A diagnostic decision support system for general practice

"How do we think?" is a question which has fascinated mankind for many centuries. This book is not an attempt to answer that question, but it does look at one aspect of thinking. An aspect which is just a tiny fraction of the whole field of research on thinking, yet an enormous field in itself, namely medical diagnosis: the way in which doctors try to find out what is wrong with their patients. This task is not performed logically, although logical, scientific methods have been worked out to mimic, improve or even replace it. Yet although humans do not perform this task in a very logical way, they do perform it in some way, following some rules, and being prone to certain types of error. It is very hard to know what goes on in the human mind. The invention of the computer, the "thinking machine", has greatly contributed to the interest in the way we think and to research into mimicking thought processes. This research has yielded many successes and surprising results, yet the "artificial intelligence" displayed by current machines is not very impressive from a human point of view. The developing field of neural network research, by attempting to copy to some extent the hardware in our own heads, may offer additional insights into thinking. This book is the description of an attempt to improve the diagnostic process, by combining human and machine capabilities into a close partnership. It is a report of the underlying philosophy, design, implementation, and a first test of a diagnostic decision support system for general practitioners. The system was designed to assist and improve human reasoning rather than replace it.