I agree but someone (?) needs to get that app out and certified and known ASAP because the ones already in the market
If you want to give your health care provider a small panic attack, just mention the word “migration” as is “migrate to a new computer system”. Nobody wants to do it. Once something “works” (using the broadest definition of the term) it will be kept as long as possible.
they are very good at being human. and using the human side of tech, like social media, chats, zooms.
they are not good at being technicians. or selecting someone else to be a technician on their behalf.
There are models for this that are better in theory, they require you to type rough notes which are refined which sidesteps the gigantic issue of transmitting such sensitive audio.
I have never seen anything that starts with typing?
Also for some reason many of the companies offering systems that audio record sessions, transcribe, and summarize (over just summarizing text notes) are Israeli and that is sketch to me.
have also noticed it. I have on my to do list to look into these. hmu if you find anything. fuck knows what the training data is for. We have campaigns to divest HOOP from israel and I think AI should be included.
making it about individual choice is an error. 98% of people agree because the person who is asking permission has so much power. we need regulations.
I work in a place that is using these. The reason they are sending it out is that there is a very low level of comfort with tech and the outside company is “taking care of” a lot of messiness. The fact of the matter is there is no in-house capability for this, You should have seen us trying to replace the toner (or was it the drum?) in the fax machine yesterday. The first company with a “nice” UI will have market dominance. Not for technical, but for social reasons.
The people I work with don’t relay want an exact transcription, they like how it formats it into a SOAP or other standard way.
They are “supposed” to review each and every note but I see things in the notes I know are incorrect like for example it gets the gender wrong or there are health problems raised which I know the person does not have. Many other times I see things which may be wrong but I can’t for sure tell. It could have happened that way. But it’s a bit weird.
A responsible design trick might actually be to hide intentional errors in the provided text and then have the charts audited at a later date (by a human) to see if those were left in.
I have warned the people I work with that when the VC makes them go public, all the “agreements” made about confidentiality of data will be out the window, to say nothing of the so-far acceptable cost of services.
My hypothesis is that this is a way of getting training data. The business is not really doing ai scribe, it’s convincing HCPs to make patients comfortable allowing ML on their visits. Especially in the set ups where the vendor can then review the note the clinician actually creates and digest that as well. Like how google is a search engine but it’s really a business to profile people for targeted ads.
Also I agree with you that making notes is not purely an administrative afterthought. Documentation is part of the cognitive work. That’s why there are so many forms, tools, scales, flowcharts, acronyms, standards, templates, and such. It isn’t just about recording what happened. It is about structuring your thoughts prior, during and after the visit. And being able to compare them over time.