NSW’s Medicinal Cannabis Driving Reform Is Scientifically Indefensible

Yesterday, in an announcement made without the Government identifying any meaningful consultation with scientists or other competent persons, the NSW Government proposed a radical change: the effective legalisation of driving while impaired, provided that a person takes a few minutes out of their day to obtain a prescription over the phone from one of the dozens of cannabis dispensaries operating in Australia and registers as a “medicinal cannabis” patient with Transport for NSW.

The primary issue is the excessively high threshold.

Before June 2026, the effective laboratory confirmation threshold for THC in NSW was 10 ng/mL of oral fluid. That was already higher than the confirmatory cutoff of 5 ng/mL specified in AS/NZS 4760:2019, Procedure for specimen collection and the detection and quantification of drugs in oral fluid. However, I would argue that 10 ng/mL was a defensible and comparatively fair threshold.

Ramaekers et al.¹ found that an initial significant shift towards impairment occurred at plasma THC concentrations of approximately 2–5 ng/mL. Huestis and Cone² reported oral-fluid-to-plasma ratios ranging from 0.5 to 2.2 after the immediate post-smoking contamination phase had substantially cleared. This places the corresponding broad oral-fluid range at approximately 1–11 ng/mL, with a mean-ratio estimate of approximately 2.4–5.9 ng/mL.

In that context, a 10 ng/mL oral-fluid cutoff makes sense. It is analogous to setting the prescribed concentration of alcohol at 0.050 g/210 L of breath or 0.050 g/100 mL of blood: not because every individual is necessarily impaired at precisely that concentration, but because it is a rational population-level threshold.

A person who has never consumed alcohol, or who has little tolerance for it, may be impaired at 0.020 g/210 L. A seasoned alcoholic may not display obvious signs of impairment until 0.080 g/210 L or higher. We nevertheless establish cutoffs at levels that make sense for the majority of road users.

The war cry of the pro-drug parties, whether driven by political bias or scientific ignorance, has been: “test for impairment”.

The problem is that there is no reliable roadside method capable of quantitatively measuring cannabis impairment. Any clinical assessment would be subjective and qualitative: effectively, the opinion of an individual police officer. Nobody should want a system in which police (who, like all humans, are capable of bias) commence proceedings based solely on a subjective view that a driver appears intoxicated.

NSW had a comparatively fair system. The proposed reforms would replace that system with a 50 ng/mL threshold for registered medicinal-cannabis patients: five times the previous laboratory confirmation threshold and ten times the confirmatory cutoff specified in AS/NZS 4760:2019.

The proposal goes further. A registered medicinal-cannabis patient who returns a laboratory result below 50 ng/mL would face no charge or further action. An offender who returns a result at or above 50 ng/mL would receive warnings on the first and second occasions. Only on the third result at or above that threshold within two years would the ordinary penalty apply.

We may as well create the same system for alcohol: allow a person with alcohol dependence to drive with an extraordinarily high blood-alcohol concentration (say 0.250), give them two warnings when they exceed it, and charge them only on the third occasion.

Di Ciano et al.³ investigated THC concentrations and simulated-driving performance. They measured standard deviation of lateral position (SDLP, or lane weaving) and found dramatically increased SDLP when oral-fluid THC exceeded the specified 25 ng/mL threshold.

The NSW Government is proposing a no-penalty threshold of 50 ng/mL: twice the oral-fluid concentration at which Di Ciano et al. demonstrated increased lane weaving in a controlled driving-simulator study. Twice the level that was found to cause a driver not to be able to control their car.

I do not see a rational scientific basis for this sudden change. It has not made testing fairer. It has created a privileged class of drug-impaired driver and established an extraordinarily high threshold that sits well beyond concentrations attributable to impairment.

This is not a minor administrative adjustment. It is a road-safety decision.

At the very least, it discriminates against people dependent on opioids or benzodiazepines, who are not granted carte blanche to drive while impaired by their drug of choice merely because a doctor prescribed it.

I do not support granting any driver permission to drive while impaired. This proposed law⁴ is reckless, scientifically indefensible, and entirely foreseeable deaths will follow.

References

¹J. Ramaekers et al., ‘Cognition and motor control as a function of Δ⁹-THC concentration in serum and oral fluid: Limits of impairment, Drug and Alcohol Dependence, vol. 85, 2006, pp. 114-122.

²M. Huestis, E. Cone. ‘Relationship of Delta 9-tetrahydrocannabinol concentrations in oral fluid and plasma after controlled administration of smoked cannabis’, Journal of Analytical Toxicology, vol. 28, no. 6, 2004, pp. 394-399.

³P. Di Ciano et al., ‘The Utility of THC Cutoff Levels in Blood and Saliva for Detection of Impaired Driving’, Cannabis Cannabinoid Research, vol. 8, no. 3, 2023, pp. 408-413.

⁴Ministerial media release, ‘Minns Labor Government to introduce commonsense medicinal cannabis driving reforms’, NSW Government, Sydney, NSW, 2026, https://www.nsw.gov.au/ministerial-releases/cannabis-drivers-not-immediately-penalised, (accessed 5 June 2026)