With cannabis legalization spreading, the Massachusetts-based company is billing its mobile, all-purpose impairment app as an answer to a growing need for a validated test to keep stoned drivers in check.
Recreational cannabis is legal in 11 states and Washington, D.C., and while statistics on cannabis-impaired driving are hard to come by, piecemeal data seems to suggest it’s on the rise.
And yet, there are no technological solutions available for reliably testing cannabis-related impairment like police can for alcohol.
The state of Washington legalized recreational cannabis in 2012, and from 2013-14, the AAA Foundation in Washington, D.C., found that the proportion of cannabis-influenced drivers involved in fatal crashes in the state rose from 8 percent to 17 percent. The California Office of Traffic Safety found that in 2018, 42 percent of drivers killed in motor vehicle crashes in California, who were tested, had either legal or illegal drugs in their system. It doesn’t specify how many of those cases involved cannabis, but the figure was a sharp increase from 11 percent in 2017, which was a small increase from 10 percent in 2016. According to the 2017 National Survey on Drug Use and Health (NSDUH), the number of people who drove under the influence of cannabis rose from 10.9 million in 2016 to 11.8 million in 2017.
Nationwide DUI data for cannabis is scarce for several reasons: it’s still being collected, agencies are still figuring out how to assess and categorize incidents, and above all, officers don’t yet have a reliable roadside test of cannabis impairment.
A company in Cambridge, Mass., called DRUID — short for Driving Under the Influence of Drugs — aims to change that with a self-titled mobile app for assessing overall impairment using hand-eye coordination tests.
According to its website, the company was founded and launched in 2018 by Michael Milburn, a now-retired professor of psychology at the University of Massachusetts Boston for almost 40 years. DRUID Chief Operating Officer John Hayes said Milburn, now the company’s chief scientific officer, developed the app in collaboration with a programmer in China by experimenting with various methods and tests for nearly a year. With further support from a cannabis researcher at Johns Hopkins University and a Small Business Innovation Research grant, they produced an app that’s still being tweaked but is available for download now. Hayes joined DRUID a little over a year ago, helping to develop an enterprise version of the app and a back-end portal to support enterprise use. He said he appreciated Milburn’s goal of saving lives.
“(Milburn) envisioned it as a personal-use app that folks, when they thought they might be impaired, would take the test themselves and potentially not drive,” Hayes said. “It’s a combination of four tests. Each does a split-decision test that combines decision-making with some hand-eye coordination, and it involves a balance test as well. The four different tasks each require you to think about making a decision, and following up with tapping some shapes on the screen in the appropriate order.”
Some of DRUID’s partners and employees believe the app’s usefulness in government, particularly law enforcement, is uncertain but promising. Hayes said it has become the “gold standard” measure of impairment in over a dozen studies at universities and hospitals, including some in England and Australia. He said the Massachusetts State Police and the State Police Academy have also used DRUID in their training for field officers alongside sobriety tests, finding that a DRUID score of 57 corresponds with a blood-alcohol level of 0.08 for most people. The higher the score, the more impaired the user, although Hayes said it’s difficult to unintentionally score below 30 or above 70. He also distinguished DRUID from a similar app called AlertMeter, which administers a longer test focused on pattern recognition and memorization, as opposed to a two-minute test of reaction time and decision-making.
“The quote from our researcher at Johns Hopkins is, [DRUID] is the best measure of impairment that he’s ever used,” he said. “There are very few objective measures of impairment. There’s scarcely anything else that will give you a number to tell you how impaired you are, so DRUID is finding its way into a lot of research studies where they want to combine a measure of impairment with something they’re testing — opioid use, sleep deprivation, exhaustion.”
Paul Armentano, deputy director of the National Organization for the Reform of Marijuana Laws (NORML), said he was involved in the development of a similar, now-defunct app called My Canary. He said there’s an unmet need for a validated, accurate impairment test which will only grow as more states legalize cannabis. But law enforcement agencies who want to test cannabis impairment with saliva or other chemical means, as they do with alcohol, shouldn’t hold their breath. The correlation between impairment and blood or saliva concentration of THC — tetrahydrocannabinol, the psychoactive compound in cannabis — is inconsistent and depends on many other variables.
“To continue to think we’ll eventually have some sort of accurate, roadside breath-type tool that determines whether someone is impaired or not impaired by marijuana is an unrealistic expectation,” Armentano said. “I think the DRUID model, where we’re using actual measurements of performance that have been validated to be sensitive … is a much more sensible approach to this issue.”
Hayes said roadside use is a long way off, because DRUID would have to be fully tested, approved and legalized before it could be used in the field. But the company is piloting the app with a mining company now, and Hayes foresees a time when it might catch on as the nation deals with stoned drivers.
“We are getting closer to some trials and tests, and deeper in with some departments to begin those qualifying steps, but it will literally have to be written into the law before we’re using it on the roadside,” he said.
Looking for the latest gov tech news as it happens? Subscribe to GT newsletters.