Measuring the performance of audible alarms for anaesthesia.

The ergonomic performance of an integrated set of 17 audible alarm sounds, divided into low, medium and high priority classes has been undertaken. The sounds were tested for their ease of learning/recall, and how closely their intrinsic perceived urgency matched to a clinical assessment of urgency. The tests were computer-administered and performed on 21 volunteers aged from 18 to 52, in two sessions a few days apart. Session 1 taught the meanings of the alarm sounds and session 2 measured the performance of the sounds. The mean correct identification rate for the sounds was 48.4% (range 10.3-90.0%) with 97.5% of misidentifications within sound priority class. The urgency correlation was statistically significant (r=0.85, p<0.001) with all priority classes included but within priority class correlations were not statistically significant. Poor within priority class performances were ascribable to a priori aspects of the design of the sound system.