I'm not sure if this is going to be relevant to you but it popped into my head whilst reading through the comments and in particular your note about speech to text.
It's this cool pad thing where the paper is timecoded. You write on the paper, and draw, with a special pen, and it records sound along with the time code. If you're familiar with Serato Scratch Live it's a similar concept to that - basically taking a "real world" interface and semi-digitising it using some form of time encoding.
Speech to text is HARD and what's worse is inaccuracies in a medical field could result in some really serious consequences (misdiagnosis, improper dosage, incorrect medication being prescribed etc.)
Doctors are pretty used to shorthand, and also my understanding is they quite often have a typist for dictations - perhaps an alternative to speech to text is to record speech by proximity in time to what has been typed. That way they could be saying something, take minimal notes but have a sound file linked to those notes. That way they can elaborate on the notes later if required, but also if they go back to their notes and can't remember what the hell they meant when they wrote "PKR instead", they can touch "PKR instead" to hear what they were saying at that time in the conversation, and fastfoward, rewind to get the context.
Have you ever heard of live scribe?
http://www.livescribe.com/en-us/
It's this cool pad thing where the paper is timecoded. You write on the paper, and draw, with a special pen, and it records sound along with the time code. If you're familiar with Serato Scratch Live it's a similar concept to that - basically taking a "real world" interface and semi-digitising it using some form of time encoding.
Speech to text is HARD and what's worse is inaccuracies in a medical field could result in some really serious consequences (misdiagnosis, improper dosage, incorrect medication being prescribed etc.)
Doctors are pretty used to shorthand, and also my understanding is they quite often have a typist for dictations - perhaps an alternative to speech to text is to record speech by proximity in time to what has been typed. That way they could be saying something, take minimal notes but have a sound file linked to those notes. That way they can elaborate on the notes later if required, but also if they go back to their notes and can't remember what the hell they meant when they wrote "PKR instead", they can touch "PKR instead" to hear what they were saying at that time in the conversation, and fastfoward, rewind to get the context.