Decoding disparities: evaluating automatic speech recognition system performance in transcribing Black and White patient verbal communication with nurses in home healthcare

Maryam Zolnoori, Sasha Vergez, Zidu Xu, Elyas Esmaeili, Ali Zolnour, Krystal Anne Briggs, Jihye Kim Scroggins, Seyed Farid Hosseini Ebrahimabad, James M Noble, Maxim Topaz, Suzanne Bakken, Kathryn H Bowles, Ian Spens, Nicole Onorato, Sridevi Sridharan, Margaret V McDonald · 2024

Read Decoding disparities: evaluating automatic speech recognition system performance in transcribing Black and White patient verbal communication with nurses in home healthcare by Maryam Zolnoori, Sasha Vergez, Zidu Xu, Elyas Esmaeili, Ali Zolnour, Krystal Anne Briggs, Jihye Kim Scroggins, Seyed Farid Hosseini Ebrahimabad, James M Noble, Maxim Topaz, Suzanne Bakken, Kathryn H Bowles, Ian Spens, Nicole Onorato, Sridevi Sridharan, Margaret V McDonald (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-11631515.txt.

Description

Home healthcare visits are often recorded, and software that turns speech into text could feed those words into risk models. Maryam Zolnoori and co-authors test whether four such systems transcribe Black and White patients equally well. They use recordings of nurse visits to 35 patients in New York City, 16 Black and 19 White, covering 860 utterances checked against careful human transcripts.

Amazon's general system did best overall, yet its median word error rate was still 39 percent. Every system performed worse for Black patients than for White patients. The gaps showed in word choice measures covering feelings, social talk, and drives, and all systems stumbled on filler words, repeats, and everyday proper names. Amazon's medical system handled medical terms best. The authors urge broader training data and better dialect handling before such transcripts guide care decisions.

Catalog record

CREATORMaryam Zolnoori, Sasha Vergez, Zidu Xu, Elyas Esmaeili, Ali Zolnour, Krystal Anne Briggs, Jihye Kim Scroggins, Seyed Farid Hosseini Ebrahimabad, James M Noble, Maxim Topaz, Suzanne Bakken, Kathryn H Bowles, Ian Spens, Nicole Onorato, Sridevi Sridharan, Margaret V McDonald
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE50 KB
VERSION1.0
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC11631515

Browse the full library · About the Tower of Babel