4 Comments
Feb 16Liked by Kunle

Fascinating, thanks for sharing! I'm personally amenable to audio because I can't write down the details I need as a patient and also think of questions I should ask. Having that transcript is helpful and a less distracted clinician would be very helpful!

One downside I can see here is that the docs / insurers are (hopefully) counter positioned. I want my doc to be like my lawyer and break down whatever auth barriers are between me and successfully executing their treatment plan. In my experience, docs may misrepresent reality a bit to get a test they felt was relevant paid for, but which an insurer following care guidelines views as wasteful.

If audio is a default and insurers get their hands on the audio, won't the docs and patients self-censor? How much will be said to trigger some claims authorization rule, rather than a discussion of symptoms and objective reality?

Very excited by the potential but a bit wary given my experience with chronic disease and high deductible, auth heavy plans.

Thanks again!

Expand full comment
author

So few thoughts

1. I dont think payers will get audio by default. It's too much + not necessary. But they might request it in place of requesting a medical record (Which happens a nontrivial amount of times). Self censoring obviates a lot of the documentation advantage from AI scribes. My intuition at the moment is that a) AI scribes will eventually evolve into copilots that nudge providers towards the standards of care (which is ultimately better for providers and patients) and b) Payors will use AI tools to review any content ingested from AI scribes and it won't materially change the financial landscape/tilt the balance in any way.

2. The real advantage of payors getting access (to the extent they do) is that they have access to patterns not available to the provider (eg the context from two visits where something is referenced twice but not picked up by either provider because it appears in isolation).

just my 2c

Expand full comment

Gotcha. Thanks again for the thought provoking content. So much potential here.

A project you may be interested in following...

https://www.med.upenn.edu/observer/reduce.html

Expand full comment
Feb 13·edited Feb 13Liked by Kunle

Kunle this is the best so far. I think your point on adoption is particularly strong if insurance providers discover their malpractice suit risk is reduced by having this additional audio evidence, or if premiums go down if you have audio because insurance providers can sell raw audio at scale or audio insights to LLMs or other scale data users. If you had cough and respiratory data across america in audio many drug companies would love to know how that worked against different drugs which people took over those periods. Its a totally new vector of information and a true signal that is hard to falsify at scale. Well done. Niyi

Expand full comment