‘Zero Tolerance’ Drug Driving Laws in Australia: A Gap Between Rationale and Form?

Legislation in all Australian states and territories creates offences and provides for police roadside testing in relation to ‘drug driving’. Ostensibly motivated by the same road safety objectives and impairment paradigm as drink driving laws, drug driving laws adopt a significantly different approach. Whereas random breath testing tests for all forms of alcohol and is designed to determine whether there is a sufficient concentration of alcohol in the driver’s body that s/he should be deemed to be impaired, random drug testing typically tests for the presence of any quantity of only the three most widely used illicit drugs—cannabis, methamphetamine and ecstasy—in the driver’s oral fluids, without reference to what is known about the different pharmacokinetic and pharmacodynamic qualities of different drugs. This article examines this idiosyncratic approach to the criminalisation of drug driving, highlighting its weak correlation with the important road safety objective of deterring substance‐impaired driving, and the risks of both over‐ and under‐criminalisation that it creates. It argues that public policy on the prohibition of certain drugs and the criminalisation of their use should be disentangled from public policy on impaired driving. It recommends that drug driving laws in all Australian jurisdictions should be brought back into line with drink driving laws, via legislation and testing practices that turn on substance‐specific prescribed concentrations for all drugs (illicit and licit) that have the potential to impair drivers.


Introduction
In all Australian states and territories there has been a renewed interest in drug driving offences in recent years. Even though such offences have been on the statute books since 1930, they have been modified over the last 15 years in ways that significantly expand the parameters of the category 'drug driving'. This definitional expansion-characterised by a movement away from an impairment/'under the influence' paradigm in favour of testing for the presence of any quantity of selected illicit drugs-has been combined with a significant escalation of police detection and enforcement activity. For example, in 2015 the Victorian Government announced a major investment in drug driving detection, including plans to conduct 100,000 roadside tests annually (Andrews 2015). In Western Australia, from 2014 to 2015, drug driving convictions increased by almost 300 per cent (Hickey 2016).
Given the considerable change and expansion of both the substantive offences of drug driving and enforcement practices in Australia since the early 2000s, it is unsurprising that drug driving laws and police roadside testing practices have begun to attract media attention (Butler 2015;Gulbin 2016;Visentin 2016;Wodak and McDonald 2015), as well as closer scrutiny by legal practitioners (Barns 2016;Lawrence and Zhou 2016;Redfern Legal Centre 2016). Judicial officers have also been required to confront challenges posed by testing practices for the assessment of responsibility and punishment (Bugden; Halper v R [2015] NSWDC 346; Police v Cahil (unreported, Lismore Local Court, 9 October 2014); Police v Carrall (unreported, Lismore Local Court, 1 Feb 2016)).
More surprising is the fact that drug driving laws have attracted relatively little scholarly attention in Australia. Road safety experts and criminologists have considered prevalence and deterrence questions (for example, Adams, Smith and Hind 2008;Armstrong, Watling and Davey 2014;Jones, Donnelly, Swift and Weatherburn 2005;Ramsey and Fitzgerald 2017;Woolley and Baldock 2009), but criminal law and criminalisation scholars in Australia have been slow to engage. 2 The latest edition of the most widely read Australian textbook on driving offences by D Brown (2006) was written more than 10 years ago, prior to the nation-wide adoption and increasingly heavy enforcement of 'zero tolerance' drug driving laws and associated testing.
Our own interest in the topic of drug driving laws has been triggered by a number of factors: first, our involvement in an ongoing study of the concept of 'intoxication' in Australian criminal laws (McNamara, Quilter, Seear and Room forthcoming;Quilter and McNamara forthcoming;Room 2016a, 2016b), including the realisation that criminal laws frequently fail to distinguish between the different effects of different drugs (Quilter et al. 2016b: 47-48); secondly, a commitment to addressing those criminal offences that have traditionally been regarded as too 'trivial' (McBarnet 1981) to attract scholarly attention, despite their high volume enforcement and prosecution (see McNamara andQuilter 2014, 2015;Quilter and McNamara 2013); and, thirdly, an interest in the politics of criminalisation as a public policy tool (McNamara 2015;McNamara and Quilter 2016), including in relation to alcohol-and drug-related harms (Quilter 2014(Quilter , 2015. That the goals of improving road safety and reducing road accidents and trauma are valid and important is axiomatic. Accordingly, and alongside other measures, the detection of individuals who are driving under the influence of substances that are known to diminish driver capacity is entirely meritorious. The aim of this article is not to question the legitimacy of these objectives, but to scrutinise and produce fresh insights about the way that the criminal law is configured and deployed to this end. Specifically, it investigates whether current drug driving laws and random drug testing (RDT) practices are consistent with the evidence-based impairment paradigm that has underpinned the success of random breath testing (RBT) and drink driving offences in transforming driving under the influence of alcohol from a common practice to a highly stigmatised criminal behaviour (Terer and Brown 2014). By producing new knowledge about the

The early drug driving offences
In 1930 the first drug driving offence was implemented with the enactment of the Traffic Act Amendment Act 1930 (WA). This Act repealed and replaced s 27 of the Traffic Act 1919 (WA) to expand the offence to include 'drugs', and to effect the move from the 'under the influence' model to the 'incapacity' model: 27 (1) Any person who, when driving or attempting to drive, or when in charge of a motor vehicle in motion on a road, or when in charge of a horse or other animal or drove of animals on a road, is under the influence of drink or drugs to such an extent as to be incapable of having proper control of the vehicle or the horse or other animal or drove of animals, shall be guilty of an offence under this Act. South Australia and NSW followed suit during the 1930s (Road Traffic Amendment Act 1936 (SA); Motor Traffic (Amendment) Act 1937 (NSW)). Equivalent laws were not introduced elsewhere until considerably later: 1949in Victoria, 1952in the Northern Territory (NT), 1961in Tasmania, and 1971 In South Australia, Tasmania, and Western Australia 'drug' was left undefined. Other jurisdictions introduced a statutory definition. For example, in NSW, s 5(2) of the Motor Traffic Act 1909 (NSW) (inserted in 1937) adopted the definition of 'drug' contained in Part VI of the Police Offences (Amendment) Act 1908 (NSW): morphine, cocaine, ecognine, diamorphine (commonly known as heroin) and opium.
