Using LLMs to review hospital data to improve patient safety v1

  • Research type

    Research Study

  • Full title

    Using Large Language Models (LLMs) to review hospital data to improve safety across the patient journey.

  • IRAS ID

    361432

  • Contact name

    Tom Lawton

  • Contact email

    tom.lawton@bthft.nhs.uk

  • Sponsor organisation

    Bradford Teaching Hospitals NHS Foundation Trust

  • Duration of Study in the UK

    2 years, 0 months, 30 days

  • Research summary

    Patient safety management and electronic patient record (EPR) systems contain a large amount of free-text data, which provides valuable insights into patient safety and care. However, analysis of this data often relies on staff manually searching records, extracting data and conducting analysis, limiting the ability to improve the safety of care. This programme of research is looking to see if it is possible for artificial intelligence (specifically large language models) to support this work by automating parts of the process to support the analysis of the free text within patient records.

    Large language models (LLMs) are computer programmes that have been trained to read, interpret and predict text. Existing LLMs held within the hospital’s IT system will be used to perform the research. Patient data will not be used to train the models. The models will be used across four patient safety areas:
    1. Generation of discharge summaries from EPRs.
    2. Analysis of EPRs for evidence of patient safety concepts, like harm and staff adaptations.
    3. Supporting PSIRF safety investigations in two areas:
    a. Using EPRs to assist in generating incident reports.
    b. Analysing PSIRF reports and incident records to assess the quality of the
    investigations.

    To support the research, free-text data from existing data sources will be used:
    1. Electronic patient records
    2. Incident records
    3. PSIRF safety investigation reports (supplemented by additional reports from two other NHS Trusts)

    The outputs from the LLMs will be reviewed by a number of healthcare stakeholders (staff and a patient panel) to provide feedback on its accuracy and suitability. This will be captured using 1-2-1 interviews and focus groups (depending on the stakeholder).

    The research programme will end in April 2028, over the lifecycle of the Y&H PSRC. There are three funding streams for this research programme: the MPS Foundation, CERSI-AI, and NIHR.

  • REC name

    Yorkshire & The Humber - Leeds East Research Ethics Committee

  • REC reference

    26/YH/0071

  • Date of REC Opinion

    16 Apr 2026

  • REC opinion

    Favourable Opinion