"Electronic Pediatrician", a non-machine learning prototype artificial intelligence software for pediatric...

Area of Science:

• Medical Informatics
• Artificial Intelligence in Medicine
• Pediatrics

Background:

• Knowledge-based systems (KBS) have been utilized for computer-assisted medical diagnosis since the 1970s.
• Existing systems include VisualDx, GIDEON, DXPlain, CADUCEUS, Internist-I, and Mycin.
• These systems aim to support clinical decision-making in various medical fields.

Purpose of the Study:

• To detail the Electronic Pediatrician (EPed), a prototype non-machine learning artificial intelligence (nml-AI) knowledge-based system (KBS).
• EPed is designed for the differential and positive diagnosis and treatment of ill children, with a database in Romanian.
• The system focuses on physiopathological reasoning for pediatric clinical cases.

Main Methods:

• EPed version 2.0 utilizes a physiopathological approach to pediatric case analysis.
• The system incorporates 302 defined physiopathological 'clusters'.
• It is capable of diagnosing 269 distinct pediatric diseases.

Main Results:

• The EPed prototype can diagnose 269 pediatric diseases, encompassing infectious and non-infectious conditions.
• Diagnosable conditions include common respiratory, digestive, renal, and central nervous system infections.
• The system also addresses non-infectious diseases such as autoimmune disorders, oncological conditions, genetic diseases, intoxications, and surgical pathologies.

Conclusions:

• EPed is the first and only physiopathology-based nml-AI KBS specifically for general pediatrics.
• It is also the first pediatric KBS developed in Romania for medical professionals.
• EPed uniquely provides physiopathology-based diagnosis and identifies explanatory 'clusters' to aid clinical reasoning and treatment planning.