[Methodological discussion about prevalence of the dental fluorosis on dental health surveys].

Objective: To analyze the limitations of studying dental fluorosis in cross-sectional studies.
Methods: Data from the Oral Health of the Brazilian Population (SBBrasil 2003) and the Brazilian Oral Health Survey (SBBrasil 2010) were used. Epidemiological trends for fluorosis in 12-year-old Brazilians, aspects of the reliability of the data as well as the accuracy of the estimates are assessed for these two studies. The distribution of prevalence of fluorosis was carried out according to the domains of the study (state capitals and regions) and the year in which the study took place. The confidence intervals (95%CI) were also shown for simple prevalence (without taking into account level of severity).
Results: The prevalence of dental fluorosis showed considerable variation, between 0% and 61% in 2003 and 0% and 59% in 2010. Inconsistencies were observed in the data in individual terms (for year and for domain) and in the behavior of the trend. Considering the expected prevalence and the data available in the two studies, the minimum sample size should be 1,500 individuals in order to obtain 3.4% and 6.6% confidence intervals, considering the minimum coefficient of variation to be 15%. Given the subjectivity in its classification, examinations for dental fluorosis may show more variation than those for other oral health conditions. The power to establish differences between the domains of the study with the sample of the SBBrasil 2010 is quite limited.
Conclusions: Based on the 2003 and 2010 studies, it was not possible to analyze patterns of dental fluorosis in Brazil; these data are merely exploratory indicators of the prevalence of dental fluorosis. It was impossible to make comparisons due to different analysis models being used in the two surveys. Investigating dental fluorosis in population-based surveys is not even an economically viable technique, using localized epidemiological studies with a sampling plan would be more suitable [corrected].