DIGITAL HEALTH TECHNOLOGY CATALYST

Swansea University

Bricolage was commissioned by the Neuroscience Research Group (NRG) within Swansea University Medical School to design a research proposal. The research was intended to  investigate how Natural Language Processing could be used to assist with creating correspondence from neurology clinicians working in the field of epilepsy.

Proposal for a proof of concept study to investigate NLP approaches to information extraction from clinical correspondence. A collaboration with Swansea University Medical School (SUMS) Neuroscience Research Group (NRG)

We developed a proposal for a feasibility study to assess the viability of using a text markup and report-authoring software tool to help create neurology clinician letters. The proof of concept would use software to extract anonymised data from existing literature to create a database of different treatments in use for people with epilepsy. Clinicians could then draw on this database to draft correspondence and reports.  Our proposal comprised a three part trial design, including the development of a corpus of medical letters relating to epilepsy diagnosis and assessment criteria; software development and integration with medical ontologies and simulated workflow testing.

Our proposal was used by the Neuroscience Research Group as a foundation from which they were able to single out various aspects of the proposal as the basis for future research and study.

  • Ministry of Defence (ACDS Logistic Operations)
  • Extraction of technical data
  • Intelligence analysis software
  • Information triage
  • Text mining
  • Extraction of financial data
  • Policing software
  • Swansea University
  • Amnesty International ‘Decoding Darfur’