Specialized Modes Raise Medical Transcription Accuracy (Full Transcript)
[00:00:05] Speaker 1: Hey Mart.
[00:00:06] Speaker 2: Hey Brittany what's up?
[00:00:08] Speaker 1: All good, what about you?
[00:00:10] Speaker 2: I'm good, thanks for joining the call today. Congrats on the big healthcare launch. So first up, honestly who is Medical Mode built for? What customers did you have in mind while designing the product?
[00:00:23] Speaker 1: Yeah, it's really for anyone who deals with any kind of medical terminology. If you're a front office ad, medical health center and you answer phone calls and speak with patients, or medical scribes, so like doctors who have had a visit with a patient and basically need to transcribe the session for documentation afterwards. Those are kind of the use cases that we're trying to target with this.
[00:00:50] Speaker 2: And what kind of problems were the customers facing that made Medical Mode in particular worth building?
[00:00:56] Speaker 1: Yeah, so the thing that healthcare providers were finding is that when it comes to transcription, even if like 90% of the transcript is correct, there's still those key important details that need to be correct in order for the transcript to be usable. And so those examples are like dosage, medication or drugs, both generic or branded drugs, procedures, symptoms.
[00:01:26] Speaker 2: Yeah, thanks for sharing that, Brittany. What about Vet Mode? What kind of problems were customers facing there and was it like different from Medical Mode?
[00:01:34] Speaker 1: Yeah, great question. The general speech models is mostly trained on human data, human-related data. And so when it comes to veterinarian terms, when it comes to veterinarian terms, a lot of the times it would actually hear human symptoms or human medication or drugs instead of the vet ones. One pair is ibuprofen and caroprofen. And caroprofen is a safe drug for animals, whereas ibuprofen is actually toxic to animals. And so Vet Mode helps correct for those.
[00:02:07] Speaker 2: How should a team decide if they should use Vet Mode or Medical Mode?
[00:02:10] Speaker 1: Yeah, honestly, testing it out on their own data. But I would recommend testing it out on the base model to see where the delta lies and then trying out Vet Mode and Medical Mode to see what the boost on their accuracy is.
[00:02:26] Speaker 2: Well, sounds super exciting. Thanks for your time, Brittany. We just released Medical Mode and Vet Mode. You can check it out in the description. Give it a try and let us know how it works for you. All right, bye.
- Medical Mode supports healthcare personnel and medical scribes who work with specialized terminology.
- Accurate transcription of dosages, drug names, procedures, and symptoms is essential for usable clinical documentation.
- Vet Mode reduces errors caused when general models confuse animal-health terminology with human medical terms.
- A confusion between carprofen and ibuprofen can have serious safety implications because ibuprofen is toxic to animals.
- Teams should compare the base model, Medical Mode, and Vet Mode using their own real-world audio and data.