Problems With Auditory Alarms in Anesthesia and Tests of a Proposed Solution: Multimodal Multitask Performance With an Auditory Display

Auditory alarms are used pervasively for patient-monitoring equipment to warn of potential problems, despite widely reported shortcomings in their design. Problems include high rates of false alarms; interruptive, uninformative, stress-inducing acoustic profiles for alarm sounds; and alarms that are acoustically indistinguishable from one-another. The current work examined the alarm problem in the domain of medical monitoring, both through field studies and controlled laboratory simulations. Through field studies, anesthesia careproviders' information gathering during airway management was recorded using an ambulatory eyetracker, and their visual scanning behavior analyzed. A model for operators' monitoring behavior was developed that included effects of alarms, workload, monitored-parameter bandwidth, memory limitations, information-access cost, control-actions taken, correlations among parameters, and task relevance. To test the model in the field, alarm thresholds (sensitivity levels) were manipulated on patient monitoring devices, and the effects on operator performance examined. An auditory display was proposed as a potential solution to the auditory-alarm problem. Auditory displays can communicate information redundant to or independent of visual displays, potentially offloading some visual workload to the auditory channel. The laboratory simulation study examined the effect of visual, auditory and redundant displays on the performance of a dual-task simulation of patient monitoring. A psychophysical pilot study examined the parameter levels of the auditory display necessary to provide information equivalent to the visual display. In the main experiment, subjects performed manual compensatory tracking task (simulating the visual demands of direct patient care) while monitoring six vital signs of a simulated patient, detecting deviations from normal levels. Monitoring was presented in three display conditions: auditory only, visual only, and redundant. Results indicate that the detection times of deviations in visual and redundant conditions were not significantly different, but were faster than with the auditory display. However, performance in the tracking task was degraded least in the auditory condition, and most in the redundant display condition-an example of a negative redundancy-gain. Reasons for this finding are examined through data from eye-movement recordings. Eye-data supported hypotheses regarding information access cost, effect of alarms, monitored-parameter bandwidth, and also revealed reduced visual scanning in the redundant display compared to the visual-only display.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Psychology

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2015

Author dc:creator

• Seagull, Frank Jacob

Contributors dc:contributor

• Wickens, Christopher D.

• Engineering, Biomedical

Language dc:language

eng

Identifier

(MiAaPQ)AAI3023191

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/82012