Introduction

The Issue of Drug Driving

The deterrent effect of roadside drug testing (RDT) has long been met with resistance, largely due to the low-risk perceptions and attitudes that drug users adopt towards the consequences of drug driving (Hasan et al. 2022). A significant portion of these perspectives may be derived from the perceived limitations of current drug testing policy, technology, and methodology. Within the context Australian roadside drug testing programmes, a zero-tolerance policy has been imposed towards drug driving, in that laws prohibit the presence, to any degree, of illicit drugs in the system (Davey et al. 2014; Moxham-Hall and Hughes 2020). This is in contrast to other countries which have opted to accommodate the changing landscape of legalisation surrounding legal cannabis and prescription medical cannabis products, by allowing certain threshold for the presence of delta-9 tetrahydrocannabinol (THC), or by assessing driver impairment through field sobriety tests. Secondly, research has demonstrated that frequent drug users generally perceive that (a) their driving ability is not impaired following use (Barrie et al. 2011; Davey et al. 2005; Hasan et al. 2022; Wilson and Wilson 2010), and (b) in comparison, legal substances, such as strong prescription medications (i.e. benzodiazepines and opioids) and alcohol, are comparatively treated with leniency, despite perceptually having more debilitating effects towards driving (Boehnke et al. 2016; Wickens et al. 2019). As identified in defiance theory, one of the driving forces behind offending behaviour is that of whether an individual perceives current sanctions to be just. Given the discrepancies between jurisdictions and the inconsistent sanctioning of various substances, it is critical that the perceptions that drug users have towards enforcement strategies are identified, to determine whether the intended effects of RDT are being carried forward.

In Queensland specifically, RDT involves testing for the presence of drugs using the Securetec Drugwipe II as an initial saliva screening test, and then the Drager DrugTest 5000 as a secondary roadside confirmatory oral fluid test. Accordingly, these devices screen for the presence of three drugs: THC, methylenedioxymethamphetamine (MDMA), and methamphetamines (Queensland Government 2021). Following a positive secondary roadside screening, a sample is then sent to the government laboratory for LCMS confirmation from which it is decided to charge (positive presence result) or discontinue action (negative presence result). A number of Queensland studies have examined the characteristics of all positive drug tests in Queensland (Davey et al. 2014; Mills et al. 2021b). It was found that methamphetamines were the most common drug detected (39.4; 40.8%, respectively), followed by cannabis (34.0%; 29.8%), and MDMA (0.8%; 1.70%). Poly-substance detection was also common, particularly between cannabis and methamphetamines (up to 22.5% of cases), indicating that there may be a degree of overlap between user types, and thus attitudes and perceptions. However, police data is limited in that it is representative of only those drug drivers who have been caught, and not those who have successfully avoided apprehension.

Accounting for this limitation, a number of studies have attempted to gain a better understanding of drug driving prevalence from the users’ perspective, by using a self-report methodology. Recent statistics have indicated that approximately 28% of Queensland drivers have driven within 4 h of taking an illicit substance in the prior 12 months (Mills et al. 2021b). Among drug users, cannabis (98.7%) was found as the most common drug driven on, followed by MDMA (39.5%), and methamphetamines (11.7%) (Mills et al. 2022b). This discrepancy in the drug driving rates between police apprehension data and self-reported data suggests that particular drug users (i.e. methamphetamine users) may be commonly targeted based on their characteristics, which was indicated in a recent study by Anderson et al. (2021). As such, this may mean different drug user types have adopted differing attitudes towards police testing, and the perceived likelihood of being apprehended.

Police Testing and Deterrence

As explained by classical deterrence theory (Homel 1988), three perceptual factors are thought to influence the deterrence of an offending behaviour, including the perceived: certainty of apprehension (COA), severity of the punishment, and timeliness of the punishment. Deterrence theory has received support from numerous studies investigating the impact of legal sanctions on the intention to drive under the influence, although the perceived COA has emerged as the strongest deterrent factor to both drink driving (Freeman et al. 2021a, b; Freeman, et al. 2021b; Freeman and Watson 2006; Szogi et al. 2017) and drug driving (Freeman et al. 2010; Hasan et al. 2022; Watling et al. 2010). Accordingly, two forms of deterrence can potentially result from these perceptual factors: (a) general deterrence, which refers to deterrence within the general population, and is reliant on how effectively legal sanctions are publicized and promoted (Taxman and Piquero 1998); and (b) specific deterrence, which refers to deterrence created by the apprehension and punishment of offenders (Davey and Freeman 2011; Homel 1988).

Incorporating the ability of offenders to avoid detection, Stafford and Warr (1993) later expanded on specific deterrence by accounting for four core experiential components related to punishment avoidance, and thus the intent to drug drive: (a) the direct experience of punishment, (b) the vicarious experience of punishment (social learning), (c) the personal experience of avoiding punishment, and (d) the vicarious experience of punishment avoidance. Of these four factors, experiencing punishment avoidance has demonstrated to have the largest negative impact towards COA and positive impact on the intent to drug drive (Armstrong et al. 2018, 2005; Watling et al. 2010). The influence of punishment, however, has been more ambiguous, as frequent offenders have shown to be perceptually resistant to legal sanctions (Freeman and Watson 2006). This is generally because punishment avoidance is far more likely to occur, particularly if avoidance behaviours are adopted, and thus the effects that punishment has towards COA may be negated (Homel 2017).

