Ambient Voice Technology 'AI-Scribe' FAQs

NHS clinicians spend a lot of time completing administrative tasks like typing up consultation notes and letters. Automating this task reduces the administrative burden for clinicians and improves consultation experience by enabling clinicians to give their full attention to patients.

The AI-Scribe is called TORTUS. The company is a UK based AI start-up and the tool is fully compliant with UK GDPR. No patient data are held by the organisation or used to train the AI. We ran an extensive trial of TORTUS at GOSH and across London to ensure it works and is safe, before starting to use it as part of routine care.

The AI-Scribe, TORTUS, uses ambient AI technology to record conversations between clinicians and patients during consultations and summarise the information into clinic notes, using templates designed by the clinicians. The clinician then checks the clinic note and makes any edits before saving it to the patient's electronic patient record.

Before the trial could take place, GOSH completed detailed and thorough governance and safety checks to ensure that the technology is secure.

The AI-scribe is called TORTUS is fully compliant with UK GDPR meets all of the NHS criteria for AI compliance. No patient data are held by the organisation or used to train the AI.

We ran an extensive trial of TORTUS at GOSH and across London to ensure it works and is safe before starting to use it as part of routine care.

The notes are produced using templates which clinicians can set to their own preferred style and tone.

The clinician always has final oversight to edit and approve the letters before they are finalised and saved to the patient record.

The tool is integrated into our electronic patient record, so will automatically pick up the patient’s name correctly.

The clinician will also check documentation for any errors and make any necessary edits before they are finalised and saved to the patient record.

During our extensive evaluation, we tested with a wide variety of accents and found that the tool worked well at accurately transcribing spoken words.

The clinician always has final oversight and it is their responsibility to check and edit the documentation before it uploaded to the patient record.

For the purposes of drafting the clinic note and letter, it is not necessary to distinguish between who is speaking, as long as all information is spoken out-loud, it will be recorded and used to form the documentation draft.

The clinician always has oversight of the drafted documents and will edit as necessary before saving to the patient record. If the clinician feels it is important to state who provided a certain piece of information, they can ensure to add this information in at this stage.

The tool will pick up conversation between the clinicians and parent or carer. Clinicians have the option of dictating their interpretation of body language to the tool once the patient has left the room if they feel it is important to record.

The clinician will also check and edit the documentation draft before it is saved to the patient record, so can manually type in any additional relevant information, such as body language signs, at this stage.

In our extensive evaluation of the tool, we found that it was able to correctly pick up a wide range of complex medical words. If a word is missed or misspelt, the clinician will correct this before they save the documentation to the patient record.