Introduction
CaDDiS is a computer program which attempts to aid the diagnosis of certain common tropical diseases in cattle. The system is based on a bayesian belief network approach and evaluates the relative likelihood of a cow having any of these diseases given the observation of particular clinical signs. CaDDiS is a Microsoft WindowsTM based application which was designed and developed by Dr. Yiqun Gu of the University of Strathclyde. The system was developed in conjunction with research carried out between 1994 and 1996 with scientists at the International Livestock Research Institute in Nairobi. This research was funded by the UK Overseas Development Administration under Grant No. R5405(H). The system has been parameterised using data collected from 41 veterinary experts. This data was analysed and converted into a suitable format for use in the belief network by Dr. Iain McKendrick also working as part of the project team at the University of Strathclyde. The project team gratefully acknowledge the contributions of the 41 veterinary experts whose expert opinions are codified in the system (a list of names and addresses is provided). However, responsibility for the further development of the software lies with the Veterinary Informatics and Epidemiology group of the University of Strathclyde, to whom any errors, omissions or comments should be directed. The project team can be contacted by writing to the following address:
Professor George Gettinby
Department of Statistics and Modelling Science
Livingstone Tower
26 Richmond Street
Glasgow
The project team may also be contacted by telephone, fax and e-mail at the following numbers:
Phone: (44) 141 552 4400 ext 3805
Fax: (44) 141 552 2079
E-mail:
George@stams.strath.ac.uk
Note
This system has not been validated. It is being circulated to interested
scientists to test whether the diagnoses tally with those which would be
delivered by veterinarians in the field. This system is a research prototype
which should not be used in the field to diagnose cattle disease. The University
of Strathclyde, the designers of the system and contributing veterinary
experts will not be liable for any loss or damage which results from such
unauthorised use.
Background
CaDDiS is designed to serve as a component of a larger decision support
system which will provide information to decision makers within the African
agricultural community. The objective of this decision support module is
to facilitate better differential diagnosis of tropical diseases. The linking
together of different types of techniques such as expert systems, mathematical
models, statistical analysis systems, hypertextualised media systems and
databases to handle different types of information will produce a system
which has analytical and advisory potential. In the first instance, the
decision support system will provide advice mainly in the field of animal
heath, although a fully-fledged system would require the incorporation
of information about animal husbandry and nutrition. On the basis of past
veterinary history and/or current conditions, the decision support system
will recommend courses of action which reduce disease-induced production
losses. It is also hoped that the information contained within this type
of system will find use in training programmes for veterinary assistants
and other agricultural research/advisory staff.
Lauritzen and Spiegelhalter Reference
Lauritzen, S. and Spiegelhalter, D. (1988). Local computations with probabilities on graphical structures and their application to expert systems. J. Royal Statistical Society B, 50: 157-224.