In contrast to the extensive support on specific deterrence, limited research has been afforded to understanding how a deterrent effect may be established in the general population (i.e. general deterrence). What is known, however, is that human learning is thought to be dependent on the repeated exposure and pairing of stimuli (del Carmen Sanjuan and Nelson 2019; Meeter et al. 2009). However, as explained in classical conditioning theory, extinction or the deterioration of acquired knowledge occurs if the conditioned stimulus (e.g. driving) occurs repeatedly without presence of the unconditioned stimulus (e.g. police testing) (Bouton and Moody 2004). In regard to road safety, Freeman et al. (2021b) applied such concepts to drink driving deterrence and found that (a) the number of times seeing and being tested by police positively correlated with the perceived COA, and (b) participants surveyed immediately following police testing had a higher COA. Despite this novel finding, there is no current evidence to demonstrate whether this relationship extends to RDT and drug driving COA, even considering the significantly lower number of tests being administered. According to the Department of Transport and Main Roads, the population of Queensland drivers exceeded 3.7 million in 2019, while less than 70,000 drug tests were administered (QPS 2020), translating to less than 2% of the total driver population being exposed to police testing for the year.

Knowledge and Attitudes Towards RDT

Given the limited perceived safety risks associated with drug driving, drug users’ attitudes towards police RDT have been generally negative. Specifically, it was believed that the testing methodology behind RDT was inaccurate, and that similar to drink driving, a threshold should be developed for detecting drugs in the system (Wilson and Wilson 2010). This was particularly because drink driving has been perceived by drug users as more detrimental to driving over most illicit drugs (Barrie et al. 2011; Wilson and Wilson 2010). Nonetheless, a consensus among users is that the chance of being caught for drug driving is low (Barrie et al. 2011; Davey et al. 2005), which is re-enforced by previous punishment avoidance experiences (Davey et al. 2005), and vicariously, through the observations of peers’ experiences with drug driving (Barrie et al. 2011; Hasan et al. 2022). Despite such concerns, a Melbourne-based study revealed that less than half of night club attendees indicated that being caught for drug driving would affect their drug driving habits (Degenhardt et al. 2006). In consideration of the low perceived physical and legal consequences of drug driving, drug users have reported that the benefits of drug driving, opposed to the limitations of alternative methods of transport (e.g. the costs of catching a taxi), were justified (Barrie et al. 2011).

In a subsequent quantitative study (Armstrong et al. 2018), only half (50.68%; n = 406) of the sample reported that RDT operations would be effective in detecting drivers who had recently used illicit drugs, suggesting that there was a need for improvement in capturing the entirety of the drug driving population. Despite having strong attitudes towards police testing, previous research has indicated that the knowledge surrounding the policy and legislation of RDT was limited. This suggests that perceptual factors involved in deterrence, such as COA and severity of punishment, may be having a minimal effect because many drug users are not actually aware of the laws and legislations that underlie them. As discussed and highlighted in past research (Hasan et al. 2022; Watling and Freeman 2011), low perceptions of legitimacy towards sanctions have been linked to intentions to offend.

The Present Study

In review of the current literature, and in consideration of the divergences between legislations and enforcement strategies between jurisdictions, there is a need for research to examine the current attitudes and perspectives that drug users have adopted towards police testing. Such findings would help inform whether certain strategies are potentially imposing a negative affect towards drug users’ intention to offend (i.e. drug drive). Therefore, the present study aims to conduct an exploratory qualitative investigation into drug users’ experiences, behaviours, and perceptions surrounding RDT practices in Queensland.

Method

Participants

In total, 52 Queensland drug users holding a current driver’s licence were interviewed. Participants were aged between 21 and 71 years old (M = 44.85, SD = 11.92), were prominently male (71.2%), and were currently using at least one type of illicit drug. More than half of participants (61.5%) reported that they were currently working, while a quarter of the sample (25.0%) were unemployed, 7.7% were retired, and 5.8% were on disability. Participants reported having consumed a variety of drug types. Specifically, black market cannabis (84.6%) was the most common, followed by medicinal cannabis (23.1%), methamphetamines (19.2%), MDMA (11.5%), cocaine (9.6%), heroin (3.8%), lysergic acid diethylamide (LSD; 1.9%), and ketamine (1.9%). The concurrent use of prescription medications (i.e. benzodiazepines; opioids) was also prevalently used within a quarter of the sample (25.0%). A range of offending driving behaviours were reported within in the sample. Specifically, some users reported that they would strictly regulate their driving following use, some would only drug drive rarely in specific situations (e.g. camping), while others would offend daily or not regulate their use.

Materials

The interview questions used in this study focused on a number of topics including demographical information (i.e. age, gender, occupation), drug use patterns (i.e. drug type, history, frequency, and context of use), experiences with RDT, avoidance behaviours around RDT, knowledge about RDT, and perceptions about RDT. The interview content was collaboratively developed by the authors (who specialise in the field of road safety, drug driving, and deterrence), and was informed by previous research literature, specifically the qualitative approach conducted by Davey et al. (2005), whom is an author of the current study and has more than 30 years’ experience in the road safety domain. In addition, the questions were developed in response to previous findings on drug users’ perceptions (Mills et al. 2022a) and behaviours (Mills et al. 2021b) adopted surrounding RDT. While there were a number of set questions to be asked, the delivery of the interview was in an unstructured manner, to allow for a more natural and deeper exploration of the topics. The primary questions used to initiate discussion on the topics can be found in Supplementary Table 1.

