Title: Data mining trauma injury data using C5.0 and logistic regression to determine factors associated with death

Authors: Thomas Chesney, Kay Penny, Peter Oakley, Simon Davies, David Chesney, Nicola Maffulli, John Templeton

Addresses: Nottingham University Business School, Jubilee Campus, Wollaton Road, Nottingham, NG8 1BB, UK. ' Centre of Mathematics, Napier University Craiglockhart Campus, Edinburgh, EH14 1DJ, UK. ' Department of Trauma Research, The University Hospital of North Staffordshire, Princes Road, Stoke on Trent, Staffordshire, ST4 7LN, UK. ' Academic Department of Military Emergency Medicine, Royal Centre for Defence Medicine, Institute of Research and Development, University of Birmingham Research Park, Vincent Drive, Edgbaston, Birmingham, B15 2SQ, UK. ' Freeman Hospital, High Heaton, Newcastle upon Tyne, NE7 7DN, UK. ' Department of Trauma and Orthopaedics, Keele University School of Medicine, Keele University, Staffordshire, ST5 5BG, UK. ' Keele University School of Medicine, Keele University, Staffordshire, ST5 5BG, UK

Abstract: Trauma injury data collected over 10 years at a UK hospital are analysed. The data include injury details such as patient age and gender, the mechanism of injury, various measures of injury severity, management interventions, and treatment outcome. Logistic regression modelling was used to determine which factors were independently associated with death during hospital stay. The data mining algorithm C5.0 was also used to determine those factors in the data that can be used to predict whether a patient will live or die. Logistic modelling and C5.0 show that different subsets of injury severity scores, and patient age, are associated with survival. In addition, C5.0 also shows that gender, and whether the patient was referred from another hospital, is important. The two techniques give different insights into those factors associated with death after trauma.

Keywords: C5.0 algorithm; data mining; logistic regression; trauma injury data; UK; United Kingdom; healthcare; patient age; patient gender; injury mechanism; injury severity; management interventions; treatment outcome; patient survival; death factors.

DOI: 10.1504/IJHTM.2009.023725

International Journal of Healthcare Technology and Management, 2009 Vol.10 No.1/2, pp.16 - 26

Published online: 08 Mar 2009 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article