Public health monitoring of Community Water Fluoridation (CWF) schemes requires estimates of exposure to fluoride in public water supplies (PWS). We aimed to use routine data to estimate population exposure to PWS-fluoride in England and to determine whether PWS-fluoride exposure from 2005 to 2015 could be used as a proxy for exposure for 1995–2004, when fluoride concentration data that could be linked to population health data were unavailable. We calculated annual mean water supply zone PWS-fluoride concentrations from monitoring data for 1995–2015, stratified by fluoridation scheme-flagging. We allocated annual 2005–2015 mean PWS-fluoride concentrations to small area boundaries to describe population exposure within five concentration categories (< 0.1 to ?0.7 mg/L). We compared zone-level 1995–2004 and 2005–2015 mean PWS-fluoride concentrations using Spearman correlation. Most (72%) of the population received PWS with < 0.2 mg/L fluoride and 10% with ?0.7 mg/L. Fluoride concentrations in 1995–2004 and 2005–2015 were similar (median 0.11 mg/L (lower quartile–upper quartile (LQ–UQ) 0.06–0.17) and 0.11 mg/L (LQ–UQ 0.07–0.17), respectively) and highly correlated (coefficient 0.93) if un-fluoridated but differed (1995–2004 median 0.78 mg/L (LQ–UQ 0.59–0.92); 2005–2015 0.84 mg/L (LQ–UQ 0.72–0.95)) and correlated weakly (coefficient 0.31) if fluoridated. Fluoride concentrations in 2005–2015 approximate those in 1995–2004 but with a greater risk of misclassification in fluoridation schemes.

The fluorine element and fluoride compounds (henceforth referred to simply as ‘fluoride’) are naturally occurring and likely to be found in sources of drinking water in varying amounts and are also present in some foods and drinks. Exposure to fluoride can reduce the risk of dental caries (tooth decay) (Selwitz et al. 2007), and Community Water Fluoridation (CWF) schemes that adjust fluoride concentrations in water supplies to target concentrations typically in the range of 0.7–1 mg/L have been shown to effectively reduce caries prevalence and severity in children (Iheozor-Ejiofor et al. 2015). In some parts of England, as a result of the geology of the area, fluoride concentrations in public water supplies (PWS) already reach the UK target concentration for CWF schemes (1 mg/L). In other areas that are part of fluoridation schemes, the fluoride concentration has been adjusted to reach this concentration. Currently, around 10% of the England population (six million people) live in areas with fluoridation schemes where the concentration has been adjusted.

In addition to the known benefits, harmful health effects have been attributed to fluorides; convincing evidence of a causal association with these at the levels permitted by water quality legislation is lacking, though an increase in dental fluorosis has been identified (Iheozor-Ejiofor et al. 2015). Current legislation in England (Statutory Instrument 2018 no. 647 2018) allows for up to 1.5 mg/L of fluoride to be present in PWS which mirrors EU legislation and is intended to be protective against any harmful effects from chronic exposure, including dental fluorosis which might be unsightly. In England, Public Health England (PHE) monitors the health effects of the adjustment of PWS fluoride concentrations for fluoridation schemes on behalf of the Secretary of State for Health and Social Care and in line with legislation (Water Industry Act 1991 c.56 1991). Previous monitoring (Public Health England 2014) and other epidemiological studies (McLaren and Emery 2012; Skinner 2012) have used data from routine PWS monitoring to estimate population exposure to fluoridation. However, these population exposure models were limited to simple binary exposures (i.e. fluoridated or not) rather than the PWS fluoride concentration, risking exposure misclassification and preventing dose–response analysis. The latter may be important when determining the optimal fluoride concentration to maximise caries prevention benefit and minimise dental fluorosis and also to consider evidence for causal associations with health effects for which evidence of an association is less established. Linkage of fluoride PWS concentration data with health data to assign exposure typically requires geo-referencing of PWS monitoring data onto administrative boundaries. Therefore, exposure models may also be limited by constraints in availability of geo-referenced routine monitoring data for certain time periods, meaning assumptions may have to be made about exposure in these periods. Quantification of past population exposures may be important when investigating potential associations between fluoride exposure and caries development in older children (as incorporation of fluoride into developing tooth tissue and after tooth eruption are both likely to play a role in modifying caries risk) (Singh and Spencer 2004). Additionally, quantifying prior exposure may also be useful for investigating more recently occurring health outcomes with longer induction periods, such as some cancers (Checkoway et al. 1990).

We aimed to estimate population exposure to increasing categories of fluoride concentration in PWS in England to use as an exposure indicator in public health monitoring of water fluoridation schemes for the 2018 PHE fluoridation health monitoring report (Public Health England 2018). We further aimed to determine whether contemporary (2005–2015) routine fluoride concentration monitoring data could be used as a proxy indicator of population exposure for prior years when geo-referenceable data was not easily available.