While legislation regarding medical and recreational cannabis use may have been relaxed in numerous countries over recent years, driving while stoned remains a criminal offense everywhere, and for good reason. However, enforcing this common-sense law is not as easy as it might sound, with new research indicating that roadside tests for marijuana are wildly inaccurate.

Tetrahydrocannabinol (THC), the main psychoactive compound in cannabis, has been found to slow reaction times and impair alertness in drivers, while also reducing their ability to drive in a straight line. Because of these and several other reasons, police in many jurisdictions carry roadside testing kits that are supposed to reveal if a person has any THC in their saliva.

However, an as yet unpublished paper by researchers at the University of Sydney found that the testing devices used by officers in New South Wales give false results 20 percent of the time. The study authors conducted an array of tests on hundreds of drivers who had been given varying concentrations of THC, to see how this affected their driving. When they then tested participants’ saliva, they were shocked at the imprecision of their results.

Speaking to ABC News, lead researcher Ian McGregor explained that “we had someone test positive for THC who was using a placebo,” while others who had ingested high concentrations of the compound tested negative.

This research follows on from a paper that appeared earlier this year in the journal Public Health, analyzing the performance of the eight most commonly used roadside testing devices. The authors found that on the whole, these kits were very bad at detecting high concentrations of THC in saliva, meaning people who are dangerously high could return a negative sample.

Furthermore, the tendency of these devices to detect cannabis in the saliva of those who have not smoked raises the possibility that they may be picking up THC from passive smoking. This had previously been thought impossible, and the results of this research therefore reveal just how much we don’t know about the drug’s interactions with the body.

For this reason, an Australian woman who tested positive for cannabis while driving recently had her charges dismissed after police couldn’t disprove her claims that she had ingested THC through passive smoking.

Most roadside testing devices are thought to be able to detect cannabis for up to four hours after use, and when someone tests positive, samples are sent to a laboratory for further examination. THC can remain detectable in urine for up to a month, in blood for around two weeks, and in hair for three months.

However, even these lab tests are not without their limitations, as one study found that marijuana smoke can get into the hair of nearby non-smokers. All things considered, it seems the best option is to just not drive.

[H/T: ABC News]