ Neither oral fluid nor urine testing is infallible. Where on-site tests are conducted - whether using oral fluid or urine - confirmatory tests need to be conducted by an appropriate laboratory before any firm conclusions can be drawn about the presence of a drug. However nothing has happened since the original decision and the subsequent appeal in 2012 to indicate that on-site oral fluid testing devices are unreliable. Particularly given NATA’s specific disavowal to the contrary, it would be wrong to infer from NATA’s recent decision to suspend accreditation under Section 3 of AS 4760-2006 that ‘the performance of Oral Fluid devices is suspect, employers should not use Oral Fluid as part of their testing regime’ or that ‘urine is the preferred option’.
 It was understood by the tribunal and the parties at the time of the original proceedings and the subsequent appeal that no facility had been accredited under Section 3 of AS 4760-2006. In its submissions in the original proceedings, Endeavour linked the failure by NATA to have accredited anyone under Section 3 to the lack of any validation process of onsite oral fluid testing devices in AS 4760. 6 The issue of the validation of devices was dealt with in the original decision. At paragraph 39 of that decision, it was noted that systems were in place to verify on-site testing devices for both oral fluids and urine. Dr Vine’s concerns regarding verification of on-site oral fluid testing devices and the consequent lack of accreditation under Section 3 of AS 4760 were noted and in effect rejected, having regard to other evidence that satisfactory systems were in place to ensure effective quality control of devices. The issues raised by Dr Vine in his evidence during these proceedings largely echo the concerns he expressed during the original proceedings.’
 The Senior Deputy President did however agree to make a simple variation to the decision so that it provided that drug testing should be done on the basis of AS 4760-2006 ‘as far as is practicable’. Any on-site testing was to be conducted by a technician as defined in Clause 1.3.38 of that Standard engaged by a collecting agency accredited under Section 2 of that Standard using an appropriate on-site testing device as determined by Endeavour Energy’s service provider in consultation with the accredited laboratory that would be performing the confirmatory testing.
 In Briggs v AWH Pty Ltd, 6 a Full Bench noted that:
‘The issue of whether the most appropriate method of workplace drug testing is by the collection and analysis of a urine sample or a saliva sample has proved to be controversial’.
 While the decision (which concerned an appeal by Mr Briggs against a decision that his dismissal had not been harsh, unjust or unreasonable) went on to outline the nature of the ‘controversy’ it expressly did not attempt its resolution. 7 Rather, the decision turned on whether the employer’s direction that Mr Briggs undertake a urine test was reasonable. This involved a consideration of Mr Briggs’ contract of employment, which required him to comply with his employer’s policies as amended from time to time. The decision noted that, contrary to Mr Briggs’ submission, the relevant policy ‘did not confine itself to testing for impairment from drug use.’ Indeed, the policy expressly recognised that there should be different disciplinary consequences ‘between a mere positive result from a urine test, indicative of drug use, and evidence of actual impairment…’.8
 The decision pointed to evidence that employees who had tested positive to a urine test were not dismissed or sent home; they were allowed to return to the workplace, but were not permitted to operate machinery, and were subsequently required to undergo another test. A number of other factors were found to support the reasonableness of the employer’s direction. This included contractual requirements that a number of clients had imposed on the employer. 9
 Of particular relevance to the current case, the Full Bench observed that:
‘The fact that a saliva test may be better at identifying persons who are at the time of the test likely to be actually impaired, and is more consistent with maintenance of employees’ privacy, may mean that it would be preferred as the more fair and reasonable method of testing in the context of an industrial arbitration (as it was in Shell and Endeavour Energy.’ 10
 In Harbour City Ferries Pty Ltd v Mr Christopher Toms 11 a Full Bench dealt with an appeal from a decision that Mr Toms’ dismissal had been harsh, unjust or unreasonable. Mr Toms had been dismissed following testing positive for marijuana. The Full Bench decision stated:
‘ … We are not persuaded that urine testing, the agreed method of drug testing at Harbour City, is a guide as to the actual presence of marijuana in an employee’s system or any impairment arising as a consequence. It is a testing system which in this case indicated past use and no present impairment.
 Despite our reservations concerning the usefulness of Harbour City’s policy as an effective method of drug detection, when considering leave to appeal and the merits of the appeal, we have identified and considered the misconduct of Mr Toms as his attending work in breach of policy.’
