Emulation of Community Water Fluoridation Coverage Across US Counties

Abstract

Introduction: Expansion of community water fluoridation has stalled in the United States, leaving 115 million Americans without fluoridated drinking water.

Objective: This study used spatial regression methods to assess contributions of supply-side factors (neighboring counties' fluoridation coverage) and demand-side factors (health literacy, education, and population density of the local county) in predicting the extent of fluoridation in US counties.

Methods: For this cross-sectional ecological analysis, data from the 2014 Water Fluoridation Reporting System for all 3,135 US counties were merged with sociodemographic data from the 2014 American Community Survey and county-level estimates of health literacy based on the National Association of Adult Literacy Survey. We employed multilevel geographically weighted autoregressive models to predict fluoridation coverage of each county as a function of fluoridation coverage of neighboring counties and local-county covariates: either health literacy or sociodemographic characteristics. Akaike's Information Criterion was used to distinguish the better model in terms of explanatory power and parsimony.

Results: In the best-fit model, an increase from the first to third quartile of neighboring counties' fluoridation coverage was associated with an increase of 27.76 percentage points (95% confidence limits [CI] = 27.71, 27.81) in a local county's fluoridation coverage, while an increase from the first to third quartile of local county's health literacy was associated with an increase of 2.8 percentage points (95% CL = 2.68, 2.89). The results are consistent with a process of emulation, in which counties implement fluoridation based upon their population's health literacy and the extent of fluoridation practiced in neighboring counties.

Conclusion: These results suggest that demand for community water fluoridation will increase as health literacy increases within a county. Furthermore, when considering expansion of fluoridation, non-fluoridated communities can benefit from precedents from nearby communities that are fluoridated.

Knowledge transfer statement: Expanded coverage of community water fluoridation has stalled in the United States. The economic theory of diffusion describes how, over time and space, policy enacted in one community can influence public opinion in a neighboring community. This study applies geospatial analysis of county-level data and the theory of policy diffusion to demonstrate that fluoridated counties can promote the implementation of community water fluoridation in their neighboring, non-fluoridated communities.

Keywords: United States; geographic locations; health literacy; public health dentistry; regression analysis; socioeconomic factors.

Figure 1.

Choropleth map of fluoridation coverage within California counties, 2014. Hollow dots reflect the centroids of each county, and solid lines reflect the neighboring counties’ contributions to the spatial error term.

Figure 2.

Cartogram of fluoridation coverage by county. County sizes are weighted by population following the Gastner-Newman method in ArcGIS software. Smaller polygons reflect counties with smaller populations, and larger polygons larger populations. Counties in darker blue reflect greater fluoridation coverage.

Figure 3.

Predicted county-level fluoridation from geographically weighted autoregressive models. The plots demonstrate the predicted fluoridation coverage for health literacy (Model 1, Table 1), neighboring fluoridation coverage (Model 1, Table 1), population density (Model 1, Table 2), and education (Model 2, Table 2). The independent variables are plotted on the x-axes, with a rug plot marked by ticked lines demonstrating the distribution of the independent variables of interest. Solid lines reflect predicted percent fluoridation coverage, with dashed lines reflecting the 95% CIs.