Machine learning to predict untreated dental caries in adolescents

Area of Science:

• Public Health
• Data Science
• Pediatric Dentistry

Background:

• Untreated dental caries pose a significant public health challenge for adolescents.
• Early prediction and intervention are crucial for managing adolescent oral health.
• Machine learning offers novel approaches for identifying at-risk populations.

Purpose of the Study:

• To develop and evaluate machine learning models for predicting untreated dental caries in adolescents.
• To identify key predictors associated with untreated dental caries in a Brazilian adolescent population.
• To explore the utility of artificial intelligence in optimizing family health team workflows for dental caries management.

Main Methods:

• Utilized data from an epidemiological survey of 615 adolescents in Mato Grosso do Sul, Brazil.
• Employed three machine learning algorithms: XGBoost, decision tree, and logistic regression.
• Trained and tested predictive models using sociodemographic data, dietary habits, and oral hygiene behaviors.

Main Results:

• XGBoost demonstrated superior performance with an Area Under the Curve (AUC) of 84%, outperforming the decision tree (81%).
• Significant predictors for untreated dental caries included dental floss usage, unhealthy food consumption, self-declared race, and exposure to fluoridated water.
• The models successfully identified individuals with untreated dental caries based on eight main predictor variables.

Conclusions:

• Artificial intelligence, specifically machine learning, can effectively predict adolescents with untreated dental caries.
• Family health teams can leverage these AI tools to enhance their work processes.
• Early prediction facilitates timely dental appointments and treatment, improving adolescent oral health outcomes.