Procedure

Following ethical approval by the University of the Sunshine Coast Human Research Ethics Committee (#A211629), an advertisement was placed on Facebook in which detailed the aims of the study and provided a link to an online survey. The survey included a research project information sheet, detailing the requirements of participation (i.e. being over the age of 18, an illicit drug user, and owning a Queensland driver’s licence) and the characteristics of the study. After consenting to participation, participants were asked to record their contact details, so that a phone interview could be arranged. The interview took between 15 and 30 min to complete, and participants were reimbursed with a $50 electronic gift card for their time. After completion of the interviews, a professional transcription service was used to convert the audio into text. Following transcription, thematic analysis was then employed by two separate researchers using the text analytical software NVivo (version 12). The identified themes were then compared, combined, and interpreted within the context of existing literature and theory. Specifically, the six stages of thematic analysis, as identified by Braun and Clarke (2006), were used. Thematic analysis was applied to allow flexibility in the approach to exploring the perceptual and behavioural themes that emerged in the responses.

Results and Discussion

The thematic analysis identified a number of core themes across three general dimensions: (a) perception about police testing, which included the experiences and perceptions about RDT effectiveness; (b) punishment avoidance behaviours, which identified the potential avoidance and masking behaviours adopted by users to evade punishment; and (c) perceptions about police limitations, which incorporated the perceived limitations of current legislations and future suggestions that participants had to improve enforcement strategies. These themes and the raw data factors that predetermine them are displayed in Fig. 1. For the purpose of interpreting the themes within the context of the findings, the results and discussion sections were combined. The relevant quotes from participants will be presented when discussing these themes in subsequent sections.

Fig. 1
figure 1

The proposed topics of the study and associated themes within the data

Full size image

Perceptions About Police Testing Effectiveness

Experiences with Offending and Policing

First, to gain a deeper understanding of their past experiences and perceptions regarding enforcement strategies, participants were asked about their past experience with RBT and RDT. Correspondent with the recorded quantity of RBT and RDT tests being administered in Queensland (QPS 2020), seeing and being tested was far more common for RBT, than for RDT experiences. This was similar for being caught for offending, in that comparable to findings by Davey et al. (2005) and Barrie et al. (2011), both direct and vicarious experiences of apprehension for drink driving were much more common than for drug driving (R5). In fact, of the participants who reported seeing RDT, many reported that they would have likely tested positive should they have been tested. This discrepancy may mean that drug users may have a diminished COA, particularly for rural areas, where testing was reportedly even more sparse. The experience of punishment avoidance has shown to be positively associated with future intentions to drug drive (Hasan et al. 2022; Watling et al. 2010). Given the high level of punishment avoidance present in the sample, the general consensus was that apprehension likelihood was low, as per findings by Davey et al. (2005). However, despite this finding, a number of participants expressed concerns over being caught (R6). It was also mentioned that the introduction of RDT has led to a decrease in offending behaviours (R6) among a number of participants, indicating that the introduction of testing has been at least somewhat effective in deterring drug driving in some user types.

R5 (Male, 46): I’m 46 and I’ve been stopped twice and tested. It’s a needle in a haystack at the moment.

R6 (Male, 60): It’s in the back of my mind all the time, these days. As I said, like even 10 years ago, this wasn’t such an issue, but now it is, and it plays in the back of your mind. It inhibits me being a bit more social on the weekend because of that fact.

Participants were also asked, based on their experiences with offending and policing, whether they thought drink or drug driving was more prevalent. The results were mixed, although drug driving was often perceived as more common than drink driving because it was believed to be safer and the chances of being caught were much lower, given their lack of exposure to RDT (response 1; R1). Such comparative perceptions have also been identified in prior research (Malhotra et al. 2017; Wickens et al. 2019), and re-enforce the potential increased risk for drug driving propensity. Some participants even mentioned that if RDT testing was as widespread as RBT testing numbers, cannabis would be found to be far more prevalent in the statistics over alcohol (R2). As highlighted, the current delivery of RDT compared to the driver population is underwhelming, and the proportion of self-reported cannabis drug drivers compared to RDT apprehension statistics is suggesting that there are far more cannabis offenders than police testing is currently recognising. It was also noted that if the presence of prescription drugs (i.e. benzos; opioids), prescribed or black market, was accounted for, drug driving would be considered far more common (R3). The potential prevalence of drug driving on prescription drugs has been previously illustrated by Starkey et al. (2017), in that prescription drugs were more common than illicit drugs on the road.

R1 (Male, 28): I’d say a lot more people drive high rather than driving drunk. I think driving drunk is a lot more dangerous a lot, you’re intoxicated. I think a lot more people make the decision not to do it.

R2 (Male, 29): Obviously, drink driving, but if people actually were testing positive and stuff like that, it’d definitely be drug driving.

R3 (Male, 46): There’re a lot more people that are taking stuff like Oxycontin and stuff like that and driving and I think they wouldn’t test for it, but it’s still drugs in the system.

Conversely, of those who reported drink driving as more prevalent, it was stated that this was due to alcohol being more readily available and socially acceptable (R4). This is in accordance with deterrence theories, in that offending behaviour is more likely deterred if the individual perceives that the risk of social (e.g. peer disapproval) and internal (e.g. shame) consequences is present (Homel 1988). Finally, it was also highlighted that that drink driving might be more common because there was a policed limit, and that people might easily go over that limit (R4). In support of this notion, it was recently highlighted that possible drink driving (25.5%) was more commonly reported than acknowledged drink driving (16.0%) instances (Freeman et al. 2021a, b).