The Commissioner’s decision
 During the proceedings before the Commissioner, evidence was taken from three expert witnesses. The CFMEU called Dr Michael Robertson, a clinical and forensic toxicologist. The employer called Professor MacDonald Christie, Professor of Pharmacology and Associate Dean Research in the Sydney Medical School at the University of Sydney and Mr Peter Simpson, a psychologist and Managing Director of BSS Corporate Psychology Services.
 In his decision, the Commissioner noted that:
‘In recent years there has been significant controversy surrounding what may be described as the debate as to whether urine or oral fluid (aka saliva), was the most appropriate method of sampling for workplace drug testing.’ 12
 The Commissioner, after having referred to some recent arbitral decisions on the subject commented:
‘There is undoubted controversy surrounding the argument as to which workplace drug sampling method is best, urine or oral fluid. The evidence, expert and otherwise, which was presented in this case, has clearly confirmed the ongoing and developing nature of the argument about urine versus oral fluid. In very broad terms, urine can be considered as the more established method for sampling and oral fluid sampling techniques and equipment involve the introduction of new methods and technologies. Further, in general, urine sampling will detect intoxication associated with long-term drug use while oral fluid will enable detection of more acute intoxication associated with recent drug use.
The evidence in this case, as with previous matters such as the Endeavour Decision, has established that each method has certain benefits and shortcomings. Significantly, the identified positives and negatives for each method change over time as each method is impacted by technological and scientific developments. In addition, the efficacy of each method of sampling is impacted by changed social circumstances involving issues such as increasing and decreasing use of particular classes of different drugs. The insidious proliferation of methyl amphetamine use is an inescapable case in point.’ 13
 The Commissioner noted that PKCT was proposing to use both oral fluid and urine as part of its workplace drug testing regime. He stated:
‘Consequently, much of the argument about which is best, urine or oral fluid, becomes academic if both methods are randomly utilized.’ 14
 The Commissioner continued:
‘ The applicant has maintained opposition to the use of urine sampling because of important privacy issues in circumstances where is contends that oral fluid alone provides a sufficient method of sampling to achieve the workplace safety objectives which underpin a drug testing regime. Consequently, in this case the argument has shifted to a cost benefit analysis involving assessment of the combined operation of both methods, oral fluid and urine, versus the positives and negatives of oral fluid alone.
 The expert evidence provided by Dr Robertson included a scientific research paper authored by Lee and Huestis [Current knowledge on cannabinoids in oral fluid’ by Dayong Lee and Marilyn A. Huestis, published in Wiley Online 25 August 2013] (the Lee and Huestis paper). This document also formed Exhibit 9. Although the Lee and Huestis paper focused upon cannabinoids, it also provided some helpful general commentary about the particular benefits of each testing method, oral fluid and urine. For example:
“Therefore, while urine testing is useful for long-term drug monitoring such as in workplace settings, OF testing would be preferable to identify recent drug intake in DUID settings. [Ibid: p.96]”
[Note: OF = Oral Fluid, DUID = Driving Under the Influence of Drugs]
“... because the goals of workplace drug testing are pre-employment screening and drug use deterrents over the course of employment, cut-off criteria that allow long detection windows would be beneficial. In contrast, it is important for DUID and post accident investigations to identify recent drug intake reflecting impairment. [Ibid: p.105]”
 In addition to avoiding what can be described as the identified scientific shortcomings of either sampling method, the random utilisation of both oral fluid and urine sampling provides a superior deterrent against drug use. There are various widely disseminated techniques which can be used to adulterate either an oral fluid or a urine sample. It is unquestionably more difficult to be equipped with adulteration materials and capacity if the method of sampling is unknown. The greater deterrent which is created by the utilisation of both sampling methods was acknowledged by the experts who gave evidence and in particular by Dr Robertson [See in particular, Transcript @ PN251].
 Consequently when the Lee and Huestis paper is included as part of an overall consideration of a workplace drug testing regime where both methods, oral fluid and urine would be utilised, it would appear that the combination of both methods would in general terms provide; (a) long-term drug monitoring benefits, and (b) the identification of more immediate acute drug induced impairment, and (c) a superior deterrent against drug use.
 The Commissioner then turned to what he described as ‘the privacy issues’. In particular he stated that:
‘ …any discomfort or even embarrassment that may be associated with providing a urine sample must be evaluated against important countervailing factors. Importantly, any discomfort or embarrassment about providing a urine sample would be of negligible consequence if such discomfort or embarrassment avoided death or debilitating injury suffered at work. The balance, in my view, would overwhelmingly favour the benefits of adoption of a superior drug detection and deterrent mechanism for the cost of the discomfort, inconvenience or embarrassment of having to provide a urine specimen.
