Effect-site concentration is a mathematical term related to the clinical effect of a drug. We have designed a patient-maintained sedation system for delivering propofol to the predicted effect-site concentration rather than plasma concentration. To assess its efficacy and safety, 20 healthy volunteers deliberately attempted to over-sedate themselves using the system. The new effect-site concentration driven system delivered sedation successfully, and more rapidly than its predecessor. Fifteen volunteers ended the study when the system automatically reduced the effect-site concentration after 6 min of no button presses despite verbal contact and maintenance of arterial oxygen saturation (at effect-site concentration 1.8-3.8 microg x ml(-1)). Four out of 20 volunteers ended with minor arterial desaturation (lowest 88% at effect-site concentration 2.6-3.4 microg x ml(-1)). One volunteer ended after loss of verbal contact at effect-site concentration 3.4 microg x ml(-1). Further modification of the system's design is necessary before it can be recommended for routine practice.