Transcript What is an Expert System?

What is an Expert System?


An expert system is a program
implemented in a computer which is
designed to mimic the problem-solving
ability of a specialist within a given
field
Expert systems are based on concepts
which are extensions of principles of
artificial intelligence (AI)
What Expert Systems Do


An AI program is composed of a
knowledge base, provided by an
expert, and a methodology of drawing
conclusions based on evidence (input)
and initial propositions.
Expert systems are able to interpret
information derived from basic
research and applied models and
threshold values.


Expert systems utilize their ability to
postulate solutions, which serve as a
basis for analysis, in order to solve a
problem or propose possible solutions.
In a clinical setting, the system could
be designed to propose possible
diagnoses based on symptoms
(patient input) and patient medical
background (history).
Why Apply Expert Systems
to Medicine?



Patents and doctors alike are both
looking for a “quick fix”
Physicians, in recent years, have
become increasingly alienated with the
clinical setting
Medical diagnoses are fundamentally
uncertain
Advantages






New level of patient-physician honesty
Significant money savings
Possible 24hr service
Very short wait periods
Patient medical history known and
considered in diagnosis
Automatic prescription transferral
Disadvantages/Issues




Controversy – can people get the
same quality medical advice and
attention from a program (machine),
that a person can offer?
Can human error ever be completely
eliminated?
Machine diagnosis impersonal
Lack of trust
Specific Example

MatheMEDics® Thorask

Windows-based software

Designed to assist doctors in
diagnosing the cause of chest pain
How it Works


The physician will input answers to
basic inquiries regarding the patient’s
medical history and current symptoms
The program will respond with a list of
possible diagnoses (in decreasing
order of potential) along with the
associated probability
Conclusions

When used in conjunction with a
physicians exam and diagnosis
probability-based expert medical
systems could significantly improve
the practice of medicine
References



1.) Sturman, M.F. MD, FACP; “Health
Controversies” EasyDiagnosis (January 2004)
http://easydiagnosis.com/articles/technology
2.) Unknown, “Fund for Rural America”
Pennsylvania State University (January1999)
http://www.fra.cas.psu.edu/expsys.htm
3.) Bansal, M. Dr.; “Medical Informatics: A
Promising Future” Healthcare Management
(July 2003)
http://www.expresshealthcaremgmt.com/2003
0731/ithealthcare4.html