Tech companies are racing to crack the code of marijuana testing
But it's a tricky physiological puzzle.
July 7, 2022
· 12 min read
If you smoke a joint or eat a space cake, marijuana’s active ingredient THC (technical name: Tetrahydrocannabinol) moves directly to your brain where it releases dopamine. You feel chill, happy, and hungry for cool ranch Doritos. Though your cravings subside after a few hours and that sweet 420 feeling fades, THC can linger in your bloodstream for weeks or even longer. That’s a problem if you have to take a drug test. Say, for example, you’re a pilot required to take semiregular drug tests. Even though you’re not high and haven’t been for days, a sample of your blood or saliva might tell a different story to your employer, one that says you could be impaired. That’s like failing a sobriety test a week after having a few beers at a party.
Even though marijuana is legal in 19 states and Washington DC, there’s no widely accepted way to measure real-time impairment caused by the substance. That’s because THC doesn’t affect people the same way alcohol and other drugs do. Instead of working its way quickly through the body, THC lingers, rendering standard drug testing regiments—through blood, hair, or urine tests—largely useless and unable to determine when a person last used. Plus, THC hits people differently; frequent users may not feel high despite having large amounts of THC in their blood, but a new smoker might struggle to walk a straight line after taking a hit. Pinpointing impairment to marijuana use is a tricky physiological puzzle that has so far bested researchers. But tech companies are racing to crack the code and, in the process, capitalize on a market potentially worth billions of dollars. A market analysis estimates the global drug testing market will be worth about $10.7 billion by 2030. Marijuana testing leads the market.
That’s a hefty incentive for tech companies and a sizable enough profit to make navigating all the nuances of THC testing worth it. There’s no consensus on what type of tech is best, and there’s no agreement that any options really work. That’s cleared a path for creative approaches to the problem. There are breathalyzers in development that would analyze samples for teeny tiny bits of THC that are only present soon after someone consumes. Then there are cognitive tests that use computers or phones to ask what should be easy questions. They try to see if a slower response time means someone is too high to work or drive. The developments are highly scrutinized by legal experts wary of smartphones and breathalyzers promising to solve the evasive problem. The tech will need to survive court challenges to become standard-bearers in the way that a .08 blood alcohol level has. Some wonder if it can ever live up to its promises of solving the drug testing issue.
“It’s definitely going to take many more years,” said Dori Stibolt, an employment and labor attorney. “The wheels of justice, the legal process, is very, very slow. And it’s just going to take time for those cases to come up.”
The race for reliable tech is playing out against a larger cultural shift on marijuana, legalization, and decriminalization. Fears of reefer madness are fading and the stigma around marijuana use is dissipating. But employers, especially those with workers who operate heavy machinery or drive, are stuck between a rock and a hard place; quit drug testing and risk a lawsuit following a workplace accident or hikes in insurance costs, or keep traditional drug testing, and potentially be sued for wrongful termination by an employee who tests positive for being impaired at work. It’s also made the landscape chaotic for employees who wonder if it’s okay to smoke a bowl on the weekend or if they’ll lose their jobs after a random drug test days later. The stakes are high, too, for medical marijuana patients, who are prescribed cannabis for health conditions and could fail a drug trust because of their medicine. The dilemma has led many to question whether drug testing is really worth the costs. Amazon announced last year it would end marijuana testing for its employees unless required by the US Department of Transportation. (That move benefited Amazon as well: “We’ve found that eliminating pre-employment testing for cannabis allows us to expand our applicant pool,” Beth Galetti, senior VP of human resources at Amazon, wrote in a January memo.)
Employers are split on how to handle America’s favorite drug. The National Safety Council conducted a 2021 survey of 500 employers and 1,000 workers in states that have legalized marijuana for medical or recreational use and found that just over half of those companies have decided to do away with marijuana testing altogether. Some see no equitable way to test workers for a drug now consumed at high-end dinners or in Las Vegas lounges, particularly because marijuana laws and testing have disproportionately affected Black and Latinx people both legally and economically.
Legalization is a reversal of century-old, flawed practices of outlawing marijuana, but drug testing workers is newer. In 1986, then-President Ronald Reagan signed an executive order mandating that federal employees and employees at companies with federal contracts undergo drug testing. Private employers took this as a cue and began drug testing, too, despite questions about its legality. The moral and legal panic turned drug testing companies into giants (Quest Diagnostics, which also runs medical testing, made $10.8 billion in revenue in 2021), and American employers have since crafted policies that are absent or illegal in most other countries. Stringent testing has made positive tests in the workplace fall over time. According to Quest Diagnostics, the rate of drug test positivity among workers declined from 13.6% in 1988, and reached a low of 3.5% in 2012.
For all the time and money spent on keeping workplaces “drug-free,” researchers don’t even agree on the effectiveness of workplace testing policies. One survey in 2011 commissioned by the Drug and Alcohol Testing Association found that drug testing policies led to less turnover and absenteeism and more productivity. But testing policies can also lead to distrust between workers and management. And just like arrests for drug use, testing is unevenly applied. A 2013 study from the Yale School of Medicine found that Black and Latinx workers were more likely to work in places that require drug testing.
This new generation of marijuana tests aims to take subjectivity and bias out of the process. Without such a test, police rely on drug recognition experts, or DREs, who are cops specifically trained to spot signs of impairment when pulling someone over. They’re expensive, and the method has been widely scrutinized and challenged in court as ineffective and racially biased. Despite the concerns with it, it’s the model New Jersey copied for its novel workplace impairment recognition experts, who will be trained to spot signs that a worker is high, and order a physical test to back up their claim. (New Jersey has not rolled out the regulations around these experts, but it did open marijuana dispensaries to the public in April.) If that sounds like a system rife with the potential for misunderstanding or abuse, then experts agree.
