City Research Online

The interpretation of data in intensive care medicine: An application of knowledge-based techniques

Chelsom, J.J.L. (1990). The interpretation of data in intensive care medicine: An application of knowledge-based techniques. (Unpublished Doctoral thesis, City, University of London)

Abstract

Faced with a rapid increase in the amount of data available to form the basis of diagnostic and management decisions in intensive care medicine, clinicians will require the assistance of computers, and in particular knowledge-based systems, if they are to avoid being overwhelmed. The development of knowledge-based systems in medicine can be traced through three generations, starting with systems that reasoned in an ad hoc fashion using surface level knowledge and progressing to systems that attempted to capture deeper knowledge of their specialist domains. More recently, a third generation of systems has emerged that use more rigorous methods of reasoning with surface level knowledge as a platform for the application of deeper knowledge.

A knowledge-based system has been developed for the diagnosis of disorders of acid-base balance and hypoxemic state using a blackboard architecture to control processes of data classification and belief updating in a hierarchy of hypotheses. To acquire the knowledge for the diagnostic system, a knowledge editing environment was developed and the two systems have been combined to form a set of domain independent tools. These tools were applied to a second domain - the diagnosis of hyperlipidaemia.

Evaluation studies performed in the domains of acid-base balance and hyperlipidaemia have shown that the system performs at a level comparable to that of an expert clinician.

Publication Type: Thesis (Doctoral)
Subjects: Q Science
R Medicine > RZ Other systems of medicine
Departments: Doctoral Theses
Doctoral Theses > School of Health & Psychological Sciences Doctoral Theses
[img]
Preview
Text - Accepted Version
Download (14MB) | Preview

Export

Downloads

Downloads per month over past year

View more statistics

Actions (login required)

Admin Login Admin Login