Driving while stoned: Why roadside testing for cannabis might not work

As the Government mulls plans to introduce roadside drug-testing, scientists are warning it's not as easy as it sounds.

National MP Alastair Scott has submitted a Member's Bill that would give police the power to test for drugs like marijuana, ecstasy and methamphetamine. Police Minister Stuart Nash said he was open to the idea, if the technology exists.

The technology does exist - saliva tests are already being used in New South Wales - but experts say there's a good reason they haven't been adopted more widely, particularly when it comes to marijuana.

"There is no one number that can be used for a blood level or an oral fluid level that will differentiate and prove impairment," says drug researcher Marilyn Huestis, formerly of the US-based National Institute on Drug Abuse.

"It's not like we need to say, 'Oh, let's do some more research and give you an answer.' We already know. We've done the research."

Unlike alcohol, the level of impairment a marijuana user experiences doesn't neatly correlate with how much they've ingested - it can depend on how the drug was consumed, for example, and its impact can be dulled or amplified by the presence of other substances, including alcohol.

This means saliva tests - even if they accurately show how much a person has smoked - can't reliably tell police how impaired a driver might be, says Dr Huestis.

And THC quickly leaves the bloodstream - down to virtually zero in 2.5 hours - ruling blood testing out as a reliable measure of consumption, as is done with suspected drunk drivers who refuse a breath test.

"If someone is driving impaired, by the time you get their blood sample you've lost 90 percent or more of the drug," says Dr Huestis.

Mr Scott said the current police method for detecting driver impairment - getting drivers to walk in a straight line and touch their nose - is "archaic" and "ridiculous", Newstalk ZB reported on Tuesday.

Saliva testing was last looked at by the Government in 2012. It was rejected after tests showed it failed to detect cannabis use half the time, and took five minutes to administer - which the Government said was too long, considering those being tested are chosen randomly and aren't suspected of committing a crime.

Also, saliva testing only picked up meth, cannabis and ecstasy, and failed to pick up other drugs that could impair driving, such as benzodiazepines. 

Another complication is medicinal cannabis users. When used regularly, THC - the active ingredient in cannabis - can take a month to clear from the body's fatty tissues, showing up in saliva and breath tests even if they haven't consumed any recently.

Dr Huestis instead advocates for "well-trained police officers who can identify the behavioural signs of impairment", and not a codified limit in law, as there is for alcohol.

Kevin McKernan, chief scientific officer at Medicinal Genomics, told Wired there was another way.

"If you want to gauge intoxication, pull the driver out and have him drive a simulator on an iPad. That'll tell ya."

New South Wales recently announced plans to expand its saliva testing regime.

Dr Huestis' comments were published in a special issue of journal Trends in Molecular Medicine on Friday (NZ time).