I'd take a guess that the prototypes are Bluetooth (because it's easy enough to work with for R&D) but a production device would almost certainly use a private RF band. That said, there are already Bluetooth-enabled pacemakers, so who knows.
The RF SOCs I've used, they could be easily reprogrammed to use 802.15.4 or Zigbee or proprietary protocols.
Even if it isn't the FW and the communication for an RF link is the least complicated part of a project like this.
Well, either way I'd hope they'd use a dedicated band and not 2.4GHz ISM.
I often listen to podcasts while making breakfast and the Alexa-phone bluetooth link drops out almost always when using the microwave.
Imagine not being able to move because of RF interference - could be very serious. I'd actually say two independent bands should be used to maximise the chance of transmission in crowded RF environments, or at least some kind of wide band frequency hopping.
Edit: in fact, thinking about this, it's the perfect application for something like UWB. Very short range transmission, very wide bandwidth with frequency hopping. Already used for things like intra-soldier wireless.