Most jurisdictions subsequently moved to an approach to defining 'drug' which focused on impairing effects. For example, in Queensland's Transport Operations (Road Use Management) Act 1995 (Qld), 'drug' means substances covered by the Drugs Misuse Act 1986 (Qld) and 'any other substance, article, preparation or mixture (with the exception of liquor) whether gaseous, liquid, solid, or in any other form which, when consumed or used by any person, deprives the person either temporarily or permanently of any of the person's normal mental or physical faculties' (Schedule 4 'Dictionary'). 6  As is demonstrated below, one of the noteworthy features of contemporary RDT laws is that these historical degrees of specificity-including the large number of drugs potentially covered and the focus on the impairing effects of the drugs and on driving capacity-have been lost. Their loss is an important part of the story of how drug driving laws (but not drink driving laws) have departed from the central issue of impairment and road safety.

The introduction of a 'proxy' for alcohol impairment: BAC/PCA
The first state to introduce drink driving offences based on the biological detection model (Quilter et al. 2016a: 936-937)-that is, where blood alcohol concentration (BAC) or prescribed concentration of alcohol (PCA) levels serve as statutory 'proxies' for impairment-was Victoria.
The Motor Car (Driving Offence) Act 1965 (Vic) added s 81A(1) to the Motor Car Act 1958 (Vic): '[a]ny person who drives a motor car while the percentage of alcohol in his blood expressed in grams per one hundred millilitres of blood is more than .05 per centum shall be guilty of an offence'. Other states quickly followed, though most initially preferred a 0.08 threshold. 7 Queensland adopted a 0.10 threshold (Traffic Acts 1949 to 1967 (Qld) s 16(1a)(a), as amended by the Traffic Acts Amendment Act of 1968 (Qld)).
These events are widely regarded as a milestone in Australian drink driving laws. They established, as the dominant approach, a model of using BAC/PCA levels as proxies for impairment. The legitimacy of these proxies was attributable to the significant scientific evidence that underpinned them. That the law was moving in line with this expert knowledge is evidenced by the fact that scientific evidence about alcohol effects was a notable feature of parliamentary debates (Griffith 1968

Pre-2000s approaches to testing for drugs other than alcohol
The introduction of random testing for other drugs lagged considerably behind RBT for alcohol. Prior to the introduction of routine oral fluid testing in the 2000s (discussed below), legislation originally provided for limited drug testing powers. Across Australia, a variety of regimes existed for gathering and testing blood or urine specimens for drivers suspected of drug driving, and for the admission of specimen analysis as evidence in driving offence cases. 9 In NSW the Motor Traffic (Road Safety) Amendment Act 1987 (NSW) amended the Motor Traffic Act 1909 (NSW) to allow for blood or urine testing following a failed or refused 'sobriety' assessment (s 5AA  (3)).

The move to 'presence' drug driving offences and random oral fluid testing for (selected) drugs
Between 2003 and 2015 all eight Australian jurisdictions put into place regimes for the criminalisation of drug driving that contained offences based on the 'presence' of any quantity of specified drugs, with random oral fluid testing powers for selected drugs (although, the Tasmanian government's approach was different in two respects, as discussed below). In most jurisdictions it is now a criminal offence to drive with any of three types of drugs in one's system: THC (cannabis), methylamphetamine (speed), and MDMA (ecstasy). 10 Once again, Victoria was the pioneering jurisdiction, when it enacted the Road Safety (Drug Driving) Act 2003 (Vic). Relevant amendments to the Road Safety Act 1986 (Vic) came into operation on 1 December 2004. Section 49(1)(bb) provides that a person who 'drives a motor vehicle or is in charge of a motor vehicle while the prescribed concentration of drugs or more than the prescribed concentration of drugs is present in his or her blood or oral fluid' is guilty of an offence. 11 In the case of a prescribed illicit drug, the 'prescribed concentration of drugs' is 'any [emphasis added] concentration of the drug present in the blood or oral fluid of that person' (s 3). This is why the Victorian model of drug driving laws, since adopted in all Australian jurisdictions, has been described as a 'zero-tolerance' approach (Kelly and Dillon 2005). Note, however, the language by which this was achieved. The legislation adopts the rhetorically powerful phrase 'prescribed concentration'-that is, the established proxy for impairment in the drink driving context, which has achieved such strong purchase in popular consciousness-while the reality is that the presence of any quantity of a prescribed drug meets the definition.
