Investigating the effectiveness of water fluoridation.

DESIGN

This retrospective cohort study used treatment claims data submitted over a 10-year period to explore the effect of water fluoridation on specified National Health Service (NHS) dental treatments, number of Decayed Missing and Filled Teeth (DMFT) and its cost-effectiveness. Ethical approval was granted and data was collected from NHS primary care settings via claims submitted to the NHS Business Services Authority (NHS BSA). To be included, participants must have attended dental services twice in the study period, been 12 years or over and had a valid English postcode. Those with claims related solely to orthodontic care were excluded, as were those who had requested NHS National Data Opt-out. Costs relating to water fluoridation were supplied by Public Health England. NHS BSA data was used to calculate NHS costs at 2020 prices.

COHORT SELECTION

A personalised water fluoride exposure for the 2010-2020 period was assigned to all individuals, who were then split into two groups, above 0.7 mg F/L (optimally fluoridated group) or lower (non-optimally fluoridated group). Individuals in each group were matched for analysis using propensity scores, estimated via logistic regression.

DATA ANALYSIS

Values of absolute standardised mean differences were used to determine covariate balance between the two groups, alongside a generalised linear model with matching weights and cluster robust standard errors and a patient deprivation decile as an interaction term. An Incremental Cost-Effectiveness Ratio (ICER) was calculated and differences in the overall costs to the public sector were illustrated by the return on investment estimate.

RESULTS

The cohort contained data on 6,370,280 individuals. Negative binomial regression models were used to analyse health outcomes. In the optimally fluoridated group, the rate of invasive dental treatments was 3% less than in the non-optimally fluoridated group, and the mean DMFT in the optimally fluoridated group was 2% lower. There was no evidence of a difference in the predicted mean number of missing teeth between groups. There was a small reduction in the predicted number of invasive treatments in the optimally fluoridated group but the largest predicted reduction was in the most deprived decile. DMFT did not exhibit the expected social inequalities gradient, and for the mean number of missing teeth there were small differences in each decile of deprivation between groups but the direct effect was inconsistent. Water fluoridation expenditure between 2010 and 2019 was estimated to be £10.30 for those receiving optimally fluoridated water. The marginal effects estimate illustrated savings of £22.26 per person (95% CI - £21.43, -£23.09), which is a relative reduction in costs to the NHS of 5.5% per patient. A subsequent estimation of cost effectiveness calculated the cost of water fluoridation to avoid one invasive dental treatment (the ICER) as £94.55. The estimated return on investment using a variety of NHS dental attendance estimates all lead to a positive return.

CONCLUSIONS

These results suggest that water fluoridation appears to be producing less impactful effects on oral health, with water fluoridation resulting in 'exceedingly small' health effects and very small reductions in use of NHS dental services. A positive return for the public sector was identified as the costs of NHS dentistry are high and costs of water fluoridation are low, though this study did not include the original set up costs of fluoridation programmes.