Study Finds Machine Learning Models Can Predict Caries Risk
-
By
-
News Staff
-
November 9, 2026
Clinical Report: Machine Learning Models Predict Caries Risk
Overview
An analysis of Brazil’s 2023 oral health survey found that machine-learning models predicted caries presence and severity across children, adolescents, and adults, with stronger discrimination of caries presence among adults. Socio-structural factors—particularly wealth and household density—were the leading predictors across age groups; the cross-sectional design and lack of external validation limit interpretation.
Background
Dental caries is a population health concern, and the study examined whether artificial intelligence models could help identify groups with greater disease burden. The analysis used data from SB Brasil 2023, a national epidemiological survey coordinated by Brazil’s Ministry of Health. It considered demographic, socioeconomic, behavioral, and clinical information across three age groups. The authors framed population-level risk stratification as a potential aid to targeting public health programs, rather than as individual clinical diagnosis.
Data Highlights
| Measure | Finding |
|---|---|
| Study population | 6,698 children aged 12 years; 8,041 adolescents aged 15–19 years; 8,987 adults aged 35–44 years |
| Most predictive structural factors | Wealth and household density across all age groups |
| Untreated caries and piped water among 12-year-olds | 62.5% without piped-water access versus 35.5% with access |
| Model performance | Predictive performance varied by age group; discrimination of caries presence was stronger among adults |
| Explainability analysis | SHAP analysis ranked socio-structural factors above individual clinical markers across age groups |
Key Findings
- The study applied several machine-learning algorithms to survey data from children, adolescents, and adults to predict caries presence and severity.
- Model performance varied across age groups, with stronger discrimination of caries presence among adults.
- Wealth and household density had the greatest predictive weight among the factors examined and were the primary structural predictors in the SHAP analysis.
- Across age groups, participants classified as non-White had a higher caries burden; private dental insurance and more recent dental visits were associated with lower caries prevalence.
- Among 12-year-olds, untreated caries was reported in 62.5% of participants without piped-water access and 35.5% of those with access.
- The authors identified the cross-sectional design, urban-population focus, absence of external model validation, and uncertain transferability to other countries as limitations.
Clinical Implications
The authors state that AI-based population risk stratification could help identify groups for targeted public health programs and preventive-care resource allocation. The study does not establish causal relationships or validate the models for clinical use in other populations.
Conclusion
In this Brazilian survey analysis, socio-structural variables were more influential than individual clinical markers in machine-learning predictions of caries risk. Further validation is needed, and the authors plan a parallel analysis using US NHANES data.
Related Resources & Content
- Brazilian Ministry of Health, SB Brasil 2023 — First results
- American Academy of Pediatric Dentistry — Caries-risk Assessment and Management for Infants, Children, and Adolescents
- ADA News — Can AI model boost prediction of dental caries?
- ADA News — Predicting periodontitis with machine-learning model
- The ASCO Post — Machine Learning Models Outperform Traditional Approaches to Multicancer Risk Prediction
- aace endocrine ai — Machine learning shows promise for predicting hypoglycemia
- SB Brasil 2023: First results
- AAPD: Caries-risk Assessment and Management for Infants, Children, and Adolescents
- 1. INTRODUÇÃO
- AAPD | Caries-risk Assessment and Management for Infants, Children, and Adolescents
- Clinical effectiveness and cost-effectiveness of two sodium fluoride varnish systems in preventing dental caries in children: a randomized clinical trial
- FDI POLICY STATEMENT
Based on findings from:
Study Finds Machine Learning Models Can Predict Caries Risk
News Staff. Inside Dentistry, 2026.
https://insidedentistry.net/news?newsID=f308b024-9ca5-4bd4-b524-1be566e59438
This content is an AI-generated, fully rewritten summary based on a published scholarly article. It does not reproduce the original text and is not a substitute for the original publication. Readers are encouraged to consult the source for full context, data, and methodology.