About Nelson Medical Practice

Nelson Medical Practice is a large NHS GP practice in South West London, with around 25 GPs alongside clinical pharmacists, a physician associate and allied health professionals. The practice takes clinical governance seriously: seriously enough that note quality was a formal risk register item before any AI tool entered the picture.

Note quality was on the practice's risk register

Like most GP practices, Nelson relied on clinicians typing notes during or after consultations. It worked, but documentation quality varied, and the variation was significant enough to be logged as a formal risk: an acknowledged, documented concern that something needed to change.

Dr Peter Hughes, GP Partner and Clinical Safety Officer, understood the stakes better than most. A stroke had left him with the use of one hand for typing, turning documentation into a daily limitation across a full clinical workload. He had also noticed something subtler: typing during a consultation splits the clinician's attention. The patient is speaking, but the GP's eyes are on the screen.

The risk register item, combined with his own experience, made the decision clear. It was time to find a better way, and to find it properly.

The practice ran its own evaluation, no vendor involved

Dr Hughes didn't take any vendor's word for it. He designed and ran a structured, independent evaluation of four AI scribes in live consultations, collected pre- and post-consultation surveys, measured documentation time and assessed note quality against clinical standards. He then led the onboarding himself, training every clinician in the practice.

The findings pointed one way. Accurx Scribe, powered by Tandem, came out ahead, and the practice adopted it.

Documentation fell from 4.1 to 1.7 minutes per consultation

Before the evaluation, documentation took an average of 4.1 minutes per consultation. With Accurx Scribe, powered by Tandem, that fell to 1.7 minutes, a 58 per cent reduction. Clinician satisfaction with note quality rose from 3.33 to 3.92 out of 5 in the practice's surveys. When asked whether they wanted to continue after the trial, 13 out of 13 clinicians said yes.

Dr Hughes then went a step further. As Clinical Safety Officer, he audited 20 post-deployment consultations: every one was rated good or very good.

For Dr Hughes personally, the change went beyond efficiency. As he described it:

"AI scribe has been transformational for me following my stroke. Because I only have the use of one hand for typing, it has made a real practical difference to my day-to-day clinical work and has also significantly reduced cognitive burden. We have also undertaken post-deployment audits and noted improved note quality more broadly across the Practice."

There was a quieter benefit too. With documentation handled by the scribe, clinicians could give patients their full attention during the consultation, and across the entire evaluation period not a single patient declined consent. Today, more than 20 clinicians at Nelson use Accurx Scribe as a standard part of their day.

Why trust decided it

Clinicians who have tried several AI scribes tend to say the deciding factor is trust: in the record the tool produces, and in how patient data is handled. Accurx Scribe is powered by Tandem, built for European healthcare from the start. It integrates directly with EMIS Web and more than 100 other medical record systems, is CE marked under the European Medical Device Regulation, and patient data is never used to train AI models.

For a Clinical Safety Officer who had put note quality on the practice risk register, designed the evaluation himself and then personally audited the results, those were not small details. As Dr Hughes put it:

"It's an amazing product. It really is. Thank you very much."

How these numbers were measured

The documentation-time and satisfaction figures come from the practice's own independent evaluation: pre- and post-consultation surveys and timed documentation across live consultations, designed and run by Dr Peter Hughes as Clinical Safety Officer. The satisfaction score of 3.33 to 3.92 out of 5 is clinician self-reported. The audit result comes from the practice's post-deployment clinical safety audit of 20 consultations. The figures are the practice's own findings, not a Tandem-run study.

See how it fits your practice

Accurx Scribe, powered by Tandem, is live in GP practices across the UK. Talk to our team about what an evaluation like Nelson's looks like in your practice.

→ Book a demo