Estimating outcomes in newborn infants using fuzzy logic.

Area of Science:

• Computational intelligence
• Medical informatics
• Neonatology

Background:

• Neonatal Intensive Care Units (NICUs) manage high-risk newborns.
• Accurate mortality risk assessment is crucial for optimal neonatal care.
• Existing methods may lack precision or be invasive.

Purpose of the Study:

• To develop a fuzzy logic-based linguistic model.
• To estimate the risk of death for neonates in the NICU.
• To provide a non-invasive and cost-effective predictive tool.

Main Methods:

• A computational fuzzy logic model was employed.
• Input variables included birth weight, gestational age, Apgar score, and oxygen fraction.
• The Mandani inference method in Matlab software was used for model development.

Main Results:

• The model achieved an 81.9% performance accuracy.
• It estimated an average mortality risk of 49.7% (p<0.001).
• A strong correlation (r=0.80) was observed between model predictions and outcomes.

Conclusions:

• The fuzzy logic model demonstrates good accuracy for predicting neonatal mortality.
• Its non-invasive nature and ease of use make it suitable for clinical application.
• This tool can support clinical decision-making in neonatal care.