EBME Forums Forums Maintenance Technical Assistance Alarms on Medical Devices

The technical and human factors related aspects of clinical alarms on medical devices are of particular interest to me. Did anyone get to the recent IET seminar concerning alarms?

If anyone knows where I can get some notes or a transcript of what was discussed, on the IET website, for example, or could mail me some details of what was discussed (or about future meetings or networks, etc) then I'd be grateful.

Unfortunately I had commitments elsewhere so I couldn't book a place - is any organisation looking for contributors to a special interest group or network concerning clinical alarms?

I'd be interested in getting more involved since I've done a fair bit of research into the techncial and human factors aspects of alarms as part of my academic studies in the not too distant past.

I seem to recall some work done (it could have been by the US Air Force) on standardizing alarms. They were talking about audible alarms on patient monitors, if I remember correctly. Yes, this is an interesting area, but perhaps we should take a cue from other areas of technical endeavour (the auto industry, for instance … consider the speaking Sat-Nav!).

Equipment controls are other consideration, are they not? Especially now that many bits of kit are software-driven, involving the use of menus to drill-down to what you are looking for. Some that I have come across have been well-designed, and intuitive, whilst others have been, well … horrendous! Perhaps, with all it's supposed clout, the UK's National Health Service could devise, and then insist upon, it's own alarm-set(s) for critical patient monitoring? Now, there's a thought!

It's not easy to impose standardization here, though, is it? Just look at what’s happening to one universally applied (almost) interface standard! These things evolve, and there have been many cases where the best, most elegant, approach has not been the one taken up. Usually it has been a case of the “industry standard” becoming the one adopted by the de facto market leader. Remember when British motor-bike manufacturers had to start putting the gear-change on the wrong side?

Standardisation seems to work quite well in aircraft, anaesthetics and nuclear power stations - ever thought why? I'm afraid the De-facto standard in safety critical systems is not necessarily the best for everyone but operators will tend to use what works well and is safe and can at least be trained to recognise standard alarms, even if they don't "like" using them.

A lot of research has gone into each of these areas and a lot of changes have been made, for the better, based on the human factors requirements in high-workload, safety-critical, situations such as this. I'm afraid the legislation and standards are with us Geoff, despite not being perfect, but they're much better than nothing as a guideline to manufacturers of medical devices.

Loads of stuff available on the internet about this, and related, topics. Happy surfing, Richard!

Been there done that (for 3 years or so). The best sources are restricted access, e.g. academic references, university libraries and NHS access (Athens).

I would suggest that audible alarms should be made to fall within the frequency range of 400 - 2000 Hz as there are a number of deaf people like myself who can't hear much of the range beyond 2000Hz. I have to use a scope and microphone to check the presence of a tone during PPM with several items of equipment. Now what about deaf clinical staff??

One of the problems with alarm sound(s) is that they first need to be heard, whatever the orientation of carer to patient is, then they need to draw attention to the physical location of the alarming device (by audible and visual means, but initially audible), ideally convey a sense of urgency (perceived urgency to the carer) associated with the alarm (that's related to the situational urgency in the current clincial context i.e. the condition of the patient), so that the response to the alarm is both timely and appropriate.

Ideally this means that alarm sounds should convey some information about the physiological source of the alarm, i.e. the parameter or set of parameters that generates the alarm(s) hence the status of the patient. They should not produce clinically insignificant or "false alarms", neither should they miss clinically significant alarms or "true positives". Unfortunately setting alarm limits is still a bit hit and miss because it depends upon clinical context and the ever-changing status of the patient connected to devices. Sounds are not the only problem associated with current alarm technology.

Sounds & visual indicators allow carers to locate and respond to a clinically significant alarms that may be occurring simultaneously......it's difficult enough for individuals with normal hearing characteristics who work in busy cinical areas with a wide range of devices.

Both visual and audible alarms are absolutely necessary - audible to draw the operators' attention to the alarm source, wherever they are, within earshot, and visually to locate the alarm(s) once attention has been drawn to them. Standardised alarms can be designed to address these needs but there's an element of "natural" response to carefuly designed alarm sounds , that convey important information, and "learned" responses to additional information conveyed in the alarms.

What I'm getting at is that it's difficult enough getting to grips with what's required for individuals with normal hearing - thus it's even more awkward designing for individuals with different hearing characteristics that are outside the norms. It must be extremely difficult to cope in a critical care area where there are lots of alarms. I guess there's more reliance on colleagues, to some extent, and the visual cues.