R4 (Male, 39): Well, I think drink driving is still a very, very big problem. There’s plenty of people out there that would be 0.05, 0.06 quite regularly driving home thinking that they’re perfectly fine to drive… I reckon that would happen quite regularly more than drug-driving because there’s a bottle around every second corner. Alcohol is promoted in sporting events, TV events. It’s everywhere.

Perceptions About Drug Testing Accuracy

A number of concerning reports surrounding the accuracy of police drug testing were also highlighted. Specifically, it was mentioned on multiple occasions that drug drivers and their social groups had often avoided detection from roadside oral fluid screening, despite their reportage of being drug tested just following consumption of drugs (e.g. cannabis; R7). This finding is somewhat consistent with previous evaluations of the oral testing devices, which identified small proportions (7–14.5%) of Drager DrugTest 5000 results to give false negative readings, compared to blood sample tests (Gjerde et al. 2018; Wille et al. 2010). These experiences may also be a result of higher THC concentration cut-off values being used in roadside screening, as compared to the secondary laboratory confirmation test, which is usually 5 ng/mL (personal communications with laboratory). These findings suggest that drug driver punishment avoidance is occurring not only due to a lack of RDT numbers being deployed, but also from the experience of avoiding detection once exposed to police testing, which may have significant consequences on the intentions to reoffend.

R7 (Female, 34): I was actually driven through it with someone, they’ve been pulled over. They’ve only just like smoked cones and passed the test, then they got pulled back through again not even an hour later and still got let go and passed the test.

As opposed to those who claimed to have experiences of punishment avoidance with police testing procedures (either directly or indirectly), some users reported that they had tested positive the next day, and even days after taking cannabis (R8). Subsequently, many concerns were raised over testing positive the next day, even after reporting that the impairing effects had passed (R9). This concern was more evident in participants who regulated their driving after cannabis use, as opposed to those who did not (and openly admitted to drug driving frequently). Concerningly, others reported that due to the perceived extended detection window of cannabis, some users might turn to ‘harder’ drugs to drive on, such as methamphetamine, because they believed the detection times were much shorter among those drugs (R10), despite this not being the case.

Again, while further research is required, prior investigations have shown that oral fluid testing devices can potentially a result in a false positive, in comparison to a negative blood tests (Dobri et al. 2019; Gjerde et al. 2018), which may be particularly among heavy cannabis users (Huestis et al. 2013). However as discussed, roadside testing THC detection levels are set much higher than the laboratory lowest level of detection (generally 5 ng/mL). Consequently, reports of having tested positive on the roadside after not smoking cannabis for numbers of days could be somewhat questionable. Regardless of the experiences with testing, these findings highlight a dilemma, in that while the perception of increased detection times may positively influence COA and deter some drug users from drug driving, it may also be (a) influencing a negative attitude towards police testing and (b) encouraging drug driving behaviours on potentially more debilitating and less policed drugs. In support, previous research has identified that attitudes about the legitimacy of sanctioning have a positive impact towards drug driving intent (Watling and Freeman 2011).

R8 (Male, 42): I hadn’t had cannabis that day. I was completely sober. Just as I was coming into Tweed, there was a roadside test. I got pulled over. I thought I was going to be fine, but they found a trace on a swab test.

R9 (Male, 24): To be honest, all the time. I understand the risks, especially when I’m driving stoned. It’s more so when I’m driving and I’m still positive from the night before, but I feel I’m in no way impaired.

R10 (Male, 49): Because the weed shows up in your system for a longer period afterwards, they all get on the hard stuff. I’ll get on the ecstasy and crack pipe and ice and all this. I’ve been part of the ice epidemic because that shit goes out of your system pretty quickly.

As a result of participants having mixed experiences with police testing following drug use, the perceptions of the detection window for testing positive after use were also mixed. In fact, participants reported a wide variety of perceived detection durations, ranging from hours to weeks. Participants attributed a number of factors as being responsible for influencing the length of detection times, including the body fat content, the amount used, the frequency of use, the level of physical activity, the drug type, the use of masking techniques, the degree of food consumption since drug use, and body mass. Currently, the evidence for factors affecting the detection window for drugs is very limited. Therefore, it would be beneficial to understand (a) what factors are associated with longer detection times, and (b) how these perceptual differences are influencing drug users’ COA and intentions to drug drive. A theme in past drug driving research (Hasan et al. 2022) is that apprehension certainty is mixed among drug users and not always predictive of intent to offend, particularly when accounting for other situational factors (e.g. punishment avoidance). As such, there may be underlying beliefs about the length of detection windows that may moderate the effects between COA, and the intent to offend.

Factors Associated with Apprehension Certainty

Following the questioning on perceived likelihood of apprehension, participants were probed about what circumstantial factors might contribute to the increased or decreased likelihood of detection. Four primary themes emerged. Firstly, a theme relating to the ability to avoid detection was identified and included confidence in ability to reduce visibility (R11), beliefs about the length of the drug detection window (R12), and the perceived ability to drive unimpaired (R13). While the experience of punishment avoidance and COA have been shown to predict drug driving behaviour (Hasan et al. 2022), there is currently little research that has investigated whether perceptions about such specific avoidance behaviours are associated with COA for drug driving.

