A Data-Driven Approach Identifies Subtypes of Caries From Dental Charting

Area of Science:

• Oral Health Research
• Public Health Surveillance
• Biostatistics

Background:

• The Swedish Quality Register on Caries and Periodontitis (SKaPa) collects adult dental data.
• Validating registry data and exploring advanced analytical methods are crucial for public health.
• Understanding caries patterns can inform targeted interventions.

Purpose of the Study:

• To assess the accuracy of adult dental data in the SKaPa register.
• To determine if Latent Class Analysis (LCA) can identify distinct patient groups based on caries data.
• To characterize the dental, medical, and behavioral profiles of these identified groups.

Main Methods:

• Dental data from SKaPa were validated against clinical data from dentists in the Malmö Offspring Dental Study (MODS).
• Latent Class Analysis (LCA) was used to classify 61,984 adult participants from the Västerbotten Intervention Programme (VIP) into five groups.
• Dental status, medical history, anthropometric, and behavioral data were compared across LCA-derived classes and DMFS-ranked quintiles.

Main Results:

• SKaPa dental data demonstrated excellent agreement with reference clinical data (intra-class correlations > 0.90).
• LCA successfully identified five distinct classes of participants based on DMFS scores, ranging from 10.0 to 94.4.
• Significant associations were found between LCA classes and various health/behavioral measures, with LCA capturing information not evident in DMFS-ranked groups.

Conclusions:

• Adult dental data within a 2-year window in the SKaPa registry are accurate.
• LCA provides a robust method for classifying individuals into caries subtypes.
• These caries subtypes exhibit differing health and behavioral characteristics and future caries progression, highlighting LCA's potential in personalized oral health.