Intelligent Framework for Early Detection of Severe Pediatric Diseases from Mild Symptoms

Area of Science:

• Pediatric Medicine
• Computational Health
• Artificial Intelligence in Healthcare

Background:

• Many childhood diseases leading to death or long-term illness have preventable or treatable causes.
• Early-stage symptoms of serious pediatric conditions can be mild, complicating timely diagnosis.
• Accurate and rapid assessment of pediatric disease severity is crucial for effective intervention.

Purpose of the Study:

• To develop a machine learning (ML) framework for detecting the severity of diseases in children.
• To create a system that can differentiate between urgent/severe pediatric conditions and less critical ones.
• To provide parents with guidance on whether immediate emergency room visits are necessary for their child.

Main Methods:

• The research implemented a machine learning (ML) framework utilizing nine distinct ML methods.
• The model incorporates key variables including presenting symptoms, risk factors (e.g., age), and the child's medical history.
• Performance evaluation of the ML methods was conducted to identify the most effective approaches for disease severity detection.

Main Results:

• The proposed ML framework demonstrated high performance in identifying serious pediatric diseases.
• Decision Tree and Random Forest algorithms achieved the highest accuracy, reaching 94%.
• The results indicate the framework's reliability as a pediatric decision-making support system.

Conclusions:

• The developed ML framework is a reliable tool for detecting serious pediatric illnesses.
• The system can effectively assist parents in making informed decisions about seeking emergency medical care for their children.
• This research offers a promising, parent-friendly approach to managing common childhood symptoms and assessing disease severity.