High Yesterday, Positive Today: The Problem With THC Testing

A friend once told me he had stopped smoking for more than a week before taking a pre-employment drug test. He wasn’t going to work high. He hadn’t smoked that morning, the night before or even the weekend before. He walked into the testing center feeling pretty confident, peed in the cup and went home.

He failed.

His first reaction was basically, “How can I fail a drug test when I’m not high?”

That question exposes one of the strangest problems we still have with cannabis. We have gotten very good at finding evidence that somebody consumed THC. We are considerably worse at determining whether that person is actually impaired right now.

With alcohol, most people understand the basic idea. Alcohol concentration rises and falls relatively quickly, and breath or blood testing can provide useful information about the amount of alcohol in someone's system. Cannabis doesn't cooperate nearly as nicely.

THC is fat-soluble, and after cannabis is consumed the body begins breaking it down into metabolites. Some of those metabolites can remain detectable long after the noticeable effects are gone. In federal urine testing, laboratories aren't even looking for the THC that originally produced the high. They test for a metabolite called THC-COOH, or THCC, which indicates that cannabis was consumed sometime previously. SAMHSA's current federal workplace-testing guidance confirms that urine testing reports this marijuana metabolite rather than active THC.

That difference matters.

A person can test positive after the impairment from cannabis has ended. Exactly how long depends on the person, how often they consume, the dose, their metabolism, the type of test and other factors. National Highway Traffic Safety Administration guidance notes that THC metabolites can remain detectable in blood or urine for days or even weeks after consumption—well beyond the period of impairment.

So a positive urine test can answer one question fairly well: Has this person used cannabis?

What it generally cannot answer by itself is the question employers, police officers and courts may actually care about: Was this person impaired at that particular moment?

Cannabis Isn't Alcohol

This is where attempts to create a cannabis version of the breathalyzer get complicated.

THC concentration doesn't track impairment in the clean, predictable way people often assume. NHTSA notes that research has found little evidence of a direct dose-response relationship between a person's blood THC concentration and their level of driving impairment. Cannabis can impair reaction time, divided attention, lane control and other skills important for driving, but putting a single number on that impairment is difficult.

That does not mean cannabis can't impair someone. It absolutely can.

It means detecting THC and proving impairment are two different things.

Some states have adopted numerical THC limits for driving cases, while others rely more heavily on observed impairment, toxicology and the circumstances surrounding the stop or crash. The problem is that two people with the same THC concentration may not necessarily experience identical impairment, particularly when comparing an occasional consumer with someone who uses cannabis frequently.

Newer testing methods may narrow the window. Oral-fluid testing, for example, looks for THC itself rather than the inactive marijuana metabolite used in urine testing. That can provide evidence more closely associated with recent consumption. But even a shorter detection window isn't the same thing as a machine that flashes a number and announces, “This person is impaired.”

We aren't there yet.

The Workplace Is Starting to Notice

For decades, many employers treated a positive cannabis test as the end of the conversation. But legalization has forced states to reconsider whether someone's legal activity on Saturday night should automatically affect their employment on Monday morning.

The rules now vary dramatically depending on where you live and what kind of work you perform.

California, for example, provides protections for many workers who legally use cannabis away from work and generally restricts employment decisions based on tests that detect nonpsychoactive cannabis metabolites. Those protections do not give employees permission to possess cannabis, use it or be impaired while working, and several exceptions exist.

New York goes further in many situations, generally protecting lawful off-duty cannabis use and restricting employer cannabis testing unless an exception applies. Employers can still act when an employee shows specific, observable symptoms of impairment that affect job performance or workplace safety. Washington also limits the use of tests for nonpsychoactive cannabis metabolites in many initial hiring decisions, while maintaining exceptions for certain safety-sensitive jobs and federally required testing.

Those distinctions make sense when you consider what the test is actually measuring. Finding evidence that someone consumed cannabis isn't necessarily the same thing as finding evidence that they showed up unable to do their job.

But there is a very important exception to all of this: safety-sensitive and federally regulated work can be a completely different world.

Truck drivers, pilots, train employees and other workers covered by Department of Transportation regulations remain subject to federal marijuana testing requirements. DOT reiterated in 2026 that state-authorized marijuana use does not turn a federally regulated positive marijuana test into a negative one, even when someone holds a state medical-marijuana card or bought the product legally under state law.

When you're responsible for an 80,000-pound truck, an airplane full of passengers or other safety-critical equipment, the rules are understandably much stricter.

What Are We Actually Trying to Measure?

Nobody should be driving a car, operating heavy machinery or performing dangerous work while impaired. Cannabis legalization doesn't change that, and responsible cannabis consumers shouldn't pretend otherwise.

But fairness requires asking the opposite question too.

If someone used cannabis legally two nights ago, slept eight hours, woke up sober and went to work, should a test that detects leftover metabolites treat that person the same as somebody who consumed cannabis immediately before clocking in?

That's the conversation our testing technology has forced us to have.

Maybe the future of cannabis testing won't be about proving that THC has passed through someone's body. We already know how to do that. The real breakthrough will be finding reliable ways to determine whether cannabis is affecting someone's ability to safely perform a task right now.

Until then, a positive THC test may tell us something.

It just doesn't necessarily tell us what we think it does.