In Victoria the original definition of 'prescribed illicit drug' was limited to THC or methylamphetamine; however, MDMA was added on 1 September 2006, pursuant to the Road Safety (Drugs) Act 2006 (Vic), thus forming the prescribed 'trio' of substances that has become the cornerstone of Australian drug driving laws and RDT practices today. Random 'oral fluid' testing was permitted under s 55D, which provided that a police force member could require, inter alia, 'any person he or she finds driving a motor vehicle … to undergo a preliminary oral fluid test by a prescribed device …'. If a positive result ensued, the person could be detained for the purpose of providing an oral fluid sample for further analysis, 12 a positive result of which would be used as evidence in prosecution of the person. 13 Tasmania was the next jurisdiction to move to the presence/RDT model with the enactment of the Road Safety (Alcohol and Drugs) Amendment Act 2005 (Tas). In doing so, it made two notable adjustments to the Victorian model, making it unique amongst Australian jurisdictions. First, although oral fluid testing is used as an initial screening method (see s 7B), the ultimate test must be a blood test (s 7C) with the substantive drug driving offence requiring proof of the presence of a 'prescribed illicit drug' in the blood, rather than the more common oral fluid or blood (or oral fluid, blood or urine in NSW). Thus, s 6A of the Road Safety (Alcohol and Drugs) Act 1970 (Tas) states: 'a person who drives a motor vehicle while a prescribed illicit drug is present in his or her blood is guilty of an offence'.
The second distinctive feature of Tasmania's drug driving laws is that the drug presence offence is not limited to cannabis, speed and ecstasy. There are 18 'prescribed illicit drugs' including cocaine, heroin, GHB, ketamine, LSD, morphine, PCP and magic mushrooms (Road Safety (Alcohol and Drugs) Regulations 2009 (Tas) cl 16). We will return to both of these distinctive features below.
IJCJ&SD 52 Online version via www.crimejusticejournal.com © 2017 6(3) In the second half of the 2000s, all other Australian jurisdictions adopted drug driving laws based on the Victorian model, 14 where oral fluid is tested for the presence of selected illicit drugs. It is noteworthy that other jurisdictions adopted Victoria's model rather than the Tasmanian approach. Clearly, one of the 'advantages' of oral fluid testing is that, for drivers, the experience is analogous to the procedure for RBT, with little inconvenience and minimal invasion. However, given the issues with oral fluid testing (which we discuss below) the wide-spread adoption of the Victorian model may prove to have been decisive in producing a number of problems with Australian drug driving laws which we will discuss. Table 1 summarises the current presence drug driving offences in all Australian jurisdictions.

Features of contemporary drug driving laws and enforcement practices that are odds with a best practice impairment model
The previous part of this article demonstrated that, after a long period during which the criminalisation of alcohol-and drug-related driving turned on a common impairment paradigm, a significant separation has taken place. When BAC/PCA levels were introduced for alcohol, lawmakers did not move away from an impairment paradigm. Rather they drew on expert knowledge about the effects of alcohol to create legislative 'proxies' for impairment, now reflected in lowrange (0.05), mid-range (0.08), and high-range (0.15) drink-driving offences according to the concentration and concomitant road safety risk. 15 In the case of drugs other than alcohol, lawmakers have taken a different path during the last two decades (Hall and Homel 2007). In this part of the article we show that contemporary drug driving laws and practices display three characteristics that mean that they do not align strongly with the impairment + road safety risk equation that has traditionally underpinned drink/drug driving laws. First, they turn on presence of any detectable quantity (rather than a threshold concentration). Secondly, drug presence is evidenced via oral fluid testing (rather than blood testing). Thirdly, police routinely test for only three illicit substances (rather than all drugs with the capacity to impair driving) Testing for presence RDT techniques currently employed in Australia do not test for 'active' drugs in a person's 'system', but the presence of any quantity/residue of a drug in a person's oral fluids (discussed further below, 3.2.1). It follows that it is inaccurate to suggest that RDT 'devices will only detect the active ingredient of the drugs when they are active as an impairing influence. They will not detect drugs taken days or weeks earlier' (Lucas 2006: 729), as was stated by the Queensland Transport Minster when that state moved to introduce the 'Victorian model' drug driving legislation (Transport Legislation and Another Act Amendment Act 2007 (Qld); see also D Brown 2006 above). Similarly, it is inaccurate to suggest that oral fluid testing will only detect 'recent consumption' as the South Australian Transport Minister suggested in his second reading speech on the Road Traffic (Drug Driving) Amendment Bill 2005 (SA): '[i]t will detect recent consumption of methamphetamines and THC. Drivers who have THC or methamphetamine residues in their bodies as a result of use in the previous days or weeks will not be detected' (Conlon 2005: 3359).
