The present study investigated the acute effects of orally ingested THC oil on medicinal cannabis patients’: (a) hazard perception skill performance; (b) driving-related risk-taking behaviours (i.e., speeding propensity, following distance, and gap acceptance); (c) self-ratings of their own hazard perception skill performance (including the association between subjective self-ratings and objective performance); and (d) self-ratings of their own on-road driving skills and safety. Although participants’ actual hazard perception skill performance did not significantly decline from baseline to post-consumption, their perceived performance did. Moreover, there was no significant correlation between self-rated and objective hazard perception skill performance at either baseline or post-consumption. In the video-based measures of driving-related risk-taking behaviours, participants selected significantly slower speeds and longer following distances post-consumption of THC, although there was no significant change in their gap acceptance behaviour. There was also no significant change in self-perceptions of current on-road driving skills and safety after correction for multiple tests. Overall, while the study yielded no evidence that oral ingestion of THC oils by medicinal cannabis patients impacted hazard perception skill performance, the results suggest that these consumers are unable to accurately self-assess their hazard perception skill performance, regardless of whether they have consumed THC. This lack of insight into one’s own driving skill corresponds to findings for other driver groups (Groeger and Grande 1996; Horswill et al. 2013). Further, our findings suggest that medicinal cannabis patients engage in compensatory behaviours (specifically, by reducing their speed and increasing their following distance). Such findings provide important insights into the impact of orally ingested THC on driving behaviour and has the potential to inform future drug driving policy and evidence-based advice for medicinal cannabis consumers.

As noted above, the present study yielded no significant difference in hazard perception skill performance between baseline and post-consumption of THC oil, which may suggest that hazard perception ability is unaffected by the acute administration of orally ingested THC amongst medicinal cannabis patients. To date, only two studies have examined the effects of orally ingested THC on driving performance (Bosker et al. 2012; Manning et al. 2024). Bosker et al. (2012) found 20 mg of oral synthetic THC (dronabinol) increased on-road SDLP (standard deviation of lateral position) by 4.2 cm relative to placebo, with peak effects occurring at 90-minutes post-ingestion. However, this finding is based on a small sample of 12 occasional recreational cannabis users. Using a naturalistic design similar to the present study, Manning et al. (2024) examined the effects of prescribed medicinal cannabis oils on simulated driving performance in Australian medicinal cannabis patients. Consistent with our findings, they found no notable evidence of driving performance decrements relative to baseline.

Our finding of no significant difference in hazard perception skill performance is interesting, given that a substantial portion of variance in hazard perception performance is explained by key cognitive functions (i.e., processing speed, attention, psychomotor ability, and executive function) that have previously been shown to be affected by THC (Horswill et al. 2008; McInerney and Suhr 2016). One potential explanation is that it may be attributable to a high tolerance level amongst medicinal cannabis patients. Specifically, the present sample had an average cannabis usage history of 15 years and consumed their THC oil at least once per day, and therefore likely had relatively high tolerance levels. Another potential explanation is that experienced medicinal cannabis patients might put more effort into hazard perception to compensate for a perceived impairment when they know that they are under the influence of THC. If participants perceived that the test felt more effortful in the post-consumption condition, then this may explain why they rated their performance less favourably despite the lack of a measurable decline. In either case, our results align with a growing body of evidence demonstrating that chronic THC use can mitigate the acute effects of the substance on driving performance (Bosker et al. 2012; Brooks-Russell et al. 2021) and driving-related cognitive functions (Desrosiers et al. 2015; Ramaekers et al. 2011; Schwope et al. 2012). This highlights the need to also examine the effects of orally ingested THC on hazard perception among occasional or new users, such as medicinal cannabis patients in the early stages of their treatment, who may be yet to develop tolerance and/or compensatory strategies.

Dosage and symptom relief might have also influenced the present findings. First, the present study took a naturalistic approach towards dosing, with each participant consuming a single dose of their prescribed THC oil product. This led to an average of 10.80 mg THC (0.12 mg/kg) being consumed by the sample, which is approximately half of their self-reported typical single session dose (M = 19.57 mg, SD = 24.56) and far less than the total dose that participants reported typically consuming over course of a day via oil and/or flower products (M = 179.81 mg, SD = 178.13). Furthermore, this observed dose is within Australia’s therapeutic dosage range (5-20 mg; Arnold et al. 2020) and less than fixed flower doses typically investigated within the literature (e.g., Arkell et al. 2019; Ramaekers et al. 2006; Spindle et al. 2020). Experimental research administering similar THC doses have found either modest or no significant performance deficits on measures of cognitive functions that correlate with hazard perception skill (Gray et al. 2008; Schlienz et al. 2020; Spindle et al. 2021). These prior findings suggest that the average THC dose observed in this study may not have been sufficient to produce deficits in hazard perception performance following ingestion. Notably, a small subset of participants self-reported extremely high typical THC doses, with daily THC consumption (via both oil and flower products) estimates reaching up to 625 mg THC (albeit this could not be verified). Further research is needed to understand the impacts of regular THC ingestion among medicinal patients, who require their medication on a frequent and ongoing basis.

