NEW YORK, US: Identifying people likely to have untreated caries before they present for dental care could help public health services prioritise examinations and reach those most in need. A recent study investigated whether artificial intelligence (AI) models could predict untreated caries and cumulative caries experience using socio-economic, demographic and behavioural data alone. The results suggest that this approach may have potential for public health screening to support more targeted preventive dental care based on established social determinants of health.
“These are powerful tools that could be helpful for targeting resources and delivering preventive care to specific groups that will benefit the most,” senior author Dr Fernando Neves Hugo, chair of the Department of Epidemiology and Health Promotion at the New York University College of Dentistry, said in a university press release.
The researchers analysed data from 23,726 participants in SB Brasil 2023, a national cross-sectional epidemiological survey coordinated by the Brazilian Ministry of Health. Their analysis covered three age cohorts: children aged 12, adolescents aged 15–19 and adults aged 35–44. The survey collected demographic, socio-economic and behavioural data, including household wealth, household density, private dental insurance status, access to piped water and time since the last dental visit, as well as self-rated oral health. Participants also underwent clinical examinations to assess the presence of untreated caries and their DMFT scores.
The researchers used this data to develop and evaluate several statistical and machine learning models. The models demonstrated potential for identifying individuals with untreated caries using socio-economic, demographic and behavioural data only. Their predictive performance improved with cohort age, reaching moderate predictive performance for the adult cohort. A statistical assessment of their potential use in public health screening also suggested that using these models to prioritise examinations could offer advantages over the hypothetical alternatives of examining everyone or no one.
However, the models performed less well in predicting cumulative caries experience, measured by DMFT scores. Simpler statistical models generally performed as well as or better than more complex machine learning algorithms for this outcome, suggesting that greater model complexity does not necessarily improve predictive accuracy. In a separate analysis that also included clinical predictors, household wealth and household density emerged as important contributors to the models’ predictions of caries experience.
The researchers envisaged the models as part of a broader approach to using social determinants of health to guide public health screening and planning. “Integrating machine learning into this framework goes beyond merely predicting cavity risk; it addresses the complexity and multidimensional interactions between social determinants that traditional measures often mask,” explained Dr Hugo.
The researchers noted that the findings may also be relevant to other middle-income countries, although local differences in access to dental care and insurance coverage could influence the importance of individual predictors. To explore the potential for developing country-specific prevention strategies, the researchers are now preparing to conduct a similar analysis using US data from the National Health and Nutrition Examination Survey.
The study, titled “Machine learning for predicting dental caries presence and severity: An age-cohort perspective from a nationally representative survey in Brazil”, was published online on 10 October 2026 in Caries Research, ahead of final publication.
Topics:
Tags:
Artificial intelligence (AI) and augmented intelligence are rapidly becoming part of dentistry, and applications span imaging, documentation, treatment ...
GUANGZHOU, China: Artificial intelligence (AI) has already demonstrated substantial utility in dentistry. Researchers are now exploring whether it could ...
Ask a group of dentists whether they use artificial intelligence (AI) in their clinical practice, and you will get a wide range of answers. Some will say ...
Artificial intelligence (AI), robotics and digital workflows are reshaping dentistry and challenging the profession to rethink clinical care, dental ...
CHICAGO, US: Foods and beverages are attracting increasing research interest for their potential effects on oral biofilms. In two randomised controlled ...
Live webinar
Thu. 19 November 2026
12:00 PM EST (New York)
Dr. Petros Yuvanoglu D.M.D., Prosthodontist, Boston University / Tufts University, Dr. Robert Gottlander DDS
Live webinar
Thu. 19 November 2026
1:00 PM EST (New York)
Live webinar
Fri. 20 November 2026
7:00 AM EST (New York)
Live webinar
Fri. 20 November 2026
8:00 AM EST (New York)
Live webinar
Fri. 20 November 2026
9:00 AM EST (New York)
Dr. Bruce McFarlane Certified Specialist in Orthodontics Fellow: Royal College of Dentists of Canada Diplomate: American Board of Orthodontics
Live webinar
Fri. 20 November 2026
10:00 AM EST (New York)
Live webinar
Fri. 20 November 2026
11:00 AM EST (New York)
To post a reply please login or register