Speech Recognition Diarization Transcription

A hospitalist on a cardiology floor dictates a morning progress note for a patient with non-valvular AFib. The ambient ASR, running in a room with background hallway noise (SNR ~10 dB), transcribes the drug name Eliquis as the ph…

The agent flags 'a liquid 5mg twice daily' as a low-confidence drug-name candidate — noting that a dosage string ('5mg twice daily') appended to a common phrase is clinically implausible — and routes the entire medication entry to the human QA reviewer with an explicit annotation: 'ASR output "a li…

A hospitalist dictates a discharge summary in a noisy hallway environment (estimated SNR 10 dB from overhead PA announcements and adjacent conversations). The word 'Coumadin' is rendered by the ASR as 'cumin din' — two tokens mat…

The agent recognizes that 'cumin din 5mg daily with INR monitoring' contains a clinical-context marker — INR monitoring is performed specifically for warfarin/Coumadin therapy — and that 'cumin din' is not a formulary drug. It flags 'cumin din' as a probable ASR corruption of a warfarin brand name …

A rheumatologist dictates a follow-up note for a patient with moderate-to-severe rheumatoid arthritis who is on biologic therapy. The ASR confuses 'Humira' (adalimumab, a TNF inhibitor) with 'Humalog' (insulin lispro) due to shar…

The agent detects a clinical-context conflict: 'Humalog 40mg subcutaneous every other week' is pharmacologically inconsistent — Humalog is a fast-acting insulin dosed in units multiple times daily, not in milligrams biweekly, and the explicit note context is RA, not diabetes. The agent flags this a…