 In addition, in recent years there has been widespread introduction of workplace drug testing regimes which involve urine sampling. Although urine sampling in the workplace could not be described as commonplace it has become increasingly more prevalent particularly in heavy and transport industry sectors. Consequently, there has been a steadily expanding exposure to urine sampling across the broader workforce.’
 Under the heading ‘An Innocent Worker Wronged - Fact or Fiction’ the Commissioner continued:
‘ The other aspect of privacy concern has involved the more extensive information which may be obtained from urine sample results. As was identified in the Endeavour and AWH Decisions, urine sampling when compared to oral fluid sampling, provides greater potential for an employer to obtain information about long-term drug use involving the private activities of an employee at times significantly disconnected from attendance at the workplace. Consequently there is a legitimate basis for concern that an employer would obtain information about the private activities of an employee which it had no right to intrude upon.
 The potential for urine sampling to unreasonably intrude into the private lives of employees has been seen as a realistic basis upon which to reject it as an appropriate method of workplace drug testing when oral fluid was available as a preferable alternative. In the Endeavour Decision it was stated that:
“ Not only is urine testing potentially less capable of identifying someone who is under the influence of cannabis, but it also has the disadvantage that it may show a positive result even though it is several days since the person has smoked the substance. This means that a person may be found to have breached the policy even though their actions were taken in their own time and in no way affect their capacity to do their job safely.” Emphasis added
 Unless the policy that was proposed in the case of Endeavour adopted immunoassay screening test cut-off levels below those set by the relevant Australian Standard, (AS/NZS 4308:2008) I am, with respect, unable to accept that a positive result would “in no way affect” capacity to safely perform work.
 Urine sampling will undoubtedly detect the presence of the metabolite of a “parent” drug over periods of time considerably beyond that for which oral fluid will detect the presence of the “parent” drug or an active derivative of it. Consequently, urine sampling is recognised to have a far more extensive window of detection than oral fluid, such that it is considered to provide detection of a drug at a time considerably after the period of acute intoxication.
 However, detection of a drug, or more accurately the metabolite of it, at or above the cut-off levels fixed by Table 1 of AS/NZS 4308:2008 does not, in my view, translate into the prospect that such detection does not indicate there to be no affect on the capacity of an employee to do their job safely simply because it was detected some considerable time after the drug was imbibed and its acute intoxication had subsided. Most toxicologists are understandably reluctant to proffer any suggestion of alignment of a level of presence of a drug with a particular level of functional impairment. However, the detection of a drug (or its metabolite) at or above the levels set by the relevant Australian Standard for immunoassay screening test cut-off, must represent a measure that can be logically inferred to have some impact on capacity to perform work related functions, irrespective of the time period that may have elapsed since the drug was taken.
 Workplace drug testing regimes are inherently an intrusion into the private lives of employees as they almost always involve the potential for detection of drug use which occurs in a person’s private life. Hopefully not many workers consume illicit drugs at work. It seems to me to be completely irrelevant if one, or four, or more days have elapsed between consumption of the drug and detection of it (or its metabolite) at the workplace. What matters is the detection of the drug at a level which can be reasonably inferred to create a recognised risk to the safety of that employee and others….
 As previously explained, I believe that a test result at or above the relevant Australian Standard cut-off levels must imply in general terms, some potential for or actual impairment which gives rise to a safety risk. There is general acceptance that an oral fluid sample test result at or above the cut-off level for THC set by AS4760-2006 at 25ng/mL (=25ug/L), is an appropriate safety detection trigger for workplace drug testing regimes. I am unable to understand why a urine sample result at or above the cut-off level for Cannabis metabolites set by AS/NZS 4308:2008 at 50ug/L would be considered to “in no way affect” capacity to perform work safely.
