Importance: General practitioners (GPs) sometimes initiate a treatment despite not expecting it to improve patients' symptoms by any physiological mechanism. These essentially placebo treatments are ethically controversial, and their frequency is unclear. They involve risks for patients, but to estimate these, more data are needed. Objective: To develop a more precise overview of the rate at which GPs prescribe essentially placebo treatments. Design, Setting, and Participants: This cross-sectional survey study included currently practicing GPs from 20 European countries and Israel who responded to online questionnaires. The online questionnaires were taken between December 12, 2019, and August 4, 2021, and analyzed on April 28, 2022. Respondents were contacted by national representatives, either through personal networks (convenience sampling) or an existing database (volunteer sampling). Main Outcomes and Measures: The main outcome was the rate of essentially placebo prescriptions, given as the rate per week and the proportion of consultations. Secondary outcomes were the associations between this rate and GP background characteristics (gender, age, education about placebos, years of experience, patients seen, and working hours per week). Results: A total of 952 practicing GPs responded (453 of 745 [61%] female; mean [SD] age of 48.02 [11.95] years), and 669 answered all questions. Overall, 689 of 818 respondents (84%) indicated they had prescribed an essentially placebo treatment at least once. Overall, the median (IQR) rate of essentially placebo prescriptions was 0.5 (0.1 to 2.0) per week or 0.67% (0.06% to 2.50%) of consultations. The prescription rate was higher in men (β = 1.94 [95% CI, 0.58 to 3.29]; P = .005), those with more work experience (β = 0.12 [95% CI, 0.06 to 0.18]; P < .001), and those who work fewer hours per week (β = -0.08 [95% CI, -0.13 to -0.03]; P = .001). Conclusions and Relevance: In this survey study of GPs across 21 countries, essentially placebo prescriptions featured in a small minority of consultations, but they nevertheless occurred regularly for most GPs. Rates varied only slightly by GP background characteristics. This suggests that essentially placebo prescriptions were common at a population level, which poses risks for the patient-GP relationship and creates medical risks for patients. Future research should further investigate the decision-making process behind these prescriptions and their effects on patients.

Background: The hormone cortisol plays important roles in human circadian and stress physiology and is an interesting target for interventions. Cortisol varies not only in response to stress but also as part of a diurnal rhythm. It shows a particularly sharp increase immediately after awakening, the cortisol awakening response (CAR). Cortisol can be affected by medication, but it is less clear whether it can also be affected by learning. Animal studies have consistently shown that cortisol can be affected by pharmacological conditioning, but the results in humans have been mixed. Other studies have suggested that conditioning is also possible during sleep and that the diurnal rhythm can be conditioned, but these findings have not yet been applied to cortisol conditioning. Objective: The objective of our study was to introduce a novel avenue for conditioning cortisol: by using the CAR as an unconditioned response and using scent conditioning while the participant is asleep. This study investigates an innovative way to study the effects of conditioning on cortisol and the diurnal rhythm, using a variety of devices and measures to make measurement possible at a distance and at unusual moments. Methods: The study protocol takes 2 weeks and is performed from the participant’s home. Measures in week 1 are taken to reflect the CAR and waking under baseline conditions. For the first 3 nights of week 2, participants are exposed to a scent from 30 minutes before awakening until their normal time of awakening to allow the scent to become associated with the CAR. On the final night, participants are forced to wake 4 hours earlier, when cortisol levels are normally low, and either the same (conditioned group) or a different (control group) scent is presented half an hour before this new time. This allows us to test whether cortisol levels are higher after the same scent is presented. The primary outcome is the CAR, assessed by saliva cortisol levels, 0, 15, 30, and 45 minutes after awakening. The secondary outcomes are heart rate variability, actigraphy measures taken during sleep, and self-reported mood after awakening. To perform manipulations and measurements, this study uses wearable devices, 2 smartphone apps, web-based questionnaires, and a programmed scent device. Results: We completed data collection as of December 24, 2021. Conclusions: This study can provide new insights into learning effects on cortisol and the diurnal rhythm. If the procedure does affect the CAR and associated measures, it also has potential clinical implications in the treatment of sleep and stress disorders.