Informed Clinical Management of Acute Stroke: Use of Established Statistical Methods and Development of an Expert System

Weir, Christopher John (1997) Informed Clinical Management of Acute Stroke: Use of Established Statistical Methods and Development of an Expert System. PhD thesis, University of Glasgow.

Full text available as:
[img]
Preview
PDF
Download (8MB) | Preview

Abstract

This thesis applies several statistical techniques which aim to provide informed clinical management in acute stroke. An introduction is given to issues arising in stroke management and expert systems methodology. Three linear discriminant scoring systems (the Allen, Siriraj and Besson scores) intended for the differential diagnosis between ischaemic and haemorrhagic stroke on the basis of clinical presentation are evaluated in chapter 2. Chapter 3 explores whether angiotensin converting enzyme DD genotype is a risk factor for acute stroke or influences stroke outcome as measured by lesion size. Chapters 4 and 5 assess computed tomography, mean cerebral transit time and single-photon emission computed tomography scanning in terms of their accuracy in predicting functional outcome after acute ischaemic stroke. Chapter 6 broadens the search for prognostic factors, looking at the performance of the Guy's prognostic score and established neurological scales (Canadian neurological scale, National Institutes of Health stroke scale, middle cerebral artery neurological scale) in predicting acute stroke outcome. A linear discriminant score, based on simple clinical measurements recorded in the acute stroke unit, is also developed. Chapter 7 looks specifically at the influence of plasma glucose level on survival following acute stroke, after adjusting for other known prognostic factors using Cox's proportional hazards regression model. The remainder of the thesis is concerned with two aspects of acute stroke management. The first of these is the selection of an appropriate clinical trial for an individual patient. A computer program is developed to obtain, in an efficient manner, the information required to check the entry and exclusion criteria for each available clinical trial. The second aspect of stroke management considered is the choice of a suitable method for secondary prevention of stroke in individual acute ischaemic stroke patients. Candidate methods are long-term anticoagulation with warfarin, or aspirin antiplatelet therapy. Expert system methodology is used to combine positive indications for, and contraindications to each of these therapies with clinical data available in the acute stroke unit. The annual risks of recurrent ischaemic stroke, haemorrhagic stroke, myocardial infarction, other ischaemic complications and other haemorrhagic complications are estimated to allow an informed decision on the appropriate method of secondary prevention to be made.

Item Type: Thesis (PhD)
Qualification Level: Doctoral
Additional Information: Adviser: Kennedy Lees
Keywords: Medicine, Biostatistics
Date of Award: 1997
Depositing User: Enlighten Team
Unique ID: glathesis:1997-75328
Copyright: Copyright of this thesis is held by the author.
Date Deposited: 19 Nov 2019 21:01
Last Modified: 19 Nov 2019 21:01
URI: http://theses.gla.ac.uk/id/eprint/75328

Actions (login required)

View Item View Item