 Further, it is important to consider the scientific research which has been conducted into the long-term effects of regular cannabis consumption. In this case the Lee and Huestis paper added to the body of material which supports the concerns that were persuasively expressed by Professor Christie about the inadequate recognition of the effects of long-term cannabis use and what he believed to be the need to reduce the cut-off levels for THC and its predominant metabolite, 11-nor-9-carboxy-THC (THCCOOH) in the relevant Australian Standards. Relevantly, the Lee and Huestis paper included the following:
“In chronic cannabis smokers during abstinence, low THC concentrations were detected in blood for up to 30 days, and psychomotor performance in tasks validated to predict on-the-road impairment remained impaired compared to occasional smokers for 21 days. In other studies, neurocognitive performance improved over 30 days in chronic frequent cannabis smokers, but was still impaired compared to occasional smokers for 7-28 days.” [‘Current knowledge on cannabinoids in oral fluid’ by Dayong Lee and Marilyn A. Huestis, published in Wiley Online 25 August 2013]
 There is compelling scientific evidence to conclude that the detection of cannabis (specifically THCCOOH) by way of urine sampling at levels at or above the Australian Standard AS/NZS 4308:2008 immunoassay screening cut-off level of 50 ug/L, at extensive time intervals (days or even weeks) after cannabis was consumed, represents valid and appropriate identification of a safety risk. Chronic and even occasional cannabis users can be intoxicated for considerable periods after they have stopped taking the drug.
 In summary, detection of cannabis metabolites and other drugs at or above the immunoassay screening cut-off levels established by Table 1 of AS/NZS 4308:2008 can logically be translated into a safety risk that requires action. The capacity for such detection should not be avoided upon the erroneous proposition that an innocent worker may be subjected to an unreasonable intrusion into their private lives. Detection of the drug at or above the cut-off level expunges innocence.
Both Better Than Either
 Although I am unable to accept the validity of the privacy concerns advanced as opposition to urine sampling, it must be recognised that oral fluid sampling has considerable benefits over urine sampling particularly in respect to its enhanced capacity to identify immediate acute intoxication which may not be detected by urine sampling. Consequently, if presented with an “either or scenario” oral fluid sampling would probably represent, on balance, a preferable option to urine sampling.
 As previously mentioned, the circumstances of this case did not involve an “either or scenario.” The employer has sought to, in effect, add urine sampling to its existing oral fluid sampling. Although there is an absence of any legitimate privacy concerns upon which to reject the addition of urine sampling, it is also necessary to briefly recognise the additional benefits that are derived from urine sampling.
Some Particular but Important Shortcomings of Oral Fluid
 The current level of technology does not enable oral fluid sampling devices to adequately detect for the presence of benzodiazepines. In recent years there has been a fairly rapid improvement in the specificity and sensitivity of oral fluid sampling devices and there may be, in the future, capacity for oral fluid detection of benzodiazepines. However, at the present time, a workplace drug testing regime without urine sampling will essentially fail to detect the presence of benzodiazepines at onsite screening. It must be recognised that benzodiazepines do not represent one of the more significant drugs of concern in respect to workplace safety but nevertheless it would be preferable to have a regime which included their detection as part of onsite screening.
Long-Term Drug Use - “Coming off Meth” as But One Example
 As mentioned earlier in the Decision, oral fluid sampling will not adequately detect long-term cannabis use. In something of a reverse scenario to urine sampling which may not detect recent consumption of THC, oral fluid sampling is unlikely to detect levels of THCCOOH associated with long-term cannabis use.
 In addition, the expert evidence confirmed that oral fluid sampling was an inferior means to detect long-term use of other drugs such as opioids, cocaine and amphetamine related psycho stimulants. The wider window of detection was one of the primary aspects of the opposition to urine testing.
 However, it is the wider window of detection which enables identification of long-term drug use, (via levels fixed by AS/NZS 4308:2008). Any suggestion that this is detection without relevant safety implications is further dispelled by evidence about the “hangover” effects of drugs like methylamphetamine. As just one example, the evidence of the physiological and psychological impacts of withdrawal from methylamphetamine provides compelling basis to detect long-term drug use.
THC Eaten Rather than Smoked
 Further, oral fluid sampling is unlikely to detect THC which was eaten rather than smoked. In a situation which involved only oral fluid drug testing, a chronic cannabis user could conceivably avoid detection by ensuring that he or she only smoked cannabis at times that were sufficiently before commencement of work, and perhaps ate substances containing THC at times likely to be closer to working time.
Other Important Components of a Drug Testing Regime
 There are aspects of any workplace drug testing regime other than the method of sampling which are important and which impact upon the issue of whether oral fluid or urine or both sampling methods, should be found to be reasonable and appropriate. It would be unrealistic to attempt to codify workplace drug testing by way of any universal rules. Workplaces have different safety risks. For example, it would seem to be largely unnecessary to implement a workplace drug testing regime in the case of a call centre. On the other hand, heavy and transport industries obviously require workplace drug testing.