Second, participants in the present sample were also being treated for a range of conditions (including chronic pain, mental health disorders, and insomnia) that can independently have negative effects on cognitive function and driving capacity (Bulmash et al. 2006; Iezzi et al. 2004; Perrier et al. 2014; Sheline et al. 2006; Veldhuijzen et al. 2006). While research in clinical populations is limited, one study found that 20 mg of inhaled THC resulted in no deficits (or even modest improvements) in performance on measures of processing speed, attention, and executive functioning amongst a similarly mixed patient sample (Olla et al. 2021). It is therefore possible that our clinical sample experienced the alleviation of symptoms following their medication use, which in turn mitigated the extent to which THC affected performance. Nonetheless, further research that compares medicinal cannabis patients’ objective performance to general population norms is needed to elucidate the role of symptom relief in clinical populations that use cannabis for medicinal purposes.

An alternative explanation for the present findings is that the hazard perception test utilised in present study is not sufficiently sensitive to orally ingested THC. However, prior research has found video-based tests of hazard perception ability to be sensitive to other impairing substances (i.e., alcohol, West et al. 1993), as well as conditions such as acute mild traumatic brain injury (Preece et al. 2010), age-related decline in cognitive ability (Horswill et al. 2009), fatigue (Smith et al. 2009), distraction (Horswill and McKenna 1999), driver experience (Hill et al. 2019), and crash involvement (Horswill et al. 2015). For example, West et al. (1993) found that the consumption of moderate doses of alcohol between 0.04 and 0.06% delayed hazard perception response latencies by 1.1 s (West et al. 1993). That is, hazard perception tests have been found to be sensitive to a range of crash-related conditions. The present study is the first to examine the effects of acute oral THC consumption on hazard perception test performance, and hence we cannot rule out the possibility that higher doses of THC might significantly impair performance. This possibility could be examined in future research investigating the effects of a range of higher THC doses on hazard perception test scores. Furthermore, research comparing performance on video-based hazard perception tests with other cognitive and driving performance measures that have proven sensitive to acute THC administration (e.g., SDLP), is required.

It is also important to consider that driving is a complex task requiring a dynamic interplay of motor, visual, and cognitive functions (Anstey et al. 2005), and a hazard perception test cannot incorporate all the demands inherent to real-world driving. Nevertheless, scores from hazard perception tests developed using the same methodology as the test employed in the present study provide an effective means of measuring a driver’s “situation awareness on the road” (Horswill and McKenna 2004), and are supported by evidence of experience-related performance differences, associations with crash rates (both retrospectively and prospectively), and measures of real-world driving performance such as heavy-braking frequency (Hill et al. 2019; Horswill et al. 2015; Wetton et al. 2011).

We found speeding propensity to decrease following acute THC ingestion. This is in accord with prior findings that recreational cannabis users adjust longitudinal control behaviours as a way of compensating for potential deficits in driving capacity (Brands et al. 2019; Brooks-Russell et al. 2021; Hartman et al. 2015; Lenné et al. 2010). In a recent experimental study, Brooks-Russell et al. (2021) observed a decrease in driving speed in a simulator amongst daily, but not occasional, recreational cannabis users following inhalation of cannabis flower containing THC. As our sample primarily consumed their medication at least once a day, our results align with this finding and support the notion that cannabis usage patterns can influence engagement in certain compensatory behaviours.

Similarly, participants’ minimum comfortable following distance increased following acute oral THC consumption. Whilst this result is consistent with previous findings of increased following distance after THC administration in occasional recreational cannabis users (Hartman et al. 2015; Lenné et al. 2010), it contrasts with recent survey research in which more than half of a sample of medicinal cannabis patients denied leaving a larger gap between them and the vehicle in front of them (Arkell et al. 2020a). However, Horswill et al. (2020) note that one of the reasons for developing the test used in the present study was because text-based self-report questions have been found to be ineffective for assessing following distance behaviour, likely because it is difficult to make such judgements reliably without the visual context that the traffic clips provide. Further, in the present study, participants made these contextualised judgements both with and without having consumed THC, so that the difference in their responses could be compared directly.

Nevertheless, despite an increased crash risk for individuals with poor gap acceptance behaviour (McDowell et al. 1983; Tupper et al. 2011), there was no significant change in gap acceptance wait time. Taken together, our findings suggest that, although medicinal cannabis users may engage in compensatory strategies when they know that they are potentially under the influence of THC oil, they may be more inclined to compensate for cannabis impairment through alterations in speed and following distance (which are closely related driving behaviours), rather than gap acceptance.

