Josephine Drum says her daughter was "cheated out of life" when she was killed while driving to work in downtown Seattle in 2012, hit by a man in a Jeep whose blood tested positive for marijuana.
"I feel if you smoke marijuana and you have to smoke it, that you should not be able to drive under the influence," said Drum, of Stockton, Calif. "I'm 84 years old. To have lost my daughter is something hard for me to accept."
With the push to legalize marijuana surging in popularity, states want to assure the public that roads will be safe. But they face a perplexing question: How stoned is too stoned to drive?
"The answer is: Pretty damned stoned is not as dangerous as drunk," said Mark Kleiman, professor of public policy at the University of California, Los Angeles, who served as Washington state's top pot consultant.
He said Washington state has a law that's far too strict and could lead to convictions of sober drivers, with many not even knowing whether they're abiding by the law.
With no conclusive research, states are all over the map as they try to assess intoxication by measuring blood levels of THC, the main ingredient in marijuana.
There's no easy way to do it, with marijuana stored in fat cells and detectable in blood long after it's smoked or consumed, for days or weeks, depending on individual tolerance and level of use.
Washington state and Colorado, the only two states to fully legalize marijuana, have set a limit of five nanograms of active THC per milliliter of blood. In Washington state, legalization proponents included the language in the ballot initiative approved by voters in 2012.
"It appealed to the voters, but it's nonsense _ it's not a good measure of whether somebody's impaired or not," Kleiman said. "The fact that legislatures will not do their job on this means we go through the cockamamie initiative process _ it's a lousy way to write legislation."
In California, much to Drum's disappointment, lawmakers last week rejected an even tougher standard. The state's Assembly Committee on Public Safety voted to kill a bill that would have set the limit at 2 nanograms per milliliter of blood, rejecting the pleas of police officers.
And in Arizona, the state Supreme Court last month struck down part of the state's zero-tolerance law, saying it could result in convictions of sober drivers.
Some legalization proponents ridicule the statutes as "sober DUI" laws.
"What we have to understand is that arbitrary rules or zero tolerance lead to unconstitutional policing," said Diane Goldstein of Tustin, Calif., a former police lieutenant and a member of Law Enforcement Against Prohibition, a pro-legalization group that opposes the laws.
While police can use breathalyzers to easily measure the amount of alcohol in one's bloodstream, the best way to determine marijuana intoxication is by examining a blood sample.
Last year, the U.S. Supreme Court complicated the situation for states by ruling that police must get a warrant before testing blood for a DUI.
"Drawing blood is not a roadside activity for a cop," Kleiman said. "Drawing blood is a medical procedure and you need a licensed phlebotomist. So you're not going to be able to do stoned-driving checkpoints."
Ultimately, he said, a mouth swab that uses a driver's saliva to detect the presence of marijuana may be the answer, if test results can be used to track impairment.
Like many states, Washington has a team of "drug-recognition experts" _ 200 specially trained officers who can be called in to assist with marijuana DUIs, and whose evidence can be admissible in court.
Christine Beckwith, a DUI attorney from Tacoma, Wash., said the officers attend short courses but are not medical experts. And she said the blood draws make it harder for the state to prosecute a case, requiring a phlebotomist to give a sample to a police officer, who then sends it to a laboratory for testing.
"There's so many more steps for the admissibility of the blood tests that they're easier for the defense. . . . It passes through a lot of hands, where there's lots of room for error," she said, predicting that marijuana DUIs will be a lively issue for the courts in coming years.
Goldstein, the former cop, said states should "go back and rely on things we know work." She said that includes paying for more saturation patrols, better training for all officers to spot signs of impairment and more research so that laws are "grounded in science, not just political rhetoric."
"The problem is that science is lagging really far behind with drugs versus alcohol," Goldstein said. "We're going to have to deal with this issue, not just in California, but the nation as a whole."
In March, Colorado Democratic Rep. Jared Polis, who backs pot legalization, introduced a bill to create federal guidelines and "a single federal standard" for driving under the influence of marijuana. He said that lawmakers should keep impaired drivers off the roads "no matter what impaired them." But he has not lined up a single co-sponsor.
In Washington state, organizers of Initiative 502, the ballot measure that legalized marijuana, decided to include the 5 nanogram standard in the language after California voters defeated a plan to legalize marijuana in 2010. A post-election survey found that public anxieties about impaired driving in the Golden State helped kill the measure in the campaign's final days.
Alison Holcomb, criminal justice director of the American Civil Liberties Union of Washington in Seattle and the lead architect of Initiative 502, said studies are needed to know whether the law will increase safety and whether unimpaired drivers wind up getting convicted.
"I don't think we have sufficient information to answer these questions, and we should try to get it," Holcomb said.
But she defended the 5 nanogram standard, saying it was backed by existing science as a reasonable guideline to measure impairment. She noted that studies using low-potency marijuana from the National Institute of Drug Abuse found that THC levels in infrequent users dropped below the limit within two to three hours of smoking marijuana, and several hours later for those who consumed it orally. A majority of heavy marijuana users dropped to the legal limit within 24 hours of their last use, she said.
Kleiman questioned whether the studies "reflect the realities of contemporary commercial pot," with much of it having higher potency levels than the government-supplied marijuana.
"Because the pharmacokinetics are unpredictable, the (law) fails the most basic test of justice in criminal law: that a person should be able to know whether he's breaking the law or not," he said.
As the debate heats up, both sides can point to competing research.
In February, researchers from Columbia University's Mailman School of Public Health reported that fatal crashes involving marijuana use had tripled over the past decade, with one of every nine drivers now involved in a deadly accident testing positive for pot.
Kevin Sabet, a former drug adviser for Obama who now heads the anti-legalization group Project SAM (Smart Approaches to Marijuana) said that pot intoxication doubles the risk of a car crash and that laws that focus on impairment "seem justified."
Legalization advocates cite statistics showing that the number of DUI fatalities has decreased slightly in both Washington state and Colorado since 2012.
Even in California, the first state to legalize medical marijuana in 1996, Goldstein said that DUIs for alcohol remain the top concern for police.
"If we had a crisis of fatality rates, law enforcement would have been screaming about this going back to 1996, but what we have is probably some of the safest roads in our nation," she said.
Kleiman of UCLA sees good news and bad news in the issue.
"The bad news is at the moment we don't have have anything sensible to do about stoned driving," he said. "The good news is that it's only a moderate-sized problem. . . . I would be nervous about over-criminalizing it."
Drum, whose daughter was killed in the Seattle crash, is angry that the driver who caused the eight-car pileup received a sentence of only 54 months.
"I don't think it's sufficient," she said.
Drum said her daughter, 56-year-old Rosemary Tempel, worked as a nurse at a Seattle hospital for more than 30 years. She broke her neck in the accident.
"She was going to work that morning, and to have a reckless driver hit her head-on, and the way she was killed, it's something hard to take," Drum said. "I can't understand how a person that doesn't have the ability to do it safely, why they're on the streets."