 In workplaces where occupational and public safety risks are present, drug and alcohol testing regimes are mechanisms which improve safety for workers and the general public. Individuals who attend these “high risk” workplaces under the influence of drugs or alcohol, at a level of recognised impairment, are likely to endanger the lives of others. Workplace drug testing, if properly conducted and policed, should not be misconceived as an invasive and punitive threat to the welfare of workers.
 However, the apprehension that employees often have about drug and alcohol testing regimes is understandable. In particular, the identification of the use of illicit drugs or disproportionate and unsympathetic disciplinary reactions to positive and confirmatory test results, naturally creates concerns in the mind of some workers. As a matter of general practice, drug or alcohol addiction or abuse issues which have been identified through workplace testing, should be recognised as problems that require a treatment program and not necessarily disciplinary action.
 The particular facts and circumstances of each case of drug detection in the workplace need to be carefully assessed and judged accordingly. Importantly there should be no automatic or prescribed approach to any consequent disciplinary action. In this instance the AOD Standard includes a number of important, commendable components such as:
● voluntary self testing
● assistance to an employee who commits to a recognised rehabilitation program
● no automatic or prescribed disciplinary action
● a case management approach to any positive confirmatory result’
 The Commissioner concluded:
‘ … the benefits that would be obtained by the adoption of both methods of sampling in random combination significantly outweigh any privacy detriments that could be identified.
 There are a range of important benefits that are derived from the random operation of both oral fluid and urine sampling. The use of both methods overcomes the scientific and technological deficiencies that each method cannot avoid if one method is used in isolation. Further, the use of both methods provides significantly enhanced deterrent properties. Against these significant attributes the alleged privacy intrusions are matters of little realistic consequence.
 In summary, a blunt distillation of the contest in this case and its determination can be described as a choice between private lives or saving lives and I have opted for saving lives.’
 The Commissioner accordingly dismissed the application to enable PKCT to introduce its preferred method of random drug testing.
 Under s.604 of the Fair Work Act 2009 (the FW Act), appeals require the permission of the Commission. Section 604(2) of the FW Act provides:
‘Without limiting when the FWC may grant permission, the FWC must grant permission if the FWC is satisfied that it is in the public interest to do so.’
 Factors that might invoke the public interest have been held to include where a matter raises issues of importance and general application, where there is a diversity of decisions at first instance so that guidance from an appellate court is required, or where the decision at first instance manifests an injustice, or the result is counterintuitive, or the legal principles applied appear disharmonious when compared with other recent decisions dealing with similar matters. 15
 The parties agree that the Commissioner’s decision involved the exercise of a significant level of discretion, in the sense that the decision involved the weighing of a number of factors, and in which no one particular outcome was prescribed. 16
 The principles relevant to an appeal from such a decision are those in House v The King:
‘The manner in which an appeal against an exercise of discretion should be determined is governed by established principles. It is not enough that the judges composing the appellate court consider that, if they had been in the position of the primary judge, they would have taken a different course. It must appear that some error has been made in exercising the discretion. If the judge acts upon a wrong principle, if he allows extraneous or irrelevant matters to guide or affect him, if he mistakes the facts, if he does not take into account some material consideration, then his determination should be reviewed and the appellate court may exercise its own discretion in substitution for his if it has the materials for doing so. It may not appear how the primary judge has reached the result embodied in his order, but, if upon the facts it is unreasonable or plainly unjust, the appellate court may infer that in some way there has been a failure properly to exercise the discretion which the law reposes in the court of first instance. In such a case, although the nature of the error may not be discoverable, the exercise of the discretion is reviewed on the ground that a substantial wrong has in fact occurred.’ 17
 The appellant raised three grounds of appeal. The first ground was that:
‘The Commissioner mistook the facts in positing a definite relationship between the Alcohol and Other Drugs (“AOD”) testing detection levels in AS/NZS4308:2008 and physical impairment or “intoxication” on the part of employees, contrary to the expert evidence in the proceedings.’
 In particular, the appellant submitted that the Commissioner was in error in his finding (at ) that
‘… the detection of a drug (or its metabolite) at or above the levels set by the relevant standard for immunoassay screening test cut-off, must represent a measure that can be logically inferred to have some impact on capacity to perform work related functions, irrespective of the time period that may have elapsed since the drug was taken.’
