Accent related errors in clinical speech transcription and a LLM-based remedy
Accurate clinical documentation is essential for safe, effective patient care. AI tools powered by automatic speech recognition can streamline this process. Variable performance across speakers with diverse accents leads to transcription errors and clinical risk. In testing Whisper and WhisperX on native and non-native English clinical speech, error rates were significantly higher for non-native speakers. Post-processing with GPT-4o restored lost accuracy. This chained approach (WhisperX-GPT) reduced accent-related errors.
Authors
- Kevin K. Tsang (ORCID: https://orcid.org/0000-0002-1531-3312)
- Yasaman Fatapour (ORCID: https://orcid.org/0000-0001-8660-1621)
- Ryan C King (ORCID: https://orcid.org/0000-0003-3479-0365)
- Nicha Wongjarupong (ORCID: https://orcid.org/0000-0001-8676-413X)
- Jamil S. Samaan (ORCID: https://orcid.org/0000-0002-6191-2631)
- Yi-Te Lee
- Nadine A. Friedrich (ORCID: https://orcid.org/0000-0002-6713-0614)
- Apoorva Srinivasan
- Nitin Srinivasan (ORCID: https://orcid.org/0009-0001-6472-4441)
- Hongyu Liu (ORCID: https://orcid.org/0009-0002-6718-7976)
- Sarvenaz Fatapour
- Ryan Czarny (ORCID: https://orcid.org/0000-0001-5039-5221)
- Trini Nguyen
- Arash Abiri
- Nicholas P. Tatonetti
- Michael Ziets
- Kim Hyunseok
- Shehan Thangaratnam
- Aditi Kuchi
- Yee Hui Yeo
- Jacob Berkowitz
Institutions
- University of Southern California (US)
- Cedars-Sinai Medical Center (US)
- University of California, Irvine (US)
- University of Washington Medical Center (US)
- Martin Luther King, Jr. Multi-Service Ambulatory Care Center (US)
- Martin Luther King, Jr. Community Hospital (US)
- University of California, Irvine Medical Center (US)
- Chonburi Hospital (TH)
- Columbia University (US)
Publication Details
- Journal
- npj Digital Medicine
- Published
- 2026-03-02
- DOI
- https://doi.org/10.1038/s41746-026-02490-z
- Primary Topic
- Artificial Intelligence in Healthcare and Education
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Cedars-Sinai Medical Center