InsightaaS: In the spirit of full disclosure, I’ve known John Morris since we worked together in 1988, and have admired and respected his perspicacity throughout the ensuing 26 years. The post highlighted here provides a great example of the insightful and balanced approach that John brings to analysis of complex IT issues, such as Internet of Things (IoT). In it, Morris takes examples of “alarm fatigue” from the healthcare industry to speculate on a possible, but not often acknowledged, downside associated with IoT in manufacturing.
The alarms highlighted in the healthcare example are triggered by monitoring equipment. A signal that might be helpful in isolation – which was likely a key selling point for a unit monitoring heart rate, blood pressure, or other vital signs – becomes lost in an avalanche of hundreds of similar signals from nearby devices. The professionals can’t respond to each beep/light/message, and eventually learn to tune out these alarms altogether, which, as Morris notes, “can lead to mission failure, even death.” Morris then links this experience to a possible outcome in manufacturing, where the extensive sensor network promised by IoT might also create an overwhelming cascade of signals, numbing rather than informing operators and engineers. There are already industrial examples of sensor-triggered alarms providing ‘too much of a good thing;’ Morris states that “the Deepwater Horizon oil rig disaster in the Gulf of Mexico (2010) in part was associated with ignored alarms.” He urges manufacturing firms that are embracing IoT to minimize alarm fatigue by investing in the process understanding needed to connect alarm systems with overall governance objectives. In the end, Morris finds, “Alarm fatigue doesn’t manage itself; by itself alarm fatigue will just get worse, by entropy. And when alarm fatigue gets worse, the documented result is that humans can’t do their work effectively.”
‘350 alarm conditions per bed, per day’ was a typical healthcare statistic recorded by Johns Hopkins Hospital researchers, according to a report in 2012. And in one intensive care unit, the number was 771, again per bed, per day.
You don’t have to be a medical professional to imagine that nursing and physician staff members cannot possibly process such a volume of alerts. Especially with the acceleration of technology adoption in healthcare over the past two decades, nurses and hospital staff have increasingly been afflicted with ’alarm fatigue‘, to the point that important medical information is ignored and patient care suffers.
Let’s add a little bit more to our understanding of the phenomenon of alarm fatigue itself. Alarm fatigue is a natural human response when a large number of poorly organized devices (in the example here, medical devices) each generate its own characteristic alert in the form of beeps or other sounds, flashing lights, text messages and more. And rather than support humans in important work, here of healthcare, this well-intentioned but flawed volume of alerts overwhelm human sensory and information processing capacity — and alarm fatigue sets in. The situation can lead to mission failure, even death…