Imagine a clinician handing you a capsule and saying, plainly: there is nothing in this. No drug, no herb, no active ingredient. Some people find it helps anyway. Would you take it? For about fifteen years, small groups of researchers have been asking exactly that question, and getting an answer that refuses to sit still. The field calls it open-label placebo, or OLP: an inert treatment given openly, with the rationale explained rather than hidden. It is the one corner of placebo science that does not depend on deceiving the person in front of you, which is precisely why it has drawn interest from people who care about consent as much as about outcomes.
What it has not had, until now, is a rulebook. On August 10, 2026, PAIN posted online ahead of print a review titled "Methodological considerations for designing and conducting open-label placebo trials", led by Helena Hartmann with Diana Muessgens, Livia Asan, Aleksandrina Skvortsova, Julian Kleine-Borgmann, Katharina Schmidt, Aurore Fernandez, Jules Phalip, Johannes Wessels, Marco Valerio, Meike Shedden-Mora, Ulrike Bingel and Jana L. Aulenkamp, together with the patient advisory board of the transregional Collaborative Research Center (CRC/TRR) 289 "Treatment Expectation". It is a review and guidance paper, not a trial: no new patients, no new results. Its subject is the awkward machinery of studying something you cannot conceal. If everyone knows which group they are in, what counts as a control group? What can still be blinded, and by whom? How should the rationale for a placebo be worded so that it informs without overselling? And what, at minimum, should a paper report so the next team can tell what was actually done? The authors also fold in what patients on that advisory board said about whether they would find such treatments acceptable and feasible at all, which is not a footnote: an open placebo only works as medicine if a person is willing to hold it in their hand.
The timing matters because the empirical picture is genuinely unsettled, and the paper does not pretend otherwise. A systematic review and meta-analysis published in Scientific Reports on July 5, 2025 pooled randomised trials of open-label placebo in chronic musculoskeletal pain and concluded, with low certainty evidence, that it may improve self-reported physical function and reduce pain intensity. The same analysis judged every patient-reported outcome to be at high risk of bias, and found heterogeneity in the physical function measures substantial enough to limit what could be pooled at all. In other words: a signal, reported honestly, in studies whose methods are the weak link. That is the gap a standards paper exists to close.
For families reading this newspaper, the connection is not a treatment. It is trial design, and the reason that is not a dry subject is that dementia trials live or die on their placebo arms. Every disease-modifying candidate is judged against what happens to people who received nothing active, and expectation is part of what happens to those people. There is direct evidence that this component is not fixed across brains. In PAIN in 2006, Fabrizio Benedetti and colleagues in Turin reported, in patients with Alzheimer's disease, the first demonstration that expectation and placebo related analgesic mechanisms can be disrupted by a clinical condition itself, meaning the open giving of a painkiller loses some of the boost it carries in other patients. That finding is about analgesia, and it is twenty years old, but it tells you the variable is real and measurable rather than mystical.
So the honest reading of this week's paper is narrow and useful at the same time. Nobody has shown that an openly labelled placebo helps memory, agitation, apathy or anything else in any dementia. The Hartmann review is about pain research "and beyond", and the beyond is, for now, an invitation rather than a result. What it offers is a shared vocabulary: if a future trial in Lewy body dementia or vascular dementia wants to separate the effect of the ritual from the effect of the molecule, or wants to give a placebo arm something better than nothing without lying to anyone in it, there is now a documented way to argue about how to do it properly.