Patrick Murphy saw the flashing lights come on behind him on State Route 115, and he immediately knew what it was about. He was on his way to get an estimate to fix a broken headlight, but it was the day before Thanksgiving, typically a big day for traffic enforcement. Of all the bad luck, a Pennsylvania state trooper had spotted the damage first.

Murphy didn’t know that his headlight was just the start of his problems that day in 2017. He suffers from chronic pain and PTSD, and while visiting relatives in Massachusetts six days earlier, he used cannabis to get relief.

Murphy’s drug use was legal — as Massachusetts already allowed it then — and he was driving safely. But his system still contained trace amounts of an inactive metabolite of the drug, and under Pennsylvania law, that meant he was driving “under the influence.” After a long court battle — during which even the state’s expert witness agreed the metabolite was no indication of impairment — Murphy spent three days in jail, and lost his license for a year.

As more states move to legalize cannabis for medical or recreational use, an increase in DUI arrests is likely to follow, say both law enforcement officials and advocates for legalization. Yet scientists warn that unlike with alcohol, detecting impairment due to cannabis is a challenge — meaning that some of these arrests may, as in Murphy’s case, involve motorists whose driving is perfectly safe.

“I’ve been fighting this the whole way,” he said.

» READ MORE: What to do if you're pulled over after using cannabis

Pennsylvania’s medical marijuana program took effect in 2018, yet it remains one of a dozen states with a zero-tolerance law — meaning it is illegal to drive if the motorist’s body contains any amount of THC (the main active compound in the drug) or its inactive metabolites, according to the National Conference of State Legislatures.

Some other states define impairment as having THC in the blood above a certain threshold, but scientists say that approach also falls short. There is wide variation in how people metabolize cannabis, plus there is little to no correlation between any given level of THC in the blood and the ability to drive.

Then there’s the debate underway in New Jersey, where long lines of customers greeted the legal debut of recreational cannabis in April. In the Garden State, a drugged driving conviction requires evidence of actual impairment — the officer’s observations coupled with other findings such as drugs in the car.

But typically, prosecutors also enlist the input of a second officer called a “drug recognition expert” — one trained in a 12-step protocol that is now being challenged before the state Supreme Court. These officers assess the suspect’s muscle tone, pulse, and eye movements, among other vital signs that are normally the province of a doctor’s office. A court-appointed “special master” heard more than four months of testimony on whether such findings should be admissible as expert opinion, and his findings are expected this summer.

No one wants impaired motorists on the road. But the bottom line, say experts in how the body responds to cannabis, is that the science has not yet caught up with the law.

The science of weed

Despite the state legalization efforts, cannabis remains an illicit substance in the eyes of the U.S. government.

That means a lot of paperwork for Tory Spindle, an experimental psychologist at the Johns Hopkins University School of Medicine. To study the drug’s effects in his lab, he needs special approval from three institutions: the school, the FDA, and the U.S. Drug Enforcement Administration.

These hurdles help to explain why cannabis has received less study than alcohol and other drugs, he said. And if anything, the actual science of pot is even more difficult.

Chemical components of the plant, or their metabolites, can be detected in the body weeks after it is consumed — long after any impairment has worn off, Spindle said. On the other hand, a person may be impaired despite having no detectable levels of the drug in the blood — as the active compounds can be present in the brain. And the way the drug is consumed, whether by smoking, vaping, or eating edibles, also affects the measurements.

“A given level of nanograms per milliliter in someone’s blood is basically meaningless,” he said.

In other words, there is no pot equivalent of the Breathalyzers used for alcohol, for which readings are directly correlated with the ability to drive.

That leaves measuring cannabis impairment directly with behavioral and cognitive tests. But those types of evaluations also can be a challenge, Spindle and other scientists say. Habitual users can develop a tolerance, meaning that a given amount of the drug results in less impairment than it would for novice users. Then there’s the difficulty of distinguishing impairment from the effects of medical conditions or age.

Yet there is no question that cannabis can impair the ability to drive. The drug slows reaction time, and it may impair decision-making. Drivers who are high are more likely to weave outside their lane, and they may also drive below the speed limit, though not always, said Robert Pandina, a neuropsychologist and professor emeritus at Rutgers University.