 The appellant also took issue with the Commissioner’s conclusion (at ) that a test result at or above the relevant Australian Standard cut-off levels must imply, in general terms, some potential for or actual impairment which gives rise to a safety risk.
 The second ground of appeal was that:
‘The Commissioner mistook the facts in finding that an employee who returned a positive test result as part of an AOD procedure based on the cut-off levels contained in AS/NZS 4308:2008 represents a “safety risk that requires action”.’
 In particular, the appellant submitted that the Commissioner was in error (at ) when he asserted a link between detection levels contained in AS/NZS 4308:2008 (the urine testing standard) and a risk to health and safety in the workplace when there was no, or no proper, evidentiary basis.
 The third ground of appeal was that the Commissioner had erred in failing to conclude that the respondent could only legitimately require employees to submit to an AOD testing regime that was directed to the detection of impairment on the part of employees whilst present at work.
 In particular, the appellant submitted that the AOD procedure of the respondent impermissibly intrudes into the legitimate rights of employees by seeking to regulate private conduct that is not demonstrated to compromise safety at work. The use by PKCT of urine testing as part of its AOD testing procedure would necessarily result in ‘positive’ test results that do not provide any reliable information demonstrating that an employee was impaired during the performance of their work duties.
 The appellant submitted that it was in the public interest for the Commission to grant permission for the appeal. First, it submitted that the appeal deals with an issue of significant controversy relating to the meaning of AS/NZS 4308:2008. This controversy is of general public interest and extends beyond the specific interests of the appellant and the respondent.
 Secondly, the appellant submitted that the decision of the Commissioner was inconsistent with other recent decisions of the Commission dealing with the subject matter.
 Thirdly, the appeal raises the important question of the legitimate scope of employer direction of, or interest in, the private activities of employees where there is no evidence to demonstrate that such activity or behaviour results in risks to health and safety at work.
Consideration of the appeal
 The first two grounds of appeal in essence allege that the Commissioner mistook certain key facts about the expert evidence. The first of these alleged errors was that he was mistaken in positing a definite relationship between a ‘positive’ test result in AS/NZS 4308:2008 and physical impairment or ‘intoxication’.
 At paragraph  of his decision, the Commissioner stated:
‘…in general, urine sampling will detect intoxication associated with long-term drug use while oral fluid will enable detection of more acute intoxication associated with recent drug use.’ (emphasis added)
 This is, with respect, not an accurate summary of the expert evidence. Neither urine nor oral fluid testing can detect ‘intoxication’ or impairment. Dr Robertson said (at PN219):
‘Neither urine nor oral fluid testing will detect impairment…’
 Dr Robertson was not challenged on this evidence. Dr Christie also agreed that drug tests ‘do not directly test for the immediate intoxicating effects of drugs…’ (or for any ‘hang-over’ effect from the use of drugs.) 18
 Dr Robertson’s evidence was that:
‘…whilst impairment cannot be inferred from the results of either urine or oral fluid, it is clearly demonstrated that relative to urine, oral fluid better reflects the presence of drug in the blood stream and therefore is a better indicator of recent drug use and therefore possible impairment. Given oral fluid reflects the presence or absence of drugs in the blood stream, like blood, the window of detection of drugs is therefore shorter (relative to urine) however this should not be seen as a negative feature of oral fluid testing but rather that when a sample is found to contain drugs i.e. ‘positive’, there is a greater likelihood that the individual may be impaired relative to a ‘positive’ result in urine and when no drug is detected this would suggest no use of the drug in the day or days preceding the test and therefore the likely absence of impairment.’ 19
 It is clear from the expert evidence that a ‘positive’ test result merely indicates the presence of a substance, or its (inactive) metabolite. A positive urine test (when confirmed in a laboratory) indicates that the subject has used a particular drug at some point in the past. This may have been some days earlier. In contrast to the Commissioner’s assertion, it does not tell you that the subject was impaired (‘intoxicated’) when the sample was taken, nor is it evidence that he or she is a long-term drug user.
 Likewise, the Commissioner was in error when he stated (at paragraph ):
‘…the detection of a drug (or its metabolite) at or above the levels set by the Australian Standard for immunoassay screening test cut-off, must represent a measure that can be logically inferred to have some impact on capacity to perform work related functions, irrespective of the time period that may have elapsed since the drug was taken.’
