It is eleven at night and the kitchen table is the clinic now. The leaflet from the pharmacy is unfolded next to the mug. Someone has typed the drug name into a search bar and added the word "side" and the rest has autocompleted itself. Nausea. Dizziness. Falls. Confusion. By the time the kettle has gone cold, the person who was going to start a new medicine on Monday has read twenty accounts of strangers who stopped it, and the medicine has acquired a biography before the first dose. This is the ordinary evening that a group of psychologists in Ohio has just tried to measure.
Their question was narrow and, once you see it, obvious. Nocebo effects are real symptoms that follow from expecting harm, and the usual route in is the information itself: you learn a drug can cause dizziness, and dizziness becomes easier to notice and easier to have. Most attempts to soften that have aimed at the symptom. Emily Spotts, Ananya Chatterjee, Legend Coleman and Andrew Geers, of the Department of Psychology at the University of Toledo, aimed instead at the eleven o'clock search. In a paper published online ahead of print in Psychology and Health on May 7, 2026, they report three pre-registered online experiments with what they describe as diverse community samples, testing whether teaching people about the nocebo effect changes how much side-effect information they then go and collect. In all three experiments it did: nocebo education reduced the behaviours associated with acquiring treatment side-effect information. In two of the three, that reduction was statistically mediated by a self-reported desire to avoid side-effect information, which is to say people did less looking because they wanted to look less. A review written by two of the same authors in Social and Personality Psychology Compass describes the behavioural measure plainly: people given nocebo education selected fewer pieces of side-effect information than people who watched a control video.
The results that did not happen matter as much as the one that did. Receiving nocebo education did not reduce satisfaction with the treatment information people ended up with. It did not raise their side-effect expectations. It did not raise treatment anxiety. That combination is the interesting part. The intuitive worry about telling patients "some of what you may feel comes from expecting to feel it" is that you have handed them one more thing to be uneasy about, or that you have talked them out of knowing what they are entitled to know. In these experiments, neither showed up on the measures taken.
This is also the point at which a line was crossed from wanting to doing. In 2024, the same Toledo group, with Kim Goerner as first author, published two studies in the Journal of Behavioral Medicine, N equals 220 and N equals 252, in which adults watched a nocebo education video and then read vignettes about a surgical treatment and a medication for pain. Those studies found that nocebo education reduced global side-effect expectations, increased people's belief that they could handle side effects, increased their intention to avoid side-effect information and decreased their intention to seek more of it. Intentions, though, are cheap. The 2026 paper is the attempt to show that the intention turns into an act, and within the walls of an online experiment it did.
For families living with a dementia diagnosis, the reason to care is not abstract. Adverse events reported by people taking nothing at all are a large and stubborn feature of Alzheimer's drug trials. Panagiotis Zis and Dimos Mitsikostas, analysing 20 randomised controlled trials in Alzheimer's disease in the Journal of the Neurological Sciences in 2015, found that of 3,049 patients given placebo, 57.8 percent reported at least one adverse event and 6.6 percent stopped the placebo because of those events. That is a pooled analysis of trial arms, not an experiment on expectation, and it cannot tell you which of those symptoms were caused by anticipation and which were the ordinary weather of illness and age. What it does show is how much of the side-effect burden in this field sits in the gap between a drug and a person's reading of it.
And the tension is genuine, not rhetorical. A carer who reads less is a carer who may miss the swelling, the bleed, the vomiting, the fall that is real and needs a phone call today. The Toledo experiments did not test that. They did not enrol anyone with dementia, they did not involve anyone actually taking a medicine, and they did not follow anyone to see whether a person who searched less went on to feel better, adhere longer, or miss something they should have caught. What they establish is smaller and still worth having: that a short explanation of the nocebo effect is a lever on behaviour, not only on feeling, and that in these samples it pulled without visibly costing people their peace of mind or their sense of being adequately informed.