A regression model analysis of longitudinal dental caries data

Area of Science:

• Dental Public Health
• Epidemiology
• Biostatistics

Background:

• Understanding the long-term predictability of dental caries is crucial for effective public health strategies.
• Previous caries experience is a known risk factor, but its predictive power over extended periods requires robust statistical validation.

Purpose of the Study:

• To analyze longitudinal caries data to determine the association between past and future caries experience in a cohort.
• To evaluate the utility of multiple regression analysis in dental epidemiology for bias reduction and variable effect isolation.

Main Methods:

• A cohort of 306 subjects was followed for an average of 33 years, with baseline and follow-up examinations and interviews.
• Data analysis employed contingency tables with enumeration statistics and multiple regression analysis.
• The regression model adjusted for multiple independent variables to assess individual variable effects.

Main Results:

• A strong association was found between caries experience at different time points within the same individual.
• The multiple regression model explained 35% of the variability in Decayed, Missing, and Filled Surfaces (DMFS) scores.
• The regression approach effectively reduced bias and isolated the specific effect of independent variables.

Conclusions:

• Past caries experience is a significant predictor of future caries development over long-term follow-up.
• Multiple regression analysis is a valuable tool for analyzing complex dental epidemiological data, offering robust bias reduction and variable assessment.
• These findings underscore the importance of early caries detection and management for long-term oral health outcomes.