Limitations of fluoridation effectiveness studies: Lessons from Alberta, Canada

A paper published in this journal, "Measuring the short-term impact of fluoridation cessation on dental caries in Grade 2 children using tooth surface indices," by McLaren et al had shortcomings in study design and interpretation of results, and did not include important pertinent data. Its pre-post cross-sectional design relied on comparison of decay rates in two cities: Calgary, which ceased fluoridation, and Edmonton, which maintained fluoridation. Dental health surveys conducted in both cities about 6.5 years prior to fluoridation cessation in Calgary provided the baseline. They were compared to decay rates determined about 2.5 years after cessation in a second set of surveys in both cities. A key shortcoming was the failure to use data from a Calgary dental health survey conducted about 1.5 years prior to cessation. When this third data set is considered, the rate of increase of decay in Calgary is found to be the same before and after cessation of fluoridation, thus contradicting the main conclusion of the paper that cessation was associated with an adverse effect on oral health. Furthermore, the study design is vulnerable to confounding by caries risk factors other than fluoridation: The two cities differed substantially in baseline decay rates, other health indicators, and demographic characteristics associated with caries risk, and these risk factors were not shown to shift in parallel in Edmonton and Calgary through time. An additional weakness was low participation rates in the dental surveys and lack of analysis to check whether this may have resulted in selection biases. Owing to these weaknesses, the study has limited ability to assess whether fluoridation cessation caused an increase in decay. The study's findings, when considered with the additional information from the third Calgary survey, more strongly support the conclusion that cessation of fluoridation had no effect on decay rate. Consideration of the limitations of this study can stimulate improvement in the quality of future fluoridation effectiveness studies.