In a law enforcement setting, the solution to detecting cannabis-related impairment may be some combination of behavioral and chemical tests, Spindle said. But the science is not there yet.

“This is a hard question,” he said. “Much more research is needed.”

The busted headlight

When he was pulled over in Effort, Pa., 90 miles north of Philadelphia, Patrick Murphy was not speeding or otherwise driving unsafely.

But upon approaching the car, Trooper Matthew Kizis said he smelled cannabis.

Not possible, Murphy replied. There were no drugs in the car, as the trooper’s search would later confirm. And while Murphy admitted he used cannabis for medical reasons, he told Kizis he had not done so in nearly a week.

He now wishes he’d kept his mouth shut.

Kizis told him to get out of the car and walk in a straight line. Murphy struggled to keep his balance, explaining that he suffered from chronic back and knee pain, but Kizis was unmoved. He arrested Murphy on suspicion of driving under the influence, and said Murphy needed to come in for a blood test. Murphy agreed, after learning that if he refused, he would lose his license for a year.

The result: His blood contained 5.1 nanograms per milliliter of an inactive metabolite of THC, the lowest detectable amount. “You could not determine impairability” from that reading, a toxicologist would later testify for the state.

Murphy appealed his conviction all the way to the state Supreme Court, which declined to take the case. In addition to the three days in jail and losing his license, the Albrightsville man owed thousands in fines and legal fees. The one-year loss of his license was delayed during appeal, taking effect last fall. Now, every day, his wife drops him off at his job as a warehouse associate and picks him up in the evening, adding an hour in each direction to her own commute.

A rise in arrests?

All sides of the debate predict that expanded legalization of marijuana will lead to a rise in arrests. In Pennsylvania, DUI arrests went up after the medical marijuana program started, though the totals do not distinguish between drug and alcohol arrests.

Pandina, the former Rutgers professor, is now a consultant who has advised lawmakers in several states who are concerned about how to cope with the expected increase.

“They’re actually pulling their hair out,” he said.

In Colorado, where sales of the drug for recreational use became legal in 2014, the share of DUI arrests involving marijuana rose from 12% that year to 31% in 2020 — though many of the arrests also involved alcohol or other drugs. Less clear is whether cannabis use increases the risk of crashes. In a 2015 National Highway Traffic Safety Administration study, researchers found no correlation between THC levels and crash risk, after taking demographic variables into account.

In New Jersey, law enforcement officials are girding for an “inevitable” increase in impaired driving, Burlington County Prosecutor Scott Coffina said. Edibles that contain cannabis are of particular concern, he said.

“The sometimes-delayed impact of edibles can really compound this problem, as people start their journey thinking they are OK, and then the edible kicks in and they are suddenly significantly impaired,” he said.

The showdown in New Jersey

The Supreme Court case in New Jersey involves Michael Olenowski, who was twice convicted of driving while impaired after police drug recognition experts testified he was under the influence of a depressant and a stimulant.

Though his case did not involve marijuana, it has prompted the court to review the use of drug recognition experts for that drug and others. These officers undergo 12 days of specialized instruction in taking a person’s vital signs and other techniques, said Jackson Township Police Sgt. Michael Kelly, president of the New Jersey Association of Drug Recognition Experts.

In its 601-page filing, the state Attorney General’s Office cites evidence that DRE programs are widely accepted in the traffic safety community, and that their components are grounded in years of research. Defense attorneys counter that this type of testimony belongs solely in the hands of physicians and scientists with relevant expertise in the brain.

In Pennsylvania, where DRE programs are less common, as the mere presence of a drug compound in the blood is enough for a conviction, medical cannabis users are wary.

Quari Williams, 23, of Bensalem, said he had been arrested at age 17 for DUI, hours after any effects from the weed he’d smoked had worn off. He now has a medical marijuana card, and takes pains to keep his car in good repair, in addition to not driving when high.

“I try to do whatever I can to keep myself in the clear and not break the rules,” he said.

Murphy, the Poconos man who tried to overturn his conviction, worries that others like him are at risk.

“After everybody’s already paid their money, after the patient has legally purchased their medicine, after the patient has legally consumed their medicine — the commonwealth turns around and says, ‘That’s not a medicine,’” he said. “That sounds a bit like fraud to me.”

Staff writer Ryan W. Briggs contributed to this article.