Secondly, most roadside drug testing does not test for drugs generally, but only three illicit substances: THC, methylamphetamine and MDMA.

Oral fluid testing vs blood sample analysis
The RDT regime in Australia (with the exception of Tasmania) is based on 'oral fluid testing' both for the preliminary random test and the final oral fluid test. The preliminary test is generally conducted via a drug wipe stick (such as the Securetec DrugWipe Twin or the Securetec DrugWipe II Twin). 16 This involves the driver wiping his/her tongue along the testing stick. If this test is positive, the driver is taken to a roadside testing bus/van (or the police station) to provide a saliva sample, commonly tested by the Drager DrugTest 5000, or the Cozart Drug Detection System (DDS). If this test is positive a direction is normally given to the driver prohibiting him/her from driving (driver directions are discussed below), and this sample is then sent to a laboratory and tested by an analyst to confirm the presence of the drug(s While the Australian RDT regime relies on oral fluid sampling, the 'gold standard' for drug detection is said to be blood sample analysis (Wolff 2013: 57). This is because: 'oral fluid tests cannot be used to give a precise prediction of the concentration of a drug in blood (or plasma or serum) for confirmation testing and therefore prediction of possible drug effects' (Wolff 2013: 57).
In other words, oral fluid testing is a relatively poor mechanism for assessing whether a person is drug impaired. While oral fluid testing may be appropriate for preliminary testing, it is not wellsuited to confirmatory testing (Wolff 2013: 57, 129).
The ROSITA projects 18 showed that there was a weak correlation between the concentration of a substance in the subject's saliva and their blood. There may be personal physiological factors, varying person to person, that render the testing of saliva inadequate as a basis for estimating the level of drugs in a person's bloodstream and, therefore, unreliable as a basis for assessing impairment (see Wolff 2013: 129 There exists extensive scientific evidence which shows a direct correlation between the presence of certain drugs in the blood and the presence of those drugs in oral fluid. This correlation is particularly evident in the cases of the drug commonly referred to as delta-9-THC, and drugs containing methylamphetamine, such as those commonly referred to as speed and ecstasy. (Llewellyn 2005: 29) Another issue with oral fluid testing is that it is open to contamination, particularly in relation to THC. If cannabis is smoked, the active ingredient (THC) remains in the mouth as deposits: However, contamination of the buccal cavity is an issue for the detection of cannabis use since the drug is often used by oral, intra-nasal or smoking routes of administration (insufflations). 'Shallow depots' of cannabis may following recent use accumulate in the buccal cavity and produce elevated concentrations in oral fluid for several hours after ingestion. Unfortunately, the cannabinoids do not pass readily from blood into saliva and the detection of … (THC) in oral fluid is largely reported to be due to contamination of the oral cavity following smoking. (Wolff 2013: 65) A positive test for THC may result, therefore, from the detection of residual deposits of THC in the mouth. However, the drug may no longer be present in the driver's bloodstream and is, therefore, unlikely to have an adverse effect on driving ability (Wolff 2013: 65). The manufacturer of the device used by NSW Police, Dräger, has defended its equipment against such suggestions that oral fluid testing can yield 'false positives' 19 by asserting that 'oral fluid residue typically persists for two to four hours after smoking, which coincides with the time window for marijuana's major effects' (Dräger n.d.). This does not, however, accord with advice from the NSW Government's Centre for Road Safety which states that detection continues for 12 hours for cannabis (Transport for NSW-Centre for Road Safety 2016). Concerns about the accuracy of oral fluid testing for THC are particularly significant given that cannabis is the most widely used illicit drug in Australia (Australian Institute of Health and Welfare 2014), and one of the three drugs tested for in Australian RDT practices. Therefore, a positive test for THC is likely to account for a significant proportion of the large number of drug driving charges laid every year (see below).