R11 (Female, 38): The motto where I grew up was always, “If you’re doing one illegal thing, don’t do two,” so if you’re driving, don’t speed, use indicators, be aware of what’s going on. Don’t draw attention to yourself. My car is very worthy, I’ve got a clean driving history, and speeding tickets. I think I’d be more alert. I don’t draw attention to myself.

R12 (Male, 35): The vast, vast majority of times I would not, but I remember reading a while ago that marijuana can go up to seven days afterwards, so if I had smoked weed overnight time, and then gone to work the next day if I had seen a police (RDT) stop, then I’d be a little worried, yes.

R13 (Male, 42): No, I’m comfortable as with it. Like I said, I’ve even had police up my butt. It crossed my mind, I’m like, “If I get pulled over, I’d probably get in trouble,” but my confidence in myself while I’m operating a vehicle is absolutely 100%.

Secondly, situational and environmental factors related to the perceived level of policing were discussed by participants. These included the location (i.e. rural vs urban; R14), is the time of day and day of the week (R15), and the number of RDT seen (R16) in certain areas. This theme was perhaps more commonly mentioned among those who did not have extensive experience with police testing, and thus displayed confidence in their ability to predict and avoid RDT. As explained by Homel (2017) and Bates and Anderson (2021), through experiences with police testing, offenders begin to adopt avoidance behaviours to circumvent potential future testing sites.

R14 (Male, 71): I’m a little bit paranoid about it (being caught) but not overly because of the area I live in. If I was living in a city, I probably wouldn’t touch it at all because I think my chances of being caught would be much higher.

R15 (Male, 50): Well, I’m definitely not going to be in the middle of a bar on a Friday night after having some.

R16 (Male, 35): No, I think it’s (being caught) not very likely. I’ve never been drug tested by the police, I’ve never seen a drug test, but I’ve seen RBTs. I’d say not very much.

The third theme related to how targetable users believed they were. Specifically, it was noted among some more frequent users that due to the targeted nature of drug tests, they may be discriminated against because of their past offending (R17), their appearance (R18), or because of the vehicle that they drove (R19). Conversely, less frequent users claimed the opposite, in that they were more likely to get away with drug driving, because they drove responsibly and did not look like a drug driver (R20). One participant noted an acquaintance that took advantage of the targeting bias, in that he would only drink drive because his past offending was drug related and thus was only drug tested when stopped by police (R17). Despite the improved detection rates associated with using targeted regimes for RDT (Anderson et al. 2021), it would appear that this approach may also yield unhelpful effects towards particular drug users’ COA and punishment avoidance, and thus future intentions to offend for drug driving (Hasan et al. 2022).

R17 (Female, 49): I did meet someone the other day. They won’t smoke now until they get home, but they drink and drive. They said, “Yes, but now that it’s been done, that’s all they’re testing for now.” They don’t breath test him anymore. They’re just drug testing because that’s what he’s been done for before.

R18 (Male, 24): I’ve got friends that just from their appearance, they get drug tested every time they’re pulled over… I look quite young in the face, so I usually get by.

R19 (Male, 28): It depends on which one of my cars I’m driving. If I’m driving the Prado, unless I go through a testing facility, it’ll be next to nothing. If I’m driving my Silvia on the other hand, it’ll be probably more than likely every day that I’ll get pulled over.

R20 (Female, 57): We’re almost all 60 years of age and if we are at a barbecue and somebody swings a joint our way, we’re not going to say no, we’re going to smoke it and hope to god we don’t get pulled over… Most of the time they would leave us alone because we’re obviously in control of our faculties. It doesn’t look like we’re on drugs.

A final theme found to influence the fear of apprehension was that of the participants’ attitudes towards police testing. Users who believed that being caught would have a negative impact on their life would typically have a greater fear of being apprehended (R21). Alternatively, some users (typically frequent offenders) reported that the impact of legal enforcement was not a concern for them (R22). Furthermore, the concern for physically harming the self or others was rarely brought up among these users. This was likened because of the beliefs that they were not impaired or held a generally deviant nature. As discussed by prior research, frequent drug users are generally under the perspective that they are not impaired when driving under the influence of drugs (Davey et al. 2005; Hasan et al. 2022), and that anti-social, deviant attitudes are predictive of future offending behaviour (Jornet-Gibert et al. 2013).

R21 (Male, 60): It’s not only the fear for me and my job. It’s the fear for my employer that I’ve got that I’m going to lose my license, but I’m going to lose my job. That’s the frightening thing about it.

R22 (Male, 46): I might even decide to have one while I’m driving. It doesn’t even bother me.

Punishment Avoidance Behaviours

Avoidance Behaviours

Next, participants were questioned about the behaviours they might adopt to avoid detection. Outside of those who reported that they do not adjust their driving behaviours because of drug use (R23), two types of avoidance behaviours and one regulatory behaviour were identified. Firstly, some users reported that they would avoid using direct routes on main roads and use secondary routes, such as residential roads and back streets (R24). This was particularly if users were familiar with the area and had identified favoured RDT locations (R25). Such behaviours have recently been identified among younger offenders and were shown to decrease COA (Bates and Anderson 2019). It was also noted among some users that they would utilise social media groups or applications that shared locations of RBT and RDT, to assist in their avoidance behaviours (R26). As identified by Mills et al. (2022b), the use of social media groups to avoid police detection may result in a lower apprehension certainty, and thus higher intentions to offend. Alternatively, other users would simply regulate their driving by inhibiting drug use, particularly if they were going to be driving in an unfamiliar area or highly populated area, where the presence of enforcement activities may be more likely (R27). Furthermore, and as discussed prior, Homel (2017) explains that as offenders become accustomed to law enforcement techniques, they adapt their driving practices to either avoid committing the offence or avoid the likelihood of detection, behaviours that should be accounted for when considering deterrence-based research.