 There are at least two problems with this statement, in the context of the expert evidence presented during the case. First, ‘immunoassay’ testing refers to the initial site screening. The expert evidence was that immunoassay testing indicates that a drug (or a by-product of a drug) might be present. It is only when a test is conducted in a laboratory (using a mass spectrometer) that the presence of a drug (or its by-product) can be confirmed.
 More importantly, the expert evidence was that a positive test using urine (which detects the presence of inactive metabolites) merely shows that that ‘an individual has used the drugs in the preceding hours, days or weeks, rather than whether they may be under the influence of or impaired by a drug.’ 20
 It is not in dispute that the impairing effects of drugs wear off after a period of time. Dr Robertson’s evidence was that because urine testing ‘has a greater window of detection’ (than oral fluid testing) ‘a ‘positive’ result may be consistent with recent ingestion and associated impairment or use in the day or days prior to collection and long after impairment has subsided.’(emphasis added) 21
 Dr Christie also acknowledged in his evidence that urine testing ‘is considered to suffer the disadvantage that it detects use for considerably longer than the period of impairment.’ 22 His evidence was that the period of impairment for cannabis is up to six hours.23
 In other words, the expert witnesses for both the appellant and the respondent agreed that a positive urine test result does not lead to a ‘logical inference’ that there must have been ‘some impact on capacity to perform work related functions, irrespective of the time period that may have elapsed since the drug was taken.’
 It is clear from a reading of the Commissioner’s decision that his erroneous finding that a positive urine result must be associated with a level of impairment was central to his conclusion in favour of the respondent’s case. This is an error of the type envisaged by House v the King.
 In the circumstances, it is not necessary to consider the second and third grounds of appeal.
 The issues raised by the Commissioner’s decision – especially his findings about the relationship between positive drug tests using urine samples and impairment – are inconsistent with the conclusions upheld by the Full Bench in Endeavour Energy. The issue of oral fluid versus urine testing is of interest to parties beyond those concerned in this case. In these circumstances we consider that it is in the public interest to grant permission to appeal.
 Given the finding of error, we have decided to allow the appeal and set aside the Commissioner’s decision. There is sufficient material before us to reconsider the matter ourselves.
 There was some disagreement between the parties about the precise nature of the dispute that the Commission has been asked to resolve. We consider that the Commission has to determine whether it would be unjust or unreasonable for PKCT to adopt its preferred method for random drug testing. That method was outlined in the statement of Mr Calder, the respondent’s Health and Safety Specialist. 24
 Under the respondent’s preferred approach, the method of testing (urine or oral fluid) for drugs other than alcohol would be randomly selected by the external testing provider for the day of testing. If the random method of testing was urine then all workers randomly selected on the day of testing would undertake a urine test. Similarly, if the random method of testing was oral fluid then all workers would undertake an oral fluid test.
 As noted by the Full Bench in Endeavour Energy, the question of which testing method is to be adopted must be considered having regard to the purpose and aims of the drug testing policy.
 The AOD Standard includes the following under the heading ‘Purpose’:
‘PKCT seeks to reduce at risk individual behaviour and organisational exposure to the potentially harmful consequences of Alcohol and Other Drugs (AOD) in the workplace. Workers impaired by AOD are a safety risk to themselves and all others present at the workplace.’
 It is reasonable to infer therefore that the purpose of the policy is to reduce the risk that workers will attend the workplace impaired by alcohol and other drugs. This would primarily be achieved through deterrence (as workers are not tested every day).
 The evidence is clear that oral fluid testing has a window of detection of a few hours (depending on the equipment used and the cut-off level adopted). Urine testing, on the other hand, has a longer window of detection. Neither method tests for impairment (or for ‘hangover effects’ or long term use). However, because its window of detection more closely approximates the likely period of impairment compared with urine a ‘positive’ oral fluid test result is more likely to be associated with impairment than a ‘positive’ urine test result. The main disadvantage of urine testing is that a worker who tests ‘positive’ may have taken the drug some time ago and no longer be impaired.
 Both methods are susceptible to cheating. 25 Moreover, neither method can be used to detect all drugs that may potentially be of interest.26
 Previous Full Bench decisions have upheld decisions by single Members to the effect that the use of urine testing as opposed to oral fluid testing is unjust and unreasonable. However, the issue in the current case is whether a system which uses both methods (selected at random) has advantages that outweigh the privacy concerns that are raised by the use of urine testing on its own.