The focus on three illicit drugs
With some exceptions (as discussed above) Australian drug driving laws are heavily focused on the detection and punishment of drivers who have used one or more of three drugs-cannabis, methamphetamine and ecstasy-the use of which is itself a criminal offence in all jurisdictions (see, generally, Bronitt and McSherry 2017: Ch. 14). This selective approach does not sit comfortably with a road safety/impairment paradigm. It appears to be the product of a range of other factors, including: the fact that these are the most commonly used illicit drugs in Australia (Australian Institute of Health and Welfare 2014); duplication of the approach pioneered in Victoria in 2003; the available technology and the ease with which the presence of these drugs can be detected via a roadside oral fluid test; and the ease with which they can be differentiated from substances which it is legal to use. In Western Australia, the selective focus was explained as follows: For practical reasons, it is necessary to exclude dexamphetamine, benzodiazepines and all drugs containing opiates. Despite the fact that in some cases these drugs do constitute a road safety concern, they are legally available for therapeutic use and for this reason roadside screening is impractical. At this point in time, roadside screening technology is not able to accurately differentiate between heroin, codeine, prescription pain medication and some over-the-counter medications. Therefore, for the purpose of random drug testing in Western Australia, it is proposed that, similar to other Australian jurisdictions, the regime will be limited to THC, tetrahydrocannabinol; methamphetamine, known as speed; and MDMA, methylenedioxy-methamphetamine, known as ecstasy. The oral fluid tests will not detect the presence of prescription drugs, common over-the-counter medication or ADHD-attention deficit hyperactivity disorder-medication. (Kobelke 2006: 7206) The NSW Government's justification for focusing on cannabis, speed and ecstasy was succinct: '[t]hese drugs are illegal, they are the most commonly used drugs in the community, and they all affect the skills and sound judgment required for safe driving' (M

'No drive' directions and times for testing
Further aspects of Australian drug driving laws that indicate a widespread failure to attend adequately to pharmacological knowledge are the rules governing the time during which a driver can be tested, and the rules governing directions that a person not drive a vehicle after testing positive to the presence of one or more prescribed illicit substances. Table 1 (column 6) shows that most jurisdictions expressly provide for 'no drive' directions after a positive test. 20 On their face, the rules on directions appear to appropriately align with a road safety/impairment rational: an impaired driver should not be permitted to continue to drive, until s/he is no longer impaired. However, with the possible exception of SA, 21 a closer analysis reveals that driver impairment is assessed by way of temporal or other proxies that do not actually assess driver capacity. For example, in Victoria, while the prohibition on driving is stated to be assessed by way of an officer's opinion on reasonable grounds that the driver is incapable of having 'proper control of a motor vehicle', the reality is more circular: a driver is effectively deemed incapable on the basis of testing positive to an illicit drug. 22 Similarly, in NSW, Queensland, the Northern Territory and the ACT, the legislation effectively deems a person to be incapable of driving for the relevant period of time of the direction (that is, 24 hours; 12 hours in the ACT). The implication is that, unless further drugs are consumed, the detected drug will no longer be present in the person's body after the expiration of the nominated period. Furthermore, legislated 'no drive' periods are not necessarily consistent with government and police advice regarding the period after consumption for which a person is at risk of testing positive (see Table 1, column 6). For example, NSW Government advice on stimulants ('speed, ice and pills') is that these can be detected for one to two days yet the direction on prohibiting driving is only for 24 hours (Transport for NSW-Centre for Road Safety 2016). Worse, the Queensland Government's advice is vague, suggesting that the detection period 'varies' (Department of Transport and Main Roads 2009); yet, the fixed 'no drive' prohibition is for 24 hours. In the ACT, where the prohibition is for 12 hours, Government advice suggests 'you could test positive hours or even days after consumption'. Note also the inconsistency with claims made during parliamentary debate on relevant legislation (discussed above). In short, standard 'no drive' times contained in Australian drug driving laws-ostensibly for preventive road safety purposes-do not account for the different pharmacodynamics and pharmacokinetics of different drugs and different drug users, 23 and so are an unreliable guide to impairment duration and return to capacity.
There is a further inconsistency between 'no drive' directions and the legally mandated times in which confirmative oral fluid sampling for these offences must be undertaken. Table 1 (column 5) shows that, across Australian jurisdictions, the time allowed for testing a driver varies from two hours to up to eight hours after the occurrence of the 'event' (typically, the last time of driving). If presence is the 'proxy' for impairment in the drug driving context, it is unclear what purpose these time limits serve, let alone why there is such cross-jurisdictional variation.
Overall, the striking absence of temporal consistency across Australian jurisdictions-in relation to the allowable period for testing, the duration of 'no drive' directions, and governmental advice on the period for which a drug user is at risk of testing positive-raises further doubt about the evidence-base for current criminal law regulatory arrangements. More specifically, such uncertainty and confusion must surely undermine the educative and communicative function IJCJ&SD 58 Online version via www.crimejusticejournal.com © 2017 6(3) (Farmer 2014) which drug driving laws are meant to serve. A driver (or prospective driver) who has consumed a prescribed illicit drug would rightfully be confused as to his/her legal responsibilities and the risks of criminalisation that driving might pose.
In summary, our analysis of parliamentary debates and the content of current drug driving legislation suggest that law-makers have been insufficiently attentive to linking drug driving offences and testing practices to the established impaired driving paradigm conceived around alcohol. There has been a failure to acknowledge that the adoption of drug presence as a proxy for impairment represents a radical expansion of the criminalisation of drug use and driving behaviours. However their merits are ultimately assessed, drug driving laws are not a simple extension of drink driving laws to other drugs. Current laws and practices adopt a distinctive (and selective) approach-centred on the presence of a small number of drugs-which is an uneasy amalgam of the road safety risks posed by drug impaired drivers, and the moral taint/criminal character of the use of illicit substances, mainly cannabis, ecstasy and methamphetamine.

