This study was designed to assess driving-related behaviors and attitudes among a convenience sample of Australian MC users recruited as part of our larger CAMS-20 survey. One of the biggest differences between the present iteration of this survey and prior CAMS surveys (CAMS-16 (Lintzeris et al. 2018) and CAMS-18 (Lintzeris et al. 2020)) is the number of respondents using prescribed MC products. While this number was very small in CAMS-18 (n = 25) and even smaller in CAMS-16 (n = 1), it was much greater here (n = 376, including both dual and prescribed only categories), signifying an important transition underway in the Australian community from illicit towards legal prescribed MC products (MacPhail et al. 2022).

While there were relatively few differences in demographic characteristics across the different surveys, there were several notable differences in cannabis use characteristics. Perceived knowledge around the cannabinoid profile of the product being used was far greater in the present survey, even among illicit only users, with fewer participants reporting uncertainty around cannabinoid profile (30.9%) when compared with the total of 48% in CAMS-18 (Lintzeris et al. 2020). This presumably reflects the certainty of cannabinoid content that comes with legally prescribed products. The use of orally administered products was also higher here than previously, even among illicit only users (38.4 vs 26.6%). Among those using only prescribed products, the majority (88.5%) reported using orally administered products, with only a small number (11.5%) using inhaled products. These figures are consistent with general prescribing trends in Australia as reported by the Therapeutic Goods Administration (Therapeutic Goods Adminsitration 2022), although it is worth noting that prescriptions for inhaled products are steadily increasing (MacPhail et al. 2022).

A key safety concern with MC products is that driving shortly after the use of THC-dominant products may impair driving and therefore increase crash risk. In a recent Canadian survey, 23% of people who had used cannabis in the past 12 months reported driving within 2 h of smoking or vaporizing cannabis at some point in their life, while 14% reported driving within 4 h of ingesting cannabis at some point in their life (Government of Canada 2022). In our analysis, around half of the current cohort endorsed the idea that NMC use can impair driving and a large majority reported that the product they used had effects that lasted for 1–3 or 4–6 h. This interval of up to 6 h is the generally acknowledged window in which driving and cognitive impairment is most likely to be observed in occasional cannabis users. Notably, around 50% of all respondents in the survey typically drove within this 6-h window after use. Despite this, respondents generally thought that their MC use did not impair their ability to drive and less than 30% of illicit users and dual users admitted to driving while “high.”

Very few agreed that their MC caused them to take more risks, made it harder to drive in a straight line, or adhere to the speed limit. Approximately 10% of respondents did, however, indicate that they find it harder to remain focused while driving after using MC, and 14.7% reported being slower to react to sudden situations. Almost 50% of respondents said that they tend to leave a larger gap between their car and the car ahead and that they tend to drive more carefully with cannabis. This pattern of behavior is often observed in experimental studies and is thought to reflect an attempt to compensate for perceived impairment (Hartman and Huestis 2013). As with CAMS-18, most respondents felt they could accurately assess their driving ability after using MC. While the validity of this claim has yet to be empirically tested in patient populations, there is evidence to suggest that individuals are generally aware of their impairment after consuming cannabis (Hartman and Huestis 2013). At the same time, a recent study in healthy volunteers who were regular cannabis users showed a reduction in perceived driving impairment at 1.5 h relative to 0.5 h after smoking cannabis containing either 5.9 or 13.4% THC, even though there was no objective improvement in driving performance over this 1-h period (Marcotte et al. 2022).

Binary logistic regression analysis revealed several important predictors of DUIC, including legality of use, route of administration, cannabinoid profile, belief in whether NMC (but not MC) cannabis impairs driving, the deterrent effect of RDT, and frequency of cannabis use per day. Respondents who were using illicit products only were twice as likely to engage in DUIC relative to those using prescribed products only, and belief that NMC does not impair driving was associated with an almost 4-fold increase in the likelihood of DUIC. The use of orally administered products was associated with a close to 40% reduction in the likelihood of DUIC relative to inhaled only products, while use of CBD-dominant products was associated with a close to 60% reduction in DUIC relative to use of THC-dominant products. This latter finding is perhaps unsurprising given that CBD is non-intoxicating and does not impair driving performance (McCartney, et al. 2022; Arkell et al. 2020c) and noting that the question around DUIC specifically asked respondents if they had driven while “under the influence (i.e., while high)” in the past 12 months. Unlike CAMS-18, where unemployment was associated with a 4.7-fold increase in the likelihood of DUIC, employment status was not a significant predictor of DUIC in CAMS-20. This previous finding may therefore have been an artefact of the small number of respondents who were unemployed (n = 26), as we hypothesized at the time (Arkell et al. 2020b).

