How does the placebo effect work?

The question

In medical trials, patients who receive a dummy pill with no active ingredient often get better. Their pain decreases, they sleep better, their mood improves. This is why every new drug has to prove that it works better than a placebo, not just better than nothing.

But two very different things hide behind the word. One is real: expectation and context actually change what happens in the body. The other is an illusion of measurement: many people would have improved anyway. The question is how big the real part is, what it consists of, and how a belief in a mind becomes a change in the body.

What we know

In 1955 Henry Beecher published "The Powerful Placebo", claiming that about a third of patients respond to placebos. The figure became famous, but later analysis showed that much of it was not a placebo effect at all: illnesses naturally get better, symptoms fluctuate, and people tend to enter trials when they feel worst, so they improve on average simply afterwards.
When trials include a third group that gets no treatment at all, the picture changes. Large reviews from 2001 onwards found that compared to no treatment, placebos have little effect on objective measures like blood tests, but a moderate effect on what patients feel and report themselves — above all pain and nausea.
Placebo pain relief has a biology. In 1978 researchers showed that naloxone, a drug that blocks the body's own opioids, reduces placebo pain relief. Brain imaging shows that when people expect relief, activity drops in brain areas that process pain, and the brain releases its own painkillers.
Feeling better and being better can separate. In a 2011 asthma study, a placebo inhaler and sham acupuncture made patients feel as much better as the real drug, but their lung function didn't improve at all.
It works even without deception. In several studies, people who were openly told they were taking sugar pills still reported less pain or fewer symptoms than those who took nothing.
The reverse exists: the nocebo effect. Expecting side effects makes them more likely, and a warning about pain can increase pain.

Main hypotheses

Expectation

The brain constantly predicts what will happen, and its predictions shape what it perceives. If it expects relief, the incoming pain signals are interpreted and dampened accordingly. This is the main explanation for placebo pain relief.

Conditioning

The body learns from experience. If a pill has relieved pain before, a similar-looking pill can trigger part of the same response, even in someone who doesn't consciously expect anything. Conditioned placebo effects have been shown even for immune responses.

The ritual of care

Attention, a warm doctor, an injection rather than a pill, an expensive drug rather than a cheap one: all of these can increase the effect. Perhaps the placebo isn't the pill but the whole experience of being treated.

Where understanding ends

We know that placebo effects are real for pain and some other symptoms, and we know some of their chemistry. We don't know how far they reach. Can expectation change the course of a disease, the size of a tumour, the healing of a wound — or only how a person feels? Reliable evidence says mostly the latter, but the boundary hasn't been drawn.

It is also unclear why responses vary so much. Some people respond strongly to placebos, others hardly at all, and the same person can respond differently at different times. Attempts to find "placebo responders" by personality or genes haven't produced reliable results.

And there are strange trends. In American trials of drugs for nerve pain, the response to placebo has grown over the decades, while elsewhere it hasn't. Nobody knows exactly why: longer and bigger trials, advertising, changing expectations — all are suspected.

Finally, open-label placebos challenge the very idea. If a pill works even when the patient knows it is empty, then what exactly is the mind responding to?

What would settle it

Large trials that always include a no-treatment group, so that the true placebo effect can be separated from natural improvement.
Studies linking the brain mechanisms of expectation to objective changes in the body, showing where the effect stops.
A reliable way to predict who will respond to a placebo and when, which would also show what the effect depends on.