My question is, where has everyone been? This issue has been apparent for years to anyone who has ever worked in an intensive care unit. Walk into any ICU and within 30 seconds, you will hear an alarm sound its annoying bleat and see the staff going about their business without a second thought.
As far back as the early 1990s, studies documented the issues. There are too many alarms. In 1994, researchers from Ohio published a study in which they played some 33 different common alarm tones for 100 ICU staff members. Only 50% of them were able to identify the source of each alarm. The authors concluded that there were too many alarms for even experienced staff to identify and proposed that alarms should be used for only critical problems.
Another 1994 study from a pediatric ICU observed alarm activity over a week’s time. They noted that of 2,176 alarm soundings, 68% were false, 26.5% were induced by staff dealing with patients and only 5.5% were actually significant.
An interesting study performed on college students in 1995 showed that people respond to alarms on the basis of their expectations. In other words, if an alarm is known to usually be false, the individual is less likely to respond to it. Obviously, if 95% of alarms in an ICU are false, the staff will be conditioned to ignore them.
There have been many other papers on the subject documenting the above as well as the stress on the staff that the constant level of noise produces.
In 2011, the problem is even worse as more and more devices have related alarms. For example, the push to prevent patient falls has prompted the invention of the bed alarm. Some of these alarms are so sensitive that the mere touching of the bed by staff triggers the alarm.
The solution to the problem of alarm fatigue is not simple. Every new device seems to come with its own random noise and sound level. A concerted effort by device manufacturers, regulatory agencies and hospitals is needed. My suggestion is to 1) scrap all current alarms and develop standardized tones for a select few important events and 2) improve the sensitivity of triggers to reduce the false alarm rate. Good luck.