Bayesian Meta-Analysis with Application in Dental Studies

Abstract

Dental caries remain a persistent global health challenge, and fluoride varnish is widely used as a preventive intervention. This study synthesizes evidence from multiple clinical trials to evaluate the effectiveness of fluoride varnish in reducing Decayed-Missing-Filled (DMF) surfaces. The principal measure of efficacy is the Prevented Fraction (PF), representing the proportional reduction in caries relative to untreated controls. A comprehensive meta-analysis was conducted using fixed-effect and random-effects models, complemented by hierarchical Bayesian inference. The Bayesian framework incorporated multiple prior distributions on between-study variance, including Pareto, half-normal, uniform, beta, and scaled chi-square forms, to assess robustness under alternative heterogeneity assumptions. Across all specifications, the pooled estimate indicated an approximate 43% reduction in caries incidence, with credible intervals consistently excluding the null. Compared to classical methods, the Bayesian approach provided richer uncertainty quantification through full posterior distributions, allowed principled incorporation of prior evidence, and offered improved inference under heterogeneity and small-sample conditions. The stability of posterior estimates across diverse priors reinforces the robustness and reliability of the conclusions. Overall, findings confirm fluoride varnish as an effective and consistent preventive measure, and demonstrate the value of Bayesian hierarchical modeling as a powerful complement to traditional meta-analytic techniques in dental public health research.