Medical Expert Systems “Consilium” and “Yumix”
Project goal: to
create an expert system helping a general practitioner to “consult”
one's patients in the field of rheumatology, endocrinology,
pulmonology, gastroenterology, and skin diseases.
| “Consilium” Medical Expert System
The expert system is intended to be used for
diagnosis of chronic diseases. It covers most frequent diseases in
rheumatology, endocrinology, pulmonology, gastroenterology, and skin
diseases. Algorithms of diagnoses are described using logical formulas
of Boolean algebra. Symptoms of diseases include physical data,
laboratory tests, x-ray, etc. Software authors used as criteria
descriptions of diseases in eMedicine, Merck manual and GPnotebook.
The system “Consilium” has been created (1998—2004) as an evolving
expert system that includes all the tools necessary for development: a
language for description of events, collection of information during
the process of diagnosis, statistical processing of information the
results of which can be used for further description. A specialized
language developed based on Boolean algebra allows satisfactory
description of disease syndromes using patient parameters, numerical
data of tests. The expert system contains more than 2000 symptoms and
algorithms of more than 1500 diseases.
was developed by a programmer Alexander Gaskov and myself. I was the
author of the idea of this expert system. I was also searching for
diagnostic criteria of diseases and then described them in the terms of
Boolean algebra, applied ideas of the shortest ways of diagnostic
search. I also was the first doctor making trials of the system in
situations of practical diagnosis.
As a result, the expert system had been used for practical diagnostic
purposes in the Pavlodar Center for Immunology from 2002 to 2008. At
present “Consilium” is used by the Russian State Medical Universities
in educational purposes.
“Yumix” Medical Expert System (2005—2015)
After several years of practical
use, the next step was to create an expert system of medical diagnosis
more convenient for the practical diagnosis than mentioned above
Probable diagnoses are provided with detailed information on the
disease that is useful for doctors to make final diagnostic decisions.
In that cases when confirmation of a diagnosis requires laboratory and
other tests, the system suggests the list of these tests.
Finally the user obtains the patient description in a required form,
list of probable diagnoses, and recommended additional test. This text
is converted in Word format and can be edited and printed.
German programmer Alexander Baum has developed the code of the second
expert system. I have developed medical algorithms, orders of data
input and diagnostic search.
Expert System of Medical Diagnosis “Yumix” is still being developed and
improved. It is used in practice of doctors of the Pavlodar Regional
Diagnostic Center. The main advantage of the system is that unlike
doctors it does not “miss” diagnoses. The main disadvantage preventing
its large scale practical application is time taking. In different
diagnostic situations it can take from 15 to 90 minutes to find a
diagnosis that is 2-3 times longer than the time required for a real
skilled doctor. At present we are searching for ideas how to make data
expert system, software development, computer diagnosis, healthcare,
programing, automation, doctors, medical expert system algorithms,