Identifying early permanent teeth caries factors in children using random forest algorithm

Area of Science:

• Pediatric Dentistry
• Public Health
• Data Science

Background:

• Early permanent dental caries pose a significant threat to long-term oral health.
• Understanding the influencing factors is crucial for effective prevention strategies in children.

Purpose of the Study:

• To identify key factors associated with early permanent dental caries in first-grade Iranian children.
• To compare the performance of Random Forest and logistic regression in predicting caries risk.

Main Methods:

• A cross-sectional study of 778 first-grade children in Tehran, Iran.
• Oral health assessed using the DMFT index; data collected on maternal education, gender, previous caries (dmft index), hygiene practices, and sweet consumption.
• Random Forest and logistic regression analyses were performed to identify risk factors.

Main Results:

• Logistic regression identified dmft index, maternal education, and sweet consumption as significant factors.
• Random Forest analysis indicated male gender, higher maternal education, and lower sweet consumption were linked to being caries-free.
• Random Forest achieved a higher predictive performance (AUC=0.81) than logistic regression (AUC=0.72).

Conclusions:

• Caring for primary teeth and reducing sweet consumption are vital for managing early permanent dental caries.
• Maternal education plays a pivotal role in mitigating caries risk.
• Random Forest is a recommended algorithm for identifying early permanent teeth caries risk factors.