Exploring the association between early childhood caries, malnutrition, and anemia by machine learning algorithm

Area of Science:

• Pediatric Dentistry
• Public Health
• Machine Learning in Healthcare

Background:

• Early childhood caries (ECC) is a significant public health issue, particularly in vulnerable populations.
• Severe acute malnutrition (SAM) is associated with numerous health complications, potentially including oral health problems.
• Understanding the multifactorial nature of ECC in SAM children is crucial for effective intervention.

Purpose of the Study:

• To determine the prevalence of ECC in children with severe acute malnutrition (SAM).
• To identify the hierarchy of association between ECC, malnutrition, anemia, and other risk factors using machine learning.
• To inform integrated care strategies for pediatricians and pediatric dentists.

Main Methods:

• A hospital-based preventive and interventional study was conducted on SAM children aged 2 to <6 years.
• Oral examinations for ECC were performed using the deft index, alongside recording anthropometric and blood examination data.
• Three machine learning algorithms (Random Tree, CART, Neural Network) were employed to analyze relationships between ECC and risk factors.

Main Results:

• The Random Tree model identified age (98.75%) as the most significant predictor of ECC.
• Maternal education (29.20%) and hemoglobin level (16.67%) were also important predictors.
• Other factors like snack intake frequency, breastfeeding, and SAM status showed varying degrees of association with ECC.

Conclusions:

• Anemia and malnutrition are significant factors in the prediction and causation of ECC.
• Pediatricians must recognize the negative impact of anemia and malnutrition on children's dental health.
• Collaborative efforts between pediatricians and pediatric dentists are essential for managing ECC in SAM children.