Finally, data from the self-report items indicated that participants had little or no insight into the effects of THC on their driving skills and safety. Not only did participants as a group perceive a decline in their hazard perception skill performance that was not borne out by the objective data, but there was also no significant relationship between their self-rated and objective performance at either timepoint (baseline or the post-consumption). In addition, self-ratings of on-road driving skills and safety did not significantly change from baseline to post-consumption, after correcting for multiple tests. Without objective on-road driving data, it is difficult to determine the true accuracy of such ratings. Nevertheless, they are consistent with a lack of substantial change in objective hazard perception skill performance. However, it is worth noting that self-ratings for predicting traffic conflicts and driving safety were significantly lower in the post-consumption condition prior to correction for multiple comparisons. Interestingly, if the uncorrected finding for predicting traffic conflicts is regarded as potentially meaningful (e.g., as suggesting that the inclusion of this measure in future studies with larger samples or fewer measures may be warranted), it is consistent with participants’ perceptions of a decline in their hazard perception test performance. Prior survey research suggests that medicinal cannabis patients are confident in their ability to accurately recognise their own level of impairment (Arkell et al. 2020a), yet recent preliminary findings indicate that cannabis users may in fact have a limited capacity to recognise their own driving impairment while acutely affected by THC (Arkell et al. 2019; Marcotte et al. 2022). Building on this research, the present study is the first to reveal a poor correspondence between objective and perceived driving performance through a direct correlational analysis, thus providing further evidence that medicinal cannabis patients’ appraisals of driving safety are inaccurate, which may increase the risk of DUIC. Despite this, our sample exhibited a tendency toward risk-aversion, likely as a means of compensating for perceived effects on performance.

Importantly, the present sample’s inability to accurately appraise their hazard perception skill performance at either time point implicates generally poor self-monitoring skills rather than an acute disruption of judgement and awareness due to THC consumption. This is consistent with prior findings that the general population tends to have inflated self-ratings of driving ability that poorly correspond with objective measures (Freund et al. 2005; Horswill et al. 2004, 2013). Most of the sample also rated their hazard perception skill performance, and on-road driving skills and safety, as better than average in both conditions, which further reflects the self-enhancement bias that is common among all driver groups. Nonetheless, these findings do not exclude the possibility that judgement and awareness may also be affected at higher THC doses and in less tolerant individuals. Given that perceptions of safety have been found to influence driving under the influence of cannabis (Borodovsky et al. 2020; Jones et al. 2007; Malhotra et al. 2017; McDonald et al. 2021), the impact of prior beliefs and attitudes on objective performance and self-ratings could be investigated in future research.

This study has several important limitations that must be noted. First, although all participants reported complying with the requirement to abstain from cannabis products for 11.5 h, almost half tested positive to THC in oral fluid prior to each testing session. This indicates that some amount of residual THC was present in their systems, which may have had an effect on their neurocognitive state. However, it is unknown whether this use was recent, given that THC is a highly lipophilic compound that can remain above oral fluid detection thresholds for up to 3 days in heavy cannabis users (Niedbala et al. 2001; Odell et al. 2015). Furthermore, the absence of blood measures prevented us from quantifying THC concentrations and its relevant metabolites under each condition, as well as from examining their association with task performance. Second, there was a substantial amount of variability in the THC dose that each participant consumed. These products also differed in cannabis strain, which can influence THC-induced subjective effects and potentially task performance (Sholler et al. 2022). Although varying amounts of cannabidiol were also present in each product, this compound appears to have negligible effects on neurocognition and driving-related skills (Arkell et al. 2019; Englund et al. 2023; McCartney et al. 2022b). Third, post-consumption performance on the video-based assessments was only measured at one time point, making it difficult to determine if the peak effects of the THC were appropriately captured. Approximately twenty minutes of neuropsychological testing was also conducted prior to the commencement of these tasks. However, this may be less problematic for orally ingested THC, considering that its effects have been shown to last for up to 4 h (Schlienz et al. 2020; Spindle et al. 2021).

Finally, it is possible that performance on the video-based measures was affected by changes in clinical symptomatology (e.g., pain flare-ups), as various clinical conditions were present within the sample, and baseline and post-consumption evaluations occurred on separate days. Certainly, future research is needed to address these issues by utilising controlled THC/CBD doses and additional assessment time-points for objective and self-rated performance. Furthermore, it is recommended that future medicinal cannabis research measures clinical symptoms to control for their influence on performance outcomes, and possibly explores the role of symptom relief in mitigating the psychoactive effects of THC. Despite these limitations, the present study offers important preliminary insights into the effects of orally ingested THC oil on drivers’ hazard perception skill performance, risk-taking behaviours, and self-perceptions of driving skills and safety within medicinal cannabis populations.