47 minutes back per shift.
One more patient seen.
The largest published evaluation of ambient voice technology in an NHS emergency department. Mahdi et al., 4,664 encounters, 24 clinicians, six months — independently modelled by York Health Economics Consortium.
What the trial measured.
Three pre-specified KPIs: patients seen per shift, time to first documentation, time from clinician assignment to clinical readiness.
Six months of live use across 24 ED clinicians (Nov 2024 – Apr 2025) at one of London's largest teaching-hospital emergency departments — the largest published single-site ED dataset for ambient voice technology in the NHS.
York Health Economics Consortium (Ciara Buckley and Nick Hex) modelled the financial impact: £270.93 of added capacity per day per additional patient seen per clinician per shift.
Modelled £1.44m in documentation savings plus £5.36m in capacity gains across 90 in-scope staff at St George's, if the gains are sustained at scale.
If scaled to all 11,055 A&E doctors in England, the modelling implies 9,259 additional A&E attendances per day and £834m in combined annual benefit — at 80% redeployment of saved time.
In such a fast-paced, high-pressured environment, every second counts. This technology allows us to be more efficient, cut down on admin, and ultimately focus on patient care.
It has taken a huge administrative burden away and allowed me to complete notes accurately without struggling to recall — especially when at peak fatigue. I hope it can be a permanent fixture in the ED.
How the trial was run.
NHS-led service evaluation. Pre-post design within the same department. Three pre-registered KPIs, statistical inference, independent economic modelling.
- 01SiteLondon teaching ED
St George's University Hospitals NHS Foundation Trust Emergency Department, Tooting — the only ED in the GOSH DRIVE Phase 4 evaluation and the largest non-core site by AVT volume.
- 02SampleNov 2024 – Apr 2025
24 ED clinicians, 4,664 patient encounters analysed against a three-month pre-implementation baseline.
- 03KPIsPre-registered
(1) Patients seen per shift. (2) Time to first documentation. (3) Time from clinician assignment to patient clinical readiness. All three improved with statistical significance.
- 04Economic modellingIndependent
York Health Economics Consortium (YHEC) translated the operational gains into cashable financial benefit at trust and national scale.
- 05GovernanceClass I device
MHRA-registered Class I medical device. DPIA approved by GOSH IGSG. DCB 0129 and DCB 0160 clinical-safety sign-off complete. Registered as clinical evaluation study no. 3650. Funded by NHS England (London) Frontline Digitisation; Tortus AI provided free of charge.
What the authors flag.
Single-site service evaluation, not a randomised controlled trial. Pre-post design carries the usual confounding risks. The preprint is published via SSRN / Preprints with The Lancet and is not yet peer-reviewed. Tortus is a collaborator on the work. National-scale projections are modelled, not observed.
Bring us your hardest department.
We've published the ED data. Let's design the same study around your service.