Exploring the Impact of Ambient Voice Technology in Primary Care
As part of our mission to evaluate emerging digital tools in real-world clinical contexts, the CoDE has completed its first major programme of work: a comprehensive assessment of Ambient Voice Technology (AVT) or Clinical AI Scribes (CAIS) in general practice.
This project was driven by a clear problem statement from the primary care workforce:
“Inputting information is time-consuming and inefficient.”
To address this, CoDE designed and executed a series of controlled experiments in our simulated GP environment, using synthetic patient data and real consultation scenarios. The aim was to understand how AVT tools—designed to transcribe and summarise clinical conversations – perform under realistic conditions.
What We Did
- 50 consultations were conducted using validated scenarios.
- AVT-generated summaries were compared with traditional note-taking and clinician-reviewed transcripts.
- We tested multiple AVT platforms, configurations, and environmental conditions (e.g. background noise, accents, speech impediments).
- Both manual and automated methods were used to assess accuracy, including advanced natural language processing techniques.
What We Found
- AVT improved the fluidity of consultations, reduced administrative time, and enhanced clinician-patient engagement.
- However, accuracy varied depending on factors like microphone placement, background noise, and patient speech characteristics.
- Omissions were the most common type of error across all platforms.
- AVT use could reduce session overruns by up to 90% and save GPs between 8 and 40 minutes per session.
Why It Matters
This work provides a robust, evidence-based foundation for NHS organisations considering AVT adoption. It also demonstrates CoDE’s unique capability to test digital health tools in a safe, reproducible, and clinically informed environment.
The full report includes a hazard log, ethical considerations, and recommendations for safe implementation. It is intended as a guide – not a decision-making document – and should be read alongside national NHS guidance.