Implications for the operational reach of drug driving laws: Penalties and enforcement
In the previous part, we raised a number of questions about the legitimacy of the prevailing emphasis in Australian drug driving laws on the detection (via oral fluid testing) of the 'presence' of three illicit drugs. To be clear, we are not suggesting that drug driving laws entirely miss their mark. The individuals caught up in, and punished according to, current laws and RDT practices in Australia do include drivers who are impaired by drugs other than alcohol and who therefore represent a road safety risk. However, at the same time, there are problems of both underinclusion and over-inclusion.
In terms of under-inclusion, users of both other illicit drugs (particularly those used commonly, such as cocaine) and lawfully prescribed drugs (such as diazepam (valium)) that can impair driving ability are largely avoiding detection because police in most states are only routinely testing for cannabis, ecstasy and methamphetamine (Shoebridge 2015). This is inconsistent with the stated road safety/crash and injury prevention rationale for drug driving laws.
In terms of over-inclusion, users of the more common illicit drugs-cannabis, ecstasy and speed-may be subjected to a form of over-criminalisation. By over-criminalisation in this context, we mean two interrelated issues. First, individuals may be punished as impaired/dangerous drivers because the presence of a relevant drug is detected in their oral fluid, in circumstances where the drug is no longer sufficiently active in their system to impact adversely on driving capacity. There may be relatively little sympathy for such individuals, but to punish them as impaired drivers when their 'sin' is that they have consumed an illicit drug at some point in the past lacks legitimacy. If their crime is illicit drug use, it is that crime that should be charged, proven and punished.
Secondly, as a result of the current form of drug driving laws, drives who test positive are exposed to hefty financial penalties and licence disqualifications and, in some cases, imprisonment terms. For example, in Victoria, the state with the highest penalties, a first offence of driving with the presence of a prescribed drug (Road Safety Act 1986 (Vic) s 49(1)(bb)) attracts a maximum penalty of 12 penalty units (currently $1,902.84) and minimum three months disqualification; a second offence, 60 penalty units ($9,514.20) and minimum six months disqualification; and a third offence, 120 penalty units ($19,280.40) and minimum six months disqualification.  s 47BA 1st Offence: Fine not less than $900 and not more than $1,300. 2nd Offence: Fine not less than $1,100 and not more than $1,600. 3rd Offence: Fine of not less than $1,500 and not more than $2,000. Subsequent Offence: Fine of not less than $1,500 and not more than $2,000.  (2AA)). If 2nd offence within 5 years: The first offence includes a number of alcohol/drug driving offences under s 79. The person is liable to maximum 20 PU or 6 months' imprisonment (s 79(2F)). If 3rd offence within 5 years: Again, the first two offences include a number of alcohol/drug driving offences under s 79. The person is liable to maximum 28 PU or 9 months' imprisonment.

Road Traffic Act 1974
1st Offence: Max 10 PU fine (s 64AC(2)) 2nd/Subsequent Offence(s): Fine not less than 10 PU or more than 20 PU (s 64AC(2)(b)).  (1)) 1st offender: Automatic disqualification for 3 years, able to be reduced to 6 months (s 34(1)) Repeat Offender, If Driver: Automatic disqualification for 5 years, able to be reduced to 12 months (s 34(2)) If Driver Trainer: none NT Traffic Act 1987 1st Offence: 5 PU or imprisonment for 3 months (s 28(1)) 2nd 'relevant' Offence: 7.5 PU or imprisonment for 6 months (s 28(1)) Subsequent 'relevant' Offence: 7.5 PU or imprisonment for 6 months 1st Offence: None 2nd 'relevant' Offence: Automatic cancellation of licence and 3 months' disqualification (s 28(4)(a)); Immediate suspension (s 28(5)) Subsequent 'relevant' Offence: Automatic cancellation of licence and minimum 6 months' disqualification (s 28(4)(b)); Immediate suspension (s 28 (5)) IJCJ&SD 60 Online version via www.crimejusticejournal.com © 2017 6(3) The concerns we have raised are amplified when it is recognised that recent years have seen a major escalation in the number of roadside tests conducted, and the number of charges laid. A full empirical analysis of the operation of drug driving laws is beyond the scope of this article, 24 but here we present preliminary data from Victoria and NSW to illustrate the scale of enforcement. Victoria conducted 42,160 ( Table 3).  This statistical information provides further reason to scrutinise whether Australian law-makers have received an appropriate alignment between the road safety risks posed by drugs other than alcohol and the form of contemporary drug driving offences and testing practices.

Conclusions and recommendations
In this article we have shown that, although drug driving has been criminalised in Australia since the 1930s, the preferred approach took a significant turn in the 2000s. Eschewing the traditional requirement of evidence of impaired driving, or a driver being 'under the influence' of a drug, lawmakers have empowered the police to conduct high volume roadside testing for the presence of the three most widely used illicit drugs-cannabis, methamphetamine and ecstasy-without any requirement to establish that there is a sufficient concentration of the drug in question to produce impairment effects. At the same time, there is no routine testing for other drugs (including licit prescription drugs) which can impair driving ability. This approach stands in stark contrast to the evidence-based approach to detecting and punishing individuals who drive with alcohol in IJCJ&SD 61 Online version via www.crimejusticejournal.com © 2017 6(3) their system, which largely turns on an evidence-based 'proxy' for impairment: BACs written into legislation.
