With increased hours utilising electronic health records, GPs are unable to effectively spend increased time committed to providing exceptional patient care. Google has recently stated in a blog that it intends to transform such documentation in EHR’s by utilising voice recognition technology and medical scribes to support GPs in their everyday roles.
Through the transcription of medical conversations between doctor and patient, satisfaction is gained on both sides, where patients feel listened to and valued, and GPs feel they can deliver on their primary goal.
"In “Speech Recognition for Medical Conversations”, we show that it is possible to build Automatic Speech Recognition (ASR) models for transcribing medical conversations", explained Google in their new blog.
Working with researchers and physicians at Stanford University, Google constructed a system utilising both a Connectionist Temporal Classification (CTC) and a Listen Attend and Spell model (LAS), the company undertook a number of tests surrounding word error rates, background noises and speech patterns which would enable both models to cover 14,000 hours’ worth of speech.
“While most of the current solutions in medical domain focusing on transcribing doctor dictations (i.e., single speaker speech consisting of predictable medical terminology), Google’s research shows that it is possible to build a model which can handle multiple speaker conversations covering everything from weather to complex medical diagnosis,” the blog added.
The study will therefore help further Google’s ambitions to support the medical sector and bring the focus on patient care back to the forefront of the clinical sector.