The development of an expert advisory system for the Acquired Immunodeficiency Syndrome (AIDS)
The first cases of the Acquired Immune Deficiency Syndrome (AIDS) were reported by the Center for Disease Control in Atlanta in the United States in 1981. At the time of completion of this thesis the numbers of cases of AIDS worldwide has risen to well in excess of 100 000. Dr. Roger Brittain, District Medical Officer and a Director of the Expert Systems and Decision Support Unit (ESDSU) at the University of Warwick believed that Information Technology could play a role in overcoming some of the problems related to AIDS. Principle among these was the problem of shortages of clinical and other expertise about this disorder in the United Kingdom. This thesis describes research relating to the attempts of a development team from the ESDSU to develop a computer based tool for resolving the problems of a dearth in AIDS-related expertise via an Action Research Programme. My role in the development team was principally in the areas of design and knowledge acquisition relating to the project. This thesis therefore concentrates on the design, knowledge acquisition and evolution of the computer based decision support tool that was developed. The thesis describes the application of Expert Systems and Information Systems Approaches to the problem and how these approaches were both shown to have deficiencies. However it seemed that the deficiencies in each approach were complementary. This led to the development of a model for a computer based decision support tool that inherited features of both an expert system and an information system. This "hybrid" tool was termed an expert advisory system and was designed to address multiple user groups with varying levels of expertise within each user group. The process of developing the Expert Advisory System model led to the invention of development aids such as the Domain Dictionary. Also it highlighted an often ignored problem in computer based decision support tools, that of responsibility for the advice of the tool. The problems of validation of an "active" decision support tool in which advice is accessed via a consultation are highlighted in the complex area of Medicine. This thesis proposes a possible solution in the form of an editorial board which would be responsible for the validation and maintenance of expertise contained in the decision support tool.