Clinical Research

Told It Was THC: Inside a Balanced Placebo Experiment

·

Identical sample containers and sealed instructions for a blinded trial.

In an ordinary blinded drug trial, nobody tells volunteers whether they have received the real drug or a placebo. Jane Metrik’s research team did something much more conspicuous. They told cannabis smokers what was in a cigarette. Sometimes the description was true. Sometimes it was the opposite of what the researchers had supplied. The apparent breach of research etiquette was the point of the experiment: it gave the words spoken before a cigarette a measurable effect of their own.

The balanced placebo pilot study recruited twenty people and independently varied two things. One was what participants were told about THC. The other was whether the cigarette actually contained active THC. Crossing those decisions created four groups. Two received exactly what they had been told to expect. One expected THC and received placebo. The remaining group expected placebo and received THC. With only about a handful of people in each group, a central question was whether the arrangement would be credible at all.

Which comparison isolates the instruction?

Take the volunteers who received placebo. Some had heard that the cigarette contained THC and some had heard that it did not. A difference between those two groups cannot be attributed to one group’s cigarette containing active THC. The chemical assignment was the same. The comparison instead estimates what changed, on average, when the researchers gave a different description. The analogous comparison among recipients of active THC asks whether the description also matters in the presence of the drug.

Turn that comparison around. Among people told to expect THC, some received it and some did not. Now the wording is held constant, and the contents differ. Repeating that exercise among those told to expect placebo gives a second view of the chemical contrast. The four groups even allow a further question: does hearing that the cigarette contains THC have a larger or smaller effect when it actually does? An ordinary comparison between a correctly labeled active cigarette and a correctly labeled placebo would mix the instruction and the contents together.

Why taste became an outcome

A volunteer’s estimate of potency reflected both the cigarette received and the description given. People told they had THC also rated aspects of the cigarette’s sensory character differently. The number of puffs taken to finish it varied with the instruction. That last observation is particularly revealing. A questionnaire captures the answer a person gives after the experience. Counting puffs captures something the participant did while the experience was still under way.

There is a complication here. If the instruction changes how someone smokes, it could change the amount of smoke they actually take in, even among people assigned cigarettes of identical composition. The smoker’s behavior sits between the instruction and the outcome. Calling the observed difference an effect of the instruction makes sense. Calling it a purely mental response to an identical delivered dose would go beyond what the researchers measured.

When a volunteer changes their mind

People bring experience to the laboratory. Someone told to expect placebo might recognize familiar physical sensations and revise that expectation partway through the session. Someone expecting THC might search closely for a change in how they feel. The instruction is an experimental assignment the researcher knows. Belief is something a participant can revise. Asking afterward what the cigarette seemed to contain helps to make that distinction visible.

A review of blinding in cannabinoid trials discusses a related difficulty when patients and investigators are meant to remain unaware of treatment assignment. A later experiment on cannabis and mood used the crossed design again, giving researchers a way to investigate particular feelings rather than treating expectancy as one all-purpose explanation.

Researchers have since examined whether expecting impairment affects decisions about risk. The pilot addressed a more basic question. It showed why design matters to that question. To study the force of the words spoken before the cigarette, start with two groups given the same contents and different descriptions. The answer lies in the distance between those groups, even when both have smoked a placebo.