R23 (Male, 59): I have never changed my cannabis use because of driving.

R24 (Male, 35): I live in a small town. There’s not a lot of roads options available here and there but sometimes I might take the residential road, I might do a little bit of a loop here and there.

R25 (Male, 24): (I have seen RDT) Three times, literally in the same spot… They were in the exact same spot, same setup, and everything.

R26 (Male, 40): Probably about 12 months ago when I was still smoking daily. I had ways on my phone to try and look out for police sites and I would actively think, “Oh, there more likely chance on this road or that road”.

R27 (Male, 70): If I was aware of that (police locations), I would certainly practice some avoidance… If I have a trip planned to go to a major city, that’s say an hour and a half from me, then I would probably not have any for a couple of days beforehand.

Masking Behaviours — Targeting

Participants also reported that they had particular masking behaviours that would help them avoid suspicion or detection, even in the presence of police. Where some users reported using specific vehicle types (R19), others stated that they would use visual cues, such as hanging a nurse’s uniform on their window (R28), to appear less characteristic of a typical user. Strictly trying to follow road rules and driving safely were also practices that were mentioned by participants (R29). Finally, the masking of the users’ physical characteristics were also commonly mentioned, in that users reported using products such as Clear eyes (to reduce the appearance of red and glazed eyes), or deodorant and breath freshener (to mask the smell of cannabis; R30). Such behavioural practices and masking strategies have not been identified in previous literature, which may be an important avenue for future quantitative research to explore, particularly in relation to deterrence theories.

R28 (Female, 38): I’ve normally got a nursing uniform in the car. Sometimes if it’s a risky time to drive, I will just hang a uniform on my hanger on the back seat.

R29 (Female, 60): I obey all the road rules and speed limits and all the other things that you risk getting pulled over for.

R30 (Male, 37): I’ll put clear eyes in my eyes to get rid of the redness. That makes my eyes completely white and not so bright, and I also just put on a bit of deodorant whenever I go out… and I’ll also brush my teeth as well before I go anyway.

Masking Behaviours — Saliva Testing

Participants also reported using a number of techniques to mask the presence of drugs in saliva, should they ever be drug tested. These substances included vinegar, apple cider vinegar, lemon juice, Fisherman’s Friend, Quick-Eze, methylated spirits, greasy and fatty foods, hydrogen peroxide, bicarb soda, toothpaste, chewing gum, dry mouth oral rinse (Biotène), mouthwash, olives, cranberry juice, menthol mints, milk drinks, soft drinks, and tea. Although some users were adamant in the ability of the substances to negate or hide the presence of drugs in saliva (R31; R32), many participants claimed that despite having no confidence in the behaviour, they would still use masking substances, to reduce the associated anxiety of being caught (R33). Of particular concern, one participant even claimed receiving medical advice on how to mask medical prescribed THC in saliva (R34). No such research has yet investigated whether specific substances can change the result of a saliva drug test, and thus an experimentally controlled study to assess such substances may be informative to future drug driving detection strategies.

R31 (Male, 58): Well, salt and vinegar chips work. The vinegar in the chips has a chemical reaction, it neutralizes (the THC in saliva).

R32 (Male, 28): It’s a dry mouth paste. It basically just adds a film on your tongue, so you don’t fail the tests. Obviously the laws in Queensland are pretty backwards with it.

R33 (Female, 40): I have Quick-Eze in my car. I don’t know if it works or not, but it helps with the anxiety of thinking, “Oh no! I had a smoke last night. I’m coming up to a roadside test. I’m just going to chew a Quick-Eze”, but even then I feel like an idiot doing it and I don’t know if it works or not.

R34 (Male, 36): By the same doctor who sold me that Arizer Solo II (a vaporiser device), he told me, “If you want to pass an oral drug test, you need to have a good dental routine before you leave the house,”.

In sum of drug user behaviours surrounding RDT, there appears to be a number of behaviours that are adopted by drug users to avoid police encounters, suspicion, and detection. While the avoidance of identified police locations and the masking of personal characteristics known to be targeted by police are likely to be effective in reducing apprehension likelihood, there is more ambiguity among drug users surrounding the use of masking substances. To any degree, it is likely that having positive perceptions about that these behaviours, regardless of their effectiveness, will decrease COA and encourage drug driving behaviours.

Perceptions About Police Testing Limitations

Based on their experiences with police testing and the judicial system, participants were next asked about their perceptions towards current drug driving legislations. While a portion of participants agreed that police testing might be successful in creating a deterrent effect (R35), most (cannabis) users did not think the current police methods were fair or accurate (R36). Therefore, several themes emerged regarding potential improvements.

R35 (Female, 49): I think a lot of people that I know that are smokers wait until they get at home before they’ll have to smoke nowadays because of the new testing.

R36 (Male, 49): I’m going to say it’s flawed. It works. But It’s flawed.

