Every clinic appointment contains a small act of theatre. A clinician reaches for a leaflet, or says the words "this one can cause", and something happens in the room before anything happens in the bloodstream. Families living with a dementia diagnosis know the sequence intimately: the new tablet, the printed list of possible harms, the long weekend of watching for them. The Leicester review that landed in Frontiers in Psychiatry this spring is about exactly that moment, and about the uncomfortable finding that the watching itself can change what is felt.
The paper, by Billy Church, Pankaj Gupta, Adarsh Shivaram, Sian Jenkins and Emma Palmer of the University of Leicester and NHS England's Leicestershire, Northamptonshire and Rutland GP training programme, is a systematic review rather than a new experiment. The team followed PRISMA 2020 reporting standards, preregistered the protocol on the Open Science Framework, and searched MEDLINE, Web of Science, PsycINFO and CINAHL up to 21 August 2025. They wanted studies that had deliberately induced nocebo effects, the mirror image of placebo, in adults with physical or neurological long-term conditions, using negative verbal suggestion, conditioning, or negatively framed clinical information. Thirteen studies met the bar. Seven used within-subject designs, four were between-subject or parallel-group, and two embedded the expectancy manipulation inside real clinical treatment. Quality was appraised with the Mixed Methods Appraisal Tool. Because the induction methods and outcome measures were so different from one another, the authors did not pool the results into a meta-analysis; they grouped and described them instead.
What the thirteen showed is consistent enough to be worth a family's attention. Negative words and negatively framed information reliably produced more intense symptoms as people rated them. Several studies also picked up changes in physiological markers and in patterns of brain activation tied to expectancy, which is the part that answers the reflex objection that this is all just reporting bias. The review's own summary is blunt: "Individuals with long-term conditions can show strong nocebo responses across a range of symptoms." But the strength was not uniform. The effects were most robust in subjective symptom ratings, while physiological and neurobiological measures were selective and context-dependent, and the evidence for neurological symptoms specifically was mixed. Anxiety, the moderator most people would guess at first, was measured inconsistently across the studies, so the review cannot say who is most susceptible.
For readers of this section, the conditions list is where the story turns personal and then stops short. The thirteen studies covered chronic pain, Parkinson's disease, irritable bowel syndrome, asthma, atopic dermatitis, benign prostatic hyperplasia, chemotherapy-treated cancer, and psychiatric outpatients taking psychotropic medication, with one study using an analogue medication paradigm. Parkinson's disease is the nearest neighbour to our territory, and it is also where the review found the evidence least settled. Dementia does not appear among the conditions the authors list. Nobody has run these experiments in people living with Alzheimer's disease, Lewy body dementia or frontotemporal dementia, and this review is not evidence about them.
What it is evidence about is the machinery of the clinical encounter, which dementia care shares with every other long-term condition. The authors note in their background that people discontinue effective medicines because of symptoms they believe the treatment caused, even when those symptoms are nonspecific or unrelated, and that clinicians can inadvertently feed the process by overemphasising possible adverse events or offering risk information without context. That is a description of a communication problem, not a character flaw in patients, and the review's recommendation is aimed squarely at the professional side: more consistent methods in research, routine measurement of psychological moderators, and communication strategies with evidence behind them.
The practical shape of this for a family is narrower than the headline suggests, and honesty requires saying so. Nothing here tells anyone that a side effect they are reporting is imaginary, and nothing here is a reason to stay quiet about one. A symptom driven partly by expectation is still a symptom, felt in a real body, and it still needs to be described out loud to the person prescribing. What the review offers is a different question to bring into the room: not only what could go wrong with this medicine, but how the risks compare, how likely they are, and what is worth watching for versus what is worth letting go. That is a conversation to have with a clinician who knows the case, and it is the only place it belongs.