Medical expert systems developed in j.MD, a Java based expert system shell: application in clinical laboratories.

Area of Science:

• Clinical Chemistry
• Medical Informatics
• Artificial Intelligence in Healthcare

Background:

• Laboratories face increasing demands for automation due to complex diagnostics, high workloads, accreditation, and budget constraints.
• Expert systems, a form of artificial intelligence, are crucial for automating iterative tasks in clinical laboratories.
• Clinical interpretation of diagnostic tests is a key area benefiting from expert system implementation.

Purpose of the Study:

• To reprogram an existing expert system for amylase isoenzyme pattern interpretation using a new Java-based shell (j.MD).
• To demonstrate the integration capabilities of expert systems with laboratory information systems (LIS).
• To showcase the application of expert systems in automating complex diagnostic test interpretation.

Main Methods:

• Reprogramming an expert system from Pro.MD (Prolog-based) to j.MD (Java-based).
• Implementing a bidirectional link between the j.MD expert system and the laboratory information system.
• Focusing on the clinical interpretation of amylase isoenzyme patterns.

Main Results:

• Successful reprogramming of the expert system using the j.MD shell.
• Established a bidirectional data link between the expert system and the LIS.
• Demonstrated the feasibility of integrating expert systems for complex test interpretation into automated labs.

Conclusions:

• Expert systems, like the one developed with j.MD, can be effectively integrated into automated clinical chemistry laboratories.
• The bidirectional LIS link is a significant advantage for seamless data flow and interpretation.
• This approach enhances the automation of complex diagnostic test interpretation, requiring specialized expertise.