Consistent with the CAMS-18 survey (Arkell et al. 2020b), we observed a significant deterrent effect of RDT with most respondents saying that RDT deterred them from driving after consuming MC. This deterrent effect was similar across users of illicit only products (81.8%), prescribed only products (81%), and dual users (82.1%). Respondents who were not deterred by the presence of RDT were almost twice as likely to engage in DUIC. The legality of the product used may therefore have little influence on willingness to drive with THC in one’s system which likely reflects the fact that current laws do not discriminate between users of prescribed and illicit products. It is worth noting that the population sampled here may have excluded patients who are deterred from using MC altogether by current driving laws.

One key question with tangible implications for policy is whether patients who use MC as prescribed are less impaired than individuals who use cannabis for other reasons. Recent reviews provide little evidence of impairment in those using stable doses of cannabis products to alleviate medical conditions that themselves may cause impairment (MacCallum et al. 2022). This adds to evidence that regular cannabis users are less susceptible to THC-induced impairment than occasional users with an equivalent dose of THC (McCartney et al. 2021b; Bosker et al. 2012; Ramaekers et al. 2009). At the same time, MC patients may perceive less risk around DUIC than other users (Wickens, et al. 2019) leading to a higher likelihood of DUIC (Wickens et al. 2022). It remains to be seen whether the relatively low perception of risk associated with DUIC seen here is an accurate assessment of risk.

Our findings indicate that a significant minority (39%) of people using MC reported driving within 3 h of consuming cannabis products—the time period most likely to be associated with intoxication and driving impairment based on studies of NMC users with limited tolerance to cannabis. The extent to which impairment occurs in people who use MC regularly and develop tolerance to the effects of cannabis remains unclear. Many classes of medications (e.g., opioids, benzodiazepines, antipsychotics, antihistamines, antidepressants) are known to cause impairment, and indeed studies highlight many such medications impair driving to a similar or greater extent than cannabis (Arkell et al. 2021). Most countries have driving policies recognising that such medications can cause drowsiness and impairment, and so patients are cautioned against driving if they are experiencing impairment. Clinically, health care professionals caution patients about driving if impaired, particularly so when commencing treatment and stabilizing on their dose, or when there are significant dose increases. It has been argued that a similar approach should be applied to patients prescribed MC in place of the blanket prohibition in Australia which prevents MC users from driving altogether. Further research examining driving impairment in patients prescribed MC long term would provide critical evidence in this debate and is sorely needed.

While evidence for sex differences in acute cannabis effects is conflicting [e.g., (Sholler et al. 2021; Arkell et al. 2022; Matheson et al. 2020)], an analysis of data from the 2016–2017 National Survey on Drug Use and Health in the US indicated a significantly greater probability of DUIC among males than females for both combined MC/NMC users and NMC-only users (Lloyd et al. 2020). Overall, the probability of DUIC ranged from 20 to 25% for females and from 28 to 40% for males, suggesting a possible need for more targeted interventions (Lloyd et al. 2020). A recent study by Wickens et al. (Wickens et al. 2022) likewise observed a greater incidence of DUIC among males relative to females (20.7 vs. 8%). No such difference was observed in the present analysis, although it is important to note that our survey only included MC users while these other two studies included both MC and NMC users. Further work might seek to elucidate whether sex differences do exist in DUIC likelihood among broader groups of patients.

This study is not without limitations. Self-report data are inevitably prone to inaccuracies and response bias. The use of convenience sampling may have also introduced a selection bias toward respondents who favor the relaxation of stringent cannabis and driving laws. Nonetheless, we were able to capture a wide range of responses from individuals using a variety of products, cannabinoid formulations, and routes of administration. As noted elsewhere (Lintzeris et al. 2022), demographic data in the prescribed group here are similar to the broader demographic data from the Australian regulator around patients receiving MC prescriptions, suggesting that this cohort is a good representation of the community at large.