Increase Current Laws

Firstly, a number of participants suggested that laws and sanctions needed to be increased. In particular, it was identified by some participants that it was rare to see drug testing operations (R37) and therefore they believed there was a need to increase the frequency of RDT to create a deterrent effect within the drug user population. As demonstrated in previous research (Freeman et al. 2021b; Mills et al. 2021a, b), exposure to police testing of any kind contributes towards apprehension certainty and thus, deterrence. Alternatively, some users suggested that the current enforcement techniques were discriminatory to only a subset of users (cannabis, methamphetamine, and MDMA), and therefore, there needed to be testing of more drug types (e.g. psychedelics, heroin, cocaine, strong prescription drugs; R38). This was particularly when participants perceived non-policed drugs (e.g. benzodiazepines) to be more debilitating than regulated drugs, such as medical cannabis (R39; R40). Defiance theory (Sherman 1993) explains that people will be more inclined to offend or defy laws when they perceive the governing laws and associated sanctions to be unjust. Therefore, these findings may suggest that the perceived discrimination that some users are experiencing may increase the likelihood of offending behaviour. Currently, the option to test for a wider variety of drugs in oral fluid is available, however considerable costs are and operational issues associated with such a change. This is similar to the suggestion of increasing the quantity of testing, and further highlights a need for more cost-effective approaches to drug testing on the road.

R37 (Female, 47): The limitation is how many people they can test… I just haven’t heard of anybody getting tested or haven’t seen them set up as drug testing. Very randomly do I even see drug tests set up.

R38 (Male, 47): Once they get the laws right, they should test for impairment, and they broaden it, so it captures a whole range of other drugs that are implicated in road crashes and fatalities and make it as wide as possible.

R39 (Male, 45): I think it’s discriminating against me because if it wasn’t for me having marijuana, I’d be on Oxycodone and all that sort of stuff, and I’m allowed to drive on them, which affects me because I get drowsy and I do swerve over the road when I’m on those pain killers… When I’m on the pot, it doesn’t affect me at all.

R40 (Female, 40): I just really want to drive home the fact that I’m allowed to drive after I take Lyrica, and I cannot even speak a sentence after I take those tablets.

Improve Current Policing Methods

A second theme to emerge among participants’ responses was a critique of current testing methods, in that the current tests were perceived as not accurate enough to detect whether drug use was recent (R41), and therefore, users believed they were being falsely charged for driving when they were no longer impaired (R42). In particular, these perceptions were often expressed from participants who believed or reported experiences in which they tested positive for cannabis on days where they had not used. While this perception runs somewhat contrary to the knowledge that oral fluid testing is a good medium for identifying recent use (National Drug Driving Working Group 2018), there is also currently limited evidence for the sustained presence of THC in oral fluid, as recent investigations of oral testing devices have been somewhat questionable in regard to the potential for false positives (Dobri et al. 2019; Gjerde et al. 2018; Huestis et al. 2013).

Of particular concern was that some participants that noted that because they perceived they were going to test positive (regardless of having recently used or not), they and others in their group of friends would continue to drive under the influence because the consequences would be the same (R43). In combination, these findings suggest that (a) there is confusion around the performance of roadside testing, which may be having adverse effects on offending and attitudes towards authorities; (b) that driver education about the testing processes, equipment performance, and sample analyses may help reduce these negative appraisals; and (c) further research on oral fluid testing is needed, particularly on the factors that may increase the risk of potential false positives.

R41 (Male, 40) If there was some sort of test that could measure if I had smoked cannabis in the last six to 12 h, but not further than that. I would absolutely support that through and through because I’ve adjusted my life to make that allowance.

R42 (Male, 39): I think the laws are extremely unfair and discriminative, but they need to be there in a way, but they need to be refined and changed and modified to only punish the people that are actually under the influence.

R43 (Male, 24): We feel that we may as well just smoke and drive anyway because you’re going to get caught wherever you go. Because we smoke, at least for me, to deal with the pain and stuff, and some sleep issues.

Following suggestions for more accurate testing methods to determine actual impairment, participants also suggested that developing and identifying a threshold may be beneficial (R44). Currently, there have been an increasing number of studies using simulator-based approaches to determine a potentially acceptable cut off level for the presence of THC in the system (Peng et al. 2020). In response to the amounting evidence, there has also been a number of countries introducing THC concentration limits in oral fluid (Ginsburg 2019; Wong et al. 2014). However, with reference to both concentration and impairment, there is great variability among the cognitive impairment and motor functioning of individuals under the influence of cannabis (McCartney et al. 2021), which suggests that a one size fits all model to enforcement is highly problematic (Beirness and Porath 2019).

R44 (Male, 43): They’re (RDT) effective as in, they can say, “Oh, yes, now we’ve definitely picked up some marijuana,”… But the effects have long worn off or they are to a degree where it doesn’t physically affect you anymore… There’s a big difference between someone who’s just used drugs, and then gets out in traffic conditions versus someone who uses it, then a whole day goes past and jump in the car… I think they need to find a threshold. There needs to be more research done, the same as alcohol.