New marijuana tech companies want to eliminate the hazy gray areas altogether by replacing human bias with a one-size-fits-all test. Ian Stewart, an attorney whose work focuses on marijuana and hemp law, said he thinks it will be a long time before such a test works. “If an employer is going to take some adverse action based on unproven technology, that can be challenged,” he said. “I think that’s a problem. The technology has not been proven.” They’ll first just have to prove such a single approach works on a drug known to affect people differently—that’s no easy task.
Enter cognitive testing, or trying to figure out if someone is high by checking their skills instead of their blood. It’s like a high-tech field sobriety test where police ask drivers to recite the alphabet backwards or touch their noses. There’s Druid, an app developed by a retired psychology professor. Its creation came as legalization efforts gained steam in 2016, and it tries to gauge whether or not someone is unable to do potentially dangerous work, like operating machinery.
When you open the Druid app, a circle flashes on the black screen. You have to tap the circle quickly, as it disappears and reemerges near another corner. Soon, there are flashing squares in the mix, too, and you have to tap on a stationary oval toward the top of the screen each time one appears. It’s a dizzying, fast-paced exercise, and each time your finger meets the screen, a bell sounds. Continue to tap the rapidly appearing circles, but now try to estimate when 30 seconds have passed, instructions from the app say next. Can you follow one circle with your index finger as it zips across the screen, and count the number of squares that flash in the background? Balance on your left foot for 15 seconds, then your right one. All of this is part of a brief cognitive test conducted on a smartphone or tablet to determine whether or not you’re still high, without wading into the THC problem.
Rob Schiller, CEO of Impairment Science, the company behind Druid, said the test measures general impairment, including slowed reactions from weed, alcohol, another drug, fatigue, or illness. But people must take the test several times to create a personal baseline, and eventually, results can try to gauge if the test taker is way off that baseline. To return to the example of a pilot: If she fails, it doesn’t necessarily mean she came expecting to fly the friendly skies while stoned, it simply means she shouldn’t be flying (or doing anything else dangerous). According to Schiller, the app “found a way of assessing cognitive information that had nothing to do with whether a substance was in your body—[it’s] testing how your brain works, not a proxy for how your brain works.” A drug test “looks for some sort of chemical in your body,” Schiller explained. “And in some instances, the presence of that chemical correlates with actual cognitive impairment. It doesn’t always do so.”
Hound Labs, perhaps the most recognizable name in new marijuana testing tech, has developed a Breathalyzer-like device to detect THC down to the parts per trillion for a few hours after someone consumes marijuana. That makes it 1 billion times more sensitive than an alcohol breath test, said Jenny Lynn, the company’s co-founder. A 2019 study conducted by Hound Labs with the University of California, San Francisco, backed up the company’s tech: The breathalyzer could detect THC in breath if people had used marijuana within three hours, a timeframe in which research shows impairment would be likely. It’s an entirely different approach than having people react quickly to shifting shapes. “There needs to be a deterrent to detect: Have you used recently? Have you used within a timeframe that could impact your performance?” Lynn said.
After eight years of research and development, Hound Labs is close to bringing the tech to the market: The breathalyzers are expected to be sold later this year. (Testing the product lent itself to some unconventional methods: In 2017, Hound Labs had drivers around a 1.5-mile race track sober, then get absolutely baked. They waited 30 minutes and navigated the course again, driving up to 65 miles per hour, with a professional race car driver in the passenger seat. They hit debris from construction and standard road signs while high.) There are pre-orders already for the device, largely from employers, according to Lynn. Hound Labs said in September it had raised $20 million to scale its production. It sees its tech as a way to cut through the bias of old drug testing. “We provide objective data,” Lynn said. “We provide one more tool to help make decisions at the roadside or with employers.” But again, the breath tests don’t measure how high someone is. They just detect recent use, which does correlate to impairment for some people. It’s a start, but it’s not the silver bullet.
At least one company started a test with a different purpose in mind and is adapting it to focus on marijuana. Cognivue, a New York-based company, began as a medical device designed by a neuroscientist to detect neurological decline, like Alzheimer’s disease, which has traditionally been done via a pen-and-paper test. But the company soon realized the test, which requires people to track a group of moving dots, could also show if someone were impaired. (Cognivue is still refining the marijuana impairment version of the test.) The FDA cleared Cognivue in 2015 for testing for Alzheimer’s. “Our test eliminates most of the bias” from traditional neurological tests, said Kristin Weber, director of strategic accounts for Cognivue: “You have a lot of socioeconomic bias with that. And whenever you have a human administering a test to another human, it’s tainted, right? There’s subjectivity. That person delivering the test is having a bad day, or if you have racial differences, whatever it might be, it’s going to affect the patient’s score.”
Without a test that can say absolutely, yes, you’re impaired right now and it’s because of the weed you smoked, there are only imperfect solutions. The testing tech companies boast studies backing up their research, and are full of eager optimists—who are also eager about the potential profit. These new technologies could be the answer that lets workers blaze it on the weekends without risking their jobs on Monday, or even weeks later. But the tech is still in its infancy, and a long way from a simple solution for either workers or employers. Consuming cannabis might be legal in many places, but dated drug testing policies aren’t as chill.