Explaining Anomalous Responses to Treatment in the Intensive Care Unit

  • Authors:

  • Laura Moss;Derek Sleeman;Malcolm Booth;Malcolm Daniel;Lyndsay Donaldson;Charlotte Gilhooly;Martin Hughes;Malcolm Sim;John Kinsella

  • Affiliations:

  • Department of Computing Science, University of Aberdeen, Aberdeen, AB24 3UE;Department of Computing Science, University of Aberdeen, Aberdeen, AB24 3UE;University Section of Anaesthesia, Pain & Critical Care Medicine, University of Glasgow, Glasgow G31 2ER;University Section of Anaesthesia, Pain & Critical Care Medicine, University of Glasgow, Glasgow G31 2ER;University Section of Anaesthesia, Pain & Critical Care Medicine, University of Glasgow, Glasgow G31 2ER;University Section of Anaesthesia, Pain & Critical Care Medicine, University of Glasgow, Glasgow G31 2ER;University Section of Anaesthesia, Pain & Critical Care Medicine, University of Glasgow, Glasgow G31 2ER;University Section of Anaesthesia, Pain & Critical Care Medicine, University of Glasgow, Glasgow G31 2ER;University Section of Anaesthesia, Pain & Critical Care Medicine, University of Glasgow, Glasgow G31 2ER

  • Venue:
  • AIME '09 Proceedings of the 12th Conference on Artificial Intelligence in Medicine: Artificial Intelligence in Medicine
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

The Intensive Care Unit (ICU) provides treatment to critically ill patients. When a patient does not respond as expected to such treatment it can be challenging for clinicians, especially junior clinicians, as they may not have the relevant experience to understand the patient's anomalous response. Datasets for 10 patients from Glasgow Royal Infirmary's ICU have been made available to us. We asked several ICU clinicians to review these datasets and to suggest sequences which include anomalous or unusual reactions to treatment. Further, we then asked two ICU clinicians if they agreed with their colleagues' assessments, and if they did to provide possible explanations for these anomalous sequences. Subsequently we have developed a system which is able to replicate the clinicians' explanations based on the knowledge contained in its several ontologies; further the system can suggest additional explanations which will be evaluated by the senior consultant.