It appears that the illicit nature of drugs such as cannabis, ecstasy and speed has been seen to justify a type of 'fast-forward' from presence to deemed impairment without any requirement to establish and meet prescribed concentration thresholds of the sort that are both conventional in the drink driving context and essential to the normative legitimacy of drink driving laws. To drive while 'drunk' is regarded as unacceptable because the scientific evidence shows that this carries an elevated risk of accident due to diminished driver capacity (Fell and Voas 2014;Howat, Sleet and Smith 1991;Mann et al. 2001). By contrast, contemporary drug driving laws in Australia are not strongly linked to scientific evidence about the relationship between substance use and driver capacity. The road safety justification for treating the presence of a particular drug in a driver's oral fluid as synonymous with impairment is open to question, just as it would be if the trace detection of a minute quantity of alcohol in any driver's breath were to be regarded as a sufficient basis for criminal punishment. In addition, testing for only three illicit drugs is not consistent with a road safety/impairment paradigm.
Just as for drink driving, criminal laws governing drug driving should be underpinned by what the expert literature tells us about the drug/safety/crash risk relationship. Reviewing that large body of literature is beyond the scope of this article, but it does exist (see, for example, Bondallaz et al. 2016;Elvik 2013;Kelly, Darke and Ross 2004;Verster, van de Loo and Roth 2017) and warrants closer attention. Policy-making and law reform can and should be informed by it, just as in the 1960s, when (alcohol) drink driving laws were modernised, and road safety qualitatively improved, in line with the available evidence about the impairment effects of alcohol.
It might be assumed that presence-focussed drug testing is the only available approach because the available scientific evidence about the pharmacology of drugs lacks the sophistication of evidence about alcohol's effects. Such an assumption would be incorrect. Several countries in Europe have taken, or are investigating, such an approach (Wolff 2013). For example, in 2012, Norway introduced evidence-based prescribed concentrations for 20 non-alcohol drugs, both illicit and licit (Kristoffersen et al. 2016;Vindenes et al. 2012). These include thresholds for THC (1.3 nanograms (ng)/ml in blood), MDMA (48 ng/ml), methamphetamine (48 ng/ml) and diazepam (57 ng/ml) (Norwegian Ministry of Transport and Communications 2014). A drugspecific threshold was also recommended by a major UK review of impaired driving (Wolff 2013).
We recommend that all Australian states and territories consider three changes to existing drug driving laws. First, serious consideration should be given to assessing the cost and feasibility of roadside drug testing that screens for all drugs which are known to have the potential to impair driving ability, whether they are currently illicit or licit substances. The Norwegian experience shows that such an approach is both technically possible and, so long as there is political will, feasible. Second, oral fluid testing should only be used as a preliminary screening device. Where this relatively non-invasive method detects the presence of a potentially impairing drug, the 'gold standard' of blood sample analysis should thereafter be used as the basis for decisions about criminal charges and convictions. To be clear, our recommended approach would not involve the mass administration of random blood tests (which would raise myriad risks to health and infringements of human rights). Only where an individual tested positive on an initial oral fluid test would a blood test be authorised. Third, informed by the available scientific evidence, minimum prescribed concentrations for all impairing drugs should be set and added to Australian drug driving laws as legislated 'proxies' for impairment, just as BAC/PCA levels are recognised as legitimate proxies for alcohol impairment.
The Wolff Report explained that the reason for setting blood concentration thresholds for different drugs is that, like PCA/BAC levels for alcohol, these thresholds would 'stand-in' or serve as proxies for levels of driver impairment: IJCJ&SD 62 Online version via www.crimejusticejournal.com © 2017 6(3) Setting a concentration or 'limit' for a psychoactive drug, for the new drug driving offence, means that if a driver exceeds this threshold the driver can be prosecuted without the requirement to prove that he or she was impaired and that this impairment was caused by the drug in the body. The implications of setting such a limit in law are therefore far-reaching, and the Panel members accept that their task in advising Government on such limits is crucial. Before recommending drug thresholds the Panel have therefore properly considered both the empirical (epidemiological) and experimental evidence, in relation to blood drug concentrations and driver behavior, whilst being mindful of stakeholders, practical and ethical considerations. (Wolff 2013: 19-20) Law reform in accordance with the three recommendations we have proposed here would bring Australian drug driving laws back into line with a road safety/impairment paradigm. It would address both the over-inclusion and under-inclusion that results from current laws and testing practices. It would remedy the current conflation of what should be regarded as two discrete public policy goals: promoting road safety; and deterring the use of certain drugs. The integrity of criminal laws and police powers concerned with drug-impaired driving will be enhanced if they are disentangled from the decision of governments to criminalise the use of certain drugs (like cannabis, ecstasy and 'ice').
