Contemporary multilevel analysis of the effectiveness of water fluoridation in Australia

Water fluoridation was extended in Queensland, Australia, across 2009-2011. A research program was commenced to inform the rationale for and the outcome of this program, to estimate the effectiveness of water fluoridation in preventing caries and to predict changes in caries experience as a result of the extension of fluoridation.

Queensland children were selected through a stratified random sample selection in 2010-2012. Oral epidemiological examinations provided individual-level outcomes for decayed, missing or filled primary or permanent tooth surfaces: dmfs (among 5-8-year-olds) and DMFS (9-14-year-olds). Explanatory factors at the individual-level, school-level and area-level fluoridation status were derived. Data were weighted to represent the population. Three-level multilevel multivariable models were sequentially specified for negative binomial distribution of dmfs/DMFS to estimate rate ratios (RR). The effectiveness of area-level water fluoridation was evaluated in the full models controlling for other factors.

Data from 2,214 5-8 year-olds and 3,186 9-14 year-olds from 207 schools in 16 areas were analysed. Queensland's average dmfs was 4.23 and DMFS 1.47. The lowest levels of dental caries were observed in long-term fluoridated Townsville. In the full models, Townsville children had significantly lower caries experience (RR for dmfs: 0.61 (95%CI: 0.44-0.82); RR for DMFS 0.60 (95%CI: 0.42-0.88)) compared with children in non-fluoridated areas.

Comparison of caries experience of children at the time of the extension of water fluoridation supported the rationale for this population health measure.