NLP-based system for automatic radiology report standardisation

  • Research type

    Research Study

  • Full title

    Trust-Enhanced Natural Language Processing for Standardisation of Radiology Reports

  • IRAS ID

    353659

  • Contact name

    Julian Hough

  • Contact email

    julian.hough@swansea.ac.uk

  • Sponsor organisation

    Swansea University

  • Clinicaltrials.gov Identifier

    N/A, N/A

  • Duration of Study in the UK

    2 years, 4 months, 30 days

  • Research summary

    Radiology reports are integral to radiology and serve as a primary and indispensable conduit of information between radiologists and secondary care providers. Miscommunication in radiology can give rise to negative medical and legal consequences, ultimately adversely impacting the quality of care. Improving communication between radiologists and referring physicians could reduce errors in radiology due to miscommunication. Structured reports, preferred by referring clinicians over the most commonly used in radiology free-text reports, have been proposed as a potential approach to enhancing the communication between radiologists and secondary care practitioners. Radiologists, however, are trained and accustomed to unstructured free-text reports, which they prefer over structured ones. A possible strategy to reconcile this discrepancy would be the automatic structuring of radiology reports. This project aims to develop a system that automatically converts free-text reports into machine-readable structured format.

  • REC name

    N/A

  • REC reference

    N/A