Adoption of Behavioural Testing

A final factor that emerged among participants’ responses to improve perceptions of current testing was that of adopting field sobriety testing approach of suspected drug drivers (R45), or a more detailed approach to drug driving on medical cannabis specifically (R46). Given the inconsistent relationship between THC presence and impairment (Chow et al. 2019), a number of countries have adopted the use of roadside field sobriety tests to judge impairment among suspected cannabis drivers. While a valid and efficient behavioural test of impairment may address some possible issues associated with the interviewees’ perceptions of current police methods, there are some major limitations to this approach (particularly within Australian jurisdictions), as recently evaluated by Spindle et al. (2021). Firstly, unalike sobriety tests for alcohol impairment, functional tests of cannabis impairment have not yet been empirically developed and validated, and the immediate effects of cannabis are difficult to distinguish from factors such as advancing age or fatigue (Ginsburg 2019). Drug drivers who have been charged could very well argue that the assessment and officers’ evaluation of their impairment were inaccurate (as per current concerns with oral fluid testing). Thirdly, for enforcement and legislating jurisdictions within Australia, there would be considerable demands on time and costs associated with the training, implementation, and maintenance of such behavioural testing across policing environments. Finally, under Australia’s mass high volume testing and general deterrence approach to impaired driving, field sobriety testing does not have strong history within Australia’s policing jurisdictions, legislation, and practices (National Drug Driving Working Group 2018).

R45 (Male, 28): The old-school sobriety test… Make them perform some manoeuvres on their hands and feet. That tells you straight away if somebody’s impaired. You can’t tell if somebody’s impaired just from a drug test, because as you know with THC, you could have had it the day before.

R46 (Female, 51): I think it should be a case-by-case basis, where it is a test to prove a level of intoxication. Definitely, a broader scope because it doesn’t make sense to me that you can’t smoke marijuana, but you can take five Valium and drive and it’s not tested.

Concluding Remarks

This study examined the perceptions and behaviours that drug users adopted surrounding police drug testing. Firstly, it was found that drug users were rarely exposed to RDT and thus commonly experienced a natural punishment avoidance. A number of situational (e.g. location), perceptual (e.g. perceived ability to avoid detection), and personal (e.g. whether users believed they were targeted) factors were thought to influence participants’ perceptions of apprehension certainty. Of note, the targeted nature of RDT appeared to increase drug users’ perceptions about the predictability of testing procedures, and thus decreased apprehension certainty among most users, as they knew had become familiar with the policing procedures. Policy makers may wish to consider the benefits of implementing more random drug testing (in addition to targeted), as this testing approach may deter the broader population of drug users and be perceived as less predictable among persistent offenders.

Three types of punishment avoidance behaviours were adopted by participants and were respective of the stages of police interaction: avoiding exposure (e.g. circumnavigating potential RDT locations), avoiding targeting (e.g. masking drug user characteristics), and avoiding detection (e.g. masking the presence of drugs in oral fluid). The use of such behaviours was typically related to inflated confidence in the ability to avoid detection and thus more frequent offending. Future research would therefore benefit from investigating whether certain substances (e.g. vinegar) are effective in masking the presence of drugs, and whether confidence in these behaviours increases COA. Future research may also consider the development of a comprehensive measure of apprehension certainty involving specific punishment avoidance strategies to help ascertain different dimensions of risk perceptions. Such methods may reveal why risk perceptions are not always associated with intentions to offend. Other more frequent users (e.g. methamphetamine and cannabis users) reported that they did not regulate their driving or use around RDT. These participants held the belief that they were not impaired, and at times demonstrated typical anti-social attitudes. Future research may wish to explore how perceptions of impairment may interact with risk perceptions (regarding harm) and attitudes towards drug driving intent.

Concerns over the accuracy of police testing were also raised among participants, in that both false positives and negatives were commonly reported. Where experiences of punishment avoidance led to lower perceptions of apprehension certainty, perceptions of false positives, or testing positive following significant delays in driving after use, led to feelings of discrimination and negative attitudes towards police testing. It was commonly perceived and suggested that testing methodologies, particularly for cannabis, need to be improved to accurately determine impairment (e.g. through use of a THC thresholds or implementation of behavioural tests of impairment). This perceived discrimination and inconsistent testing methods were shown to not only potentially increase consumption of non-policed drugs, but also to increase offending, because participants believed the punishment would be identical if they were no longer impaired. Currently, there is emerging evidence for the effectiveness of such methodologies, and thus future research may benefit from the working towards an accurate threshold to determine recent use; or in developing and testing an effective (and efficient) subjective test. As discussed, the use of behavioural assessments (although not without their own limitations) would also allow the policing of drugs not currently enforced upon.

While this study identified potentially important concerns for RDT and offending, limitations were present. Specifically, a convenience sample was used and recruited via Facebook, and thus the sample may have been subject to particular bias (i.e. those users who use Facebook and are willing to share sensitive information anonymously). Nonetheless, the thematic analysis was able to capture a wide range of users with differing and polarised perceptions and behaviours regarding RDT, meaning that it is likely a significant portion of users were captured. Regardless, further research may wish to confirm such findings using both broader and more specific samples, to give a cultural context to drug user perceptions of different police enforcement strategies. Now that the relevant themes have been identified in drug user inclinations towards RDT using qualitative methodology, the findings have highlighted a number of theoretical and practical implications that future quantitative studies may wish to endeavour. In light of the current limitations, implications, and future suggestions, it is evident that there is currently a level of dissonance between the objectives of police testing and the outcome on drugs user perceptions, particular among medical cannabis users. While limitations do exist, it is vital that change is initiated with caution and within the framework of accompanying empirical research. Given the relatively new legislations surrounding cannabis legalisation and regulation in parts of the world, the current long-term and real-world evidence is yet to emerge.