We acknowledge that recommending the setting of drug thresholds to underpin driving offences might encounter resistance if it is perceived as condoning the use of substances that are currently prohibited. However, we argue that our recommended approach to drug driving laws is the most appropriate mechanism for advancing their primary goal of improving road safety. Drug driving laws should not be used as a de facto mechanism for punishing individuals who are suspected of having committed the crime of possession and/or self-administration of an illicit drug (see, for example, Drug Misuse and Trafficking Act 1985 (NSW) ss 10 and 12). Personal drug use remains a controversial site of criminalisation (Schwartz 2015), but there is little doubt that, outside the driving context, it would be regarded as unlawful for police to arbitrarily search an individual for illicit drugs, absent a 'reasonable suspicion' of possession and/or recent administration. On one reading of contemporary drug driving laws and the random testing for selected drugs that they facilitate, police have been empowered to test for illicit drug possession/use in a way that would otherwise be regarded as inconsistent with Australian society's respect for civil liberties and the presumption of innocence. Re-alignment with an evidence-based impairment paradigm would go a considerable way towards restoring the integrity of current laws.
This approach is consistent with the position advanced in the Wolff Report, which emphasised the need to distinguish the optimal impairment/road safety-focused approach to drug driving from the wider Government strategy of deterring (via criminalisation) the use of certain psychoactive substances: … the Panel has been solely concerned with the relationship between drug use while driving and this should not be confused with or taken as an extension to existing legislation about possession or supply of drugs or the Government's wider drugs strategy. Drug driving legislation is contained in the Road Traffic Act 1988 and has a separate policy aim from wider drug related legislation-namely it aims to improve road safety. The Government's drug strategy aims to reduce illicit and other harmful drug use and to increase the numbers recovering from their dependence. (Wolff 2013: 19) State and territorial governments in Australia would do well to heed this advice.
It might be considered naïve to argue for evidence-based law reform in a context where it is well known that much criminal justice policy formation and law making in Australia defies rational IJCJ&SD 63 Online version via www.crimejusticejournal.com © 2017 6(3) explanation, where 'common sense' paradigms dominate, motivated by the assumed electoral appeal of punitive 'law and order' approaches (Hogg and Brown 1998;McNamara and Quilter 2016). However, we believe it is incumbent on researchers to 'call out' unprincipled law making and encourage governments to be attentive to the normative deficits (and long-term legitimacy) of how criminal law is employed as a public policy tool (McNamara 2015).
A final consideration in our call for reconsideration of the nature of drug driving laws is the cost of current drug driving enforcement practices. As noted above, in recent years, most state and territory governments have significantly increased their investment in RDT (Andrews 2015 15 We acknowledge that, more recently, Australian governments have moved to introduce a 'zero tolerance' approach to alcohol in the case of inexperienced/high risk drivers. The standard prescribed concentration of alcohol for learner drivers and provisional licence holders is now zero and less than 0.05 for certain categories of drivers (such as drivers of taxis and other passenger vehicles): see, for example, Road Transport Act 2013 (NSW) s 108, which establishes a 'novice range' PCA threshold of zero, and a 'special range' PCA of 0.02. 16 These devices are approved in each state/territory under the relevant legislation and/or by gazette notice. 17 For instance, in Victoria, see s 57B (3) of the Road Safety Act 1986 (Vic). 18 The Rosita project was a collaborative international exercise 'to evaluate the usability and analytical reliability of the onsite oral fluid (saliva) drug testing devices' (Verstraete and Raes 2006). 19 Issues have also been raised about the number of 'false negatives' (that is, a problem of potential under-detection) (see, for example, Barone 2016). 20 Western Australia and Tasmania do not have such express powers, however, police practice in Tasmania appears to involve the giving of directions not to drive following a positive test for the presence of a prescribed illicit drug (Department of Police and Emergency Management 2010: 14.8.5(1)(k)). 21 Section 40K(7) of the Road Traffic Act 1961 (SA) appears to require a police officer to make an assessment of fitness before keys can be delivered back to a driver: 'A police officer on duty at the station must deliver possession of the keys to any person the officer is satisfied is lawfully entitled to them, and who makes a request for them at the police station, provided the officer has no reason to believe that person will drive the vehicle but not be qualified or unfit to do so '. 22 Similarly, in the ACT, impairment is deemed on the basis of the presence of a prescribed illicit drug: Road Transport (Alcohol and Drugs) Act 1977 (ACT) s 47B(2)-(3). 23 'Pharmacodynamics' refers to 'what the drug does to the body'; that is, the effect that the drug has on the body of a living person (Wolff 2013: 9). 'Pharmacokinetics' refers to what 'the body does to a drug' that is being consumed; that is, what happens to substances, in this case drugs, when they are consumed by a living person. This includes how the substance is absorbed, how it distributes in the body, how it breaks down or changes within the body, and how it is excreted (Wolff 2013: 8). 24 The operation of drug driving laws is the subject of ongoing research by the authors. 25 These 'positive' figures have been calculated by deducting from 100% the 'proportion of drivers tested who return clear result for prohibited drugs' in the respective years contained in the Victoria Police Annual Reports. We note that the information available does not distinguish whether these clear results were a product of the initial 'drug screening' tests alone, or a combination of the initial tests and secondary analyses.