Patients and methods. Three hundred patients admitted to a rheumatology department (mean age, 49.2 ± 15.5 years) and 100 nurses working in the same teaching hospital (mean age, 36 ± 8.6 years) completed a questionnaire on the placebo effect in the treatment of pain. Results. The patients believed that the percentages of subjects who responded occasionally or consistently to a placebo were 32% (± 22%) and 21% (± 17%), respectively. The figures given by the nurses were only slightly higher: 42% (± 23%) and 23% (± 17%), respectively. Only 27% of the patients and 58% of the nurses knew that pain could resolve completely under placebo therapy. The mean percentage improvement induced by placebo therapy as compared to the study analgesic was estimated at 21% (± 17%) by the patients and 30% (± 18%) by the nurses. Only 23% of patients and 24% of nurses knew about nocebo effects. Conclusion. These data suggest that during double-blind placebo-controlled trials three-quarters of the placebo arm patients who experience a marked improvement or a nocebo effect may believe they received the active drug. This is likely to reduce the difference between the placebo and active drug groups. Furthermore, most patients and nurses have a negative perception of placebo therapy. It may be useful to include a brief description of placebo and nocebo effects into the patient's information and informed consent documents used in double-blind placebo-controlled trials.