Toward Automated Clinical Transcriptions

Abstract

Administrative documentation is a major driver of rising healthcare costs and is linked to adverse outcomes, including physician burnout and diminished quality of care. This paper introduces a secure system that applies recent advancements in speech-to-text transcription and speaker-labeling (diarization) to patient-provider conversations. This system is optimized to produce accurate transcriptions and highlight potential errors to promote rapid human verification, further reducing the necessary manual effort. Applied to over 40 hours of simulated conversations, this system offers a promising foundation for automating clinical transcriptions.

Figure 1

Diagram of the full system showing how each individual component interacts

Figure 2

Example showing calculation of transcription-diarization overlap to predict the speaker

Figure 3

A graph showing the distributions of Word Error Rates

Figure 4

A graph showing the distributions of mislabeled speakers

Figure 5

Pie charts detailing the percentages of words in the original text and transcribed text

Figure 6

Bar graph showing the breakdown of incorrect words in the transcription and the percentage of words with the speaker mislabeled, separated by domain