Ten anaesthetists were observed while providing anaesthesia for two simulated surgical procedures, twice using conventional methods and twice using a prototype of a new drug administration system designed to reduce error. Aspects of each method were rated by users on 10-cm visual analogue scales (10 being best) and comments were invited. Median safety scores were 7.7 cm (range 4.3-8.9) for the new system and 4.6 cm (1.3-8.2) for conventional methods (p = 0.009). The new system was compared favourably with conventional methods in respect of safety (p = 0.005), clinical acceptability (p = 0.008), organisation and layout (p = 0.047), and acceptability for use on patients (p = 0.005). The new system saved time in the preparation of drugs both before anaesthesia (105 vs. 346 s; p < 0.001) and during anaesthesia (20 vs. 104 s; p < 0.001). Comments facilitated development of the system and the evaluation endorsed proceeding to a clinical trial.