Objective: Four sets of eight audible alarms matching the functions specified in IEC 60601-1-8 (2012) were designed using known principles from auditory cognition, with the intention that they would be more recognizable and localizable than those currently specified in the standard. Background:The audible alarms associated with IEC 60601-1-(2012), a global medical device standard, are known to be difficult to learn and retain, and there have been many calls to update them. There are known principles of design and cognition which might form the basis of more readily recognizable alarms. There is also scope for improvement in the localizability of the existing alarms.Method: Four alternative sets of alarms matched to the functions specified in IEC 60601-1-8 (2012) were tested for recognizability and localizability, and compared with the alarms currently specified in the standard Results: With a single exception, all prototype sets of alarms outperformed the current IEC set on both recognizability and localizability. Within the prototype sets, 'auditory icons' were the most easily recognized, but the other sets, using word rhythms and simple acoustic metaphors, were also more easily recognized than the current alarms. With the exception of one set, all prototype sets were also easier to localize. Conclusion:Known auditory cognition and perception principles were successfully applied to a known audible alarm problem Application: This work constitutes the first (benchmarking) phase of replacing the alarms currently specified in the standard. The design principles used for each set demonstrates the Medical audible alarms 3 relative ease with which different alarm types can be recognized and localized.Keywords: audition; auditory displays; learning; medical device technologies Precis: Four sets of audible alarms matched to the functions specified in IEC 60601-1-8, a global medical device standard, were designed using known principles of successful audible alarm design. When tested for recognizability and localizability, all sets (with one exception for localizability) outperformed the current alarms specified in the standard. This work represents the first phase of updating the alarms specified in that standard

The reserved set of audible alarm signals embodied within the global medical device safety standard, IEC 60601-1-8, is known to be problematic and in need of updating. The current alarm signals are not only suboptimal, but there is also little evidence beyond learnability (which is known to be poor) that demonstrates their performance in realistic and representative clinical environments. In this article, we describe the process of first designing and then testing potential replacement audible alarm signals for IEC 60601-1-8, starting with the design of several sets of candidate sounds and initial tests on learnability and localizability, followed by testing in simulated clinical environments. We demonstrate that in all tests, the alarm signals selected for further development significantly outperform the current alarm signals. We describe the process of collecting considerably more data on the performance of the new sounds than exists for the current sounds, which ultimately will be of use to end users. We also reflect on the process and practice of working with the relevant committees and other practical issues beyond the science, which also need constant attention if the alarms we have developed are to be included successfully in an updated version of the standard.

No abstract

IEC 60601-1-8 is a medical alarms standard concerned with equipment safety. Part of its content specifies a set of audible alarms for eight clinical functions: cardiovascular, drug administration, oxygen, artificial perfusion, power down, temperature, ventilation and a general sound, along with a medium- and a low-priority alarm. It has long been acknowledged that the alarms specified are difficult to learn and retain , (Atyeo and Sanderson 2015; Edworthy, Page, et al, 2014; Lacherez, Seah & Sanderson, 2007; Sanderson, Wee & Lacherez 2006; Wee and Sanderson 2008). The alarm sounds require updating. The project reported here is concerned with developing candidate sets of alarms aimed at replacing those currently specified in IEC 60601-1-8. Four prototype sets of alarms were designed for the eight functions specified in the standard. These were as follows: Word rhythms: Each alarm contains a large number of harmonics and the eight alarms vary in pitch pattern, number of pulses and rhythm where those patterns mimic the words of the functions (so for example the ‘Cardiovascular’ alarm possesses 6 pulses and contains stresses on the first and fourth pulses). Auditory icons: These are real-world sounds that are used as metaphors for the functions. So for example the cardiovascular alarm is a heartbeat-like sound, the “Oxygenation” sound is an aerosol sound and so on. Auditory icons with beacon: Same as the Auditory icons but possesses a short beacon at the start of the sound in order to signify that it is an alarm sound. Resilient: These sounds are simple metaphors (such as a rising pitch for temperature) created at a sampling rate of 8kHz, intended for use in equipment with poor acoustic reproduction. A fifth set, IEC, is the set currently specified in the standard and was also tested for comparison. Participants were tested on a single set of alarms. Once they had been told the meanings of the eight alarms, they were presented with each of them ten times in a block-randomised fashion, and were required to name the alarm. Analysis of variance on the correct/incorrect scores considered the effect of sound type (IEC, word rhythm, auditory icons, auditory icons plus beacon, resilient), block (blocks 1 to 10) and sound (the individual alarm sounds) on the number of correct responses. Each of these factors showed significant effects, as well as some interactions. The largest effect by a considerable margin was attributable to sound type. Performance was significantly lower for the IEC condition than any of the other conditions; performance for the word rhythm condition was significantly higher than for the IEC condition but significantly lower than the other three conditions; performance in the resilient condition was significantly higher than for both the IEC and the word rhythm conditions but significantly lower than in the auditory icons conditions; and performance in the auditory icons and the auditory icons plus beacon conditions was significantly higher than in all other conditions, and similar to one another.