Fascinating article on Boston.com about medical devices and connectivity. It talks about trying to get all the devices talking to each other and making decisions on the information they receive, which seems to me to be a recipe for disasters bigger than the lede event.
The event in the story's lede [yes, that's spelled right] is a case where they had to turn off a breathing machine to get a quick Xray, and the distracted anaesthesiologist didn't get it back on in time. Perhaps the solution for that is something simple like a "pause 45 seconds" option for the breathing machine, rather than trying to teach every machine to negotiate with all the others?
They mention the fact that to have 10 machines wired directly to each other, you have to run lots of cables. (They estimated 100 cables rather than the correct number, 45). I'd argue against using EM spectrum like bluetooth - too much chance of locking onto the device in the next operating theatre. More likely, you need a hub and device protocol, which would require only one more device and 10 cables. But first, someone has to analyze what are all the conditions that need to be communicated, and design the control programs.
Lots of work there.