 On balance, and in the particular circumstances faced by the respondent, we are satisfied that it would not be unjust or unreasonable for PKCT to implement its preferred approach to random drug testing.
 We have reached this conclusion for a number of reasons. First, it needs to be emphasised that the policy concerns a random testing regime. Whichever method of drug testing is adopted, employees attending for work will often not be tested. This means that some employees might be impaired by drugs or alcohol and not be detected through testing. The real purpose of random testing is therefore to deter employees from attending work in an impaired state because of the risk that they might be detected.
 The appellant’s own expert witness agreed that a system where workers would not know which type of drug testing method might be used would enhance the deterrent value of the testing. 27 In particular, it would be significantly more difficult for a worker to take measures to avoid detection. An additional benefit is that there is scope to test for a wider range of drugs if both methods of testing were to be used. This also adds to the deterrent value.
 An additional purpose of random testing is to detect drug use by employees in order to enable PKCT to reduce and manage workplace risks associated with drug use. As we have already stated, neither test establishes functional impairment caused by drug use.
 PKCT has a statutory duty to ensure, so far as it is reasonably practicable, the safety of its employees and contractors who might be put at risk by work that is being carried out. An essential element of this duty involves the identification of potential hazards and elimination or minimisation of risks. It seems to us that PKCT’s AOD Standard and its preferred drug testing regime is part of the method employed by PKCT to discharge this duty. Having regard to the high-risk nature of the work undertaken at the Port Kembla coal terminal by employees, the privacy concerns about urine testing must therefore give way to allow the implementation of a testing method which will enable PKCT to identify and manage workplace safety risks.
 We have also taken into account two other factors. One is Mr Calder’s uncontested evidence is that most of the respondent’s shareholder entities and other Australian coal export terminals use urine-based drug testing. 28
 Finally, we have given significant weight to the way in which PKCT has indicated it will use non-negative test results. In particular a case management approach will be adopted, which will have regard to the circumstances of individual workers. While acknowledging that in some circumstances a non-negative result could lead to disciplinary action, other outcomes could include rehabilitation, counselling, participation in the Employee Assistance Program, scheduled testing and the development of a return to work plan.
 As we have indicated, PKCT is obliged to ensure, so far as is reasonably practicable, the health and safety of its employees and contractors while they are at work. This means, inter alia, that PKCT must try to eliminate (and where this is not practicable, to minimise) the risk that employees might come to work impaired by drugs or alcohol and so pose a risk to health and safety. PKCT is certainly entitled to implement a system of random drug testing to assist it in discharging its obligation.
 Random drug testing inevitably involves a degree of intrusion by an employer into the private lives of its employees. While neither method is fool-proof, the evidence indicates that oral fluid testing will generally identify employees who have recently consumed a drug and are therefore likely to be impaired. Urine testing will identify whether an employee has taken a drug in the preceding days or even weeks – including at times when there is no serious risk that the employee will still be impaired when they attend for work. While there are privacy concerns with urine testing, we consider that in the particular circumstances of PKCT, it would not be unjust or unreasonable for PKCT to implement its proposed AOD Standard and associated testing method.
 The dispute is determined accordingly.
SENIOR DEPUTY PRESIDENT
L Doust of counsel with A Thomas for the Construction, Forestry, Mining and Energy Union.
J Kirk SC with B Rauf of counsel for Port Kembla Coal Terminal Limited.
2015. 16 June.Sydney.
1  FWC 2384.
3 Endeavour Energy v CEPU and ors.  FWA 1809.
4 Endeavour Energy v CEPU and ors.  FWAFB 4998.
5  FWC 198.
6  FWCFB 3316.
7 Ibid at .
8 Ibid at .
9 Ibid at  and .
10 Ibid at .
11  FWCFB 6249.
12 At .
13 At -.
14 At .
15 GlaxoSmithKline Australia Pty Ltd v Makin (2010) 197 IR 266.
16 See, for example, the appellant’s outline of submissions, paragraph 23.
17 House v The King (1936) 55 CLR 499.
19 Exhibit 1, page 3.
20 Exhibit 1, page 4.
21 Exhibit 2, page 3.
22 Exhibit 4, page 9.
23 Exhibit 4, page 4.
24 Exhibit 14, paragraphs 29-30.
26 Exhibit 2, page 9.
28 Exhibit 14.