[graypegg]

I wonder how much information any 1 medical device with an alert knows though. GPWS has the benefit of being part of 1 system, where (I think, no experience here) hospitals seem to treat equipment as singular items that do 1 thing or are meant for a specific area of responsibility in an operation. Like a vitals monitor might not know what the drug pump is doing.

I could imagine

    ventilation? arrhyth-*C-chord*-ARRHYTHMIA! CHECK PUMP! HEART RATE!

coming from different devices to be pretty distracting.

I think GPWS can set windows of cases where an alert is given. Like, a terrain warning isn't much help when landing. Maybe there's something like that already for medicine, but a device who's job is to consume information from other devices, and only provide alerts based on rules the staff can configure before an operation, could be a thing that's useful.

[HeyLaughingBoy]

That's really the problem many here are describing. 60601 mandates what alarms shall be active, but it spans a single device. If you have, e.g., 10 ventilators in a room that are alarming, you can't silence them all with a single button press.

[graypegg]

I can kind of understand why it ended up that way. There's some benefit to just assuming, nothing works together. It's at least a consistent state of affairs. You can just wheel in any heart rate monitor, and you only need to understand that heart rate monitor.

But it seems like a space that's really ripe for improving. We have very reliable simple protocols you could hook these all into. Imagine it was law that every medical device had to emit the numbers it displays on something like an ODB2 port. Something that can be visually checked to be plugged in, be unplugged and replugged with no handshake, and handle daisy chaining so in the event the "network" breaks in two, or a device goes down, you still get information from the remaining network/it can reroute.

For such a highly regulated industry... you kind of wish they would regulate. I guess status quo is also a regulation.

[HeyLaughingBoy]

Interoperability is definitely a problem. I've been in the industry for quite some time, and at one job we got to shadow med lab techs, since they were the operators of the machines that we built. Their workload is insane and after an hour watching them work, I could identify a number of new products that would help them. I brought it up to my management and learned that "marketing is aware of those problems and we're devising solutions." No idea if those solutions ever hit the market -- this was over 10 years ago.

At the time, the solution to interoperability was to buy all your lab equipment from one manufacturer, who would use their own (usually proprietary) protocols to tie things together. That way, at least even if they weren't actually interoperable, the UI's and workflows were mostly consistent.

A large part of the problem is that hospital IT is understandably hostile to anything connecting to their network, so all the stuff we were building at the time that talked to each other, had to use its own standalone network, or serial ports (ugh!).

Standards like IEC-60601 or 62304 (my daily bread) are easier to adopt because they address patient safety. I suspect it would be much harder to mandate an interoperability standard unless you could show that it improved safety instead of "just" making the healthcare provider's job easier. Or maybe it exists, but just never came up on my radar.

[fhsm]

Inter op has a lot of different standards but support for all those standards is limited. Here’s an example:

https://www.astm.org/f2761-09r13.html

As you point out the corporate network operators tend to have a different set of priorities but even there too standards exist. Here’s an example:

https://www.iso.org/standard/72026.html