Outcome Prediction and Decision Support for Stroke using Routine Data

  • Research type

    Research Study

  • Full title

    Outcome Prediction and Decision Support for Stroke using Routine Data

  • IRAS ID

    302301

  • Contact name

    Philip Clatworthy

  • Contact email

    Phil.Clatworthy@bristol.ac.uk

  • Sponsor organisation

    North Bristol NHS Trust

  • Duration of Study in the UK

    2 years, 11 months, 30 days

  • Research summary

    Stroke is a major cause of disability amongst adults worldwide. Treatment decisions depend on clinicians’ ability to anticipate the consequences of different courses of action. In stroke, this frequently involves combining information gained from imaging (both images and radiology reports) with knowledge of an individual patient’s current status and medical history. Decision support tools exist which, for example, analyse imaging data to provide insights, or use basic clinical data, sometimes combined with a very basic imaging parameter e.g. stroke lesion volume, to help clinicians make treatment decisions.

    This project will use a combination of clinical and imaging data to predict outcome after stroke and where appropriate estimate the effect of different treatment choices on clinical outcomes. Machine learning methods including deep learning techniques employing convolutional networks, and natural language processing, will be used alongside conventional analysis techniques.

  • REC name

    London - Queen Square Research Ethics Committee

  • REC reference

    22/PR/0122

  • Date of REC Opinion

    22 Sep 2022

  • REC opinion

    Further Information Favourable Opinion