This study employed a longitudinal ecological design using prefectural-level panel data from Japan spanning 2007 to 2018. We chose 2018 as the endpoint because it was the latest year with available data, and 2007 as the starting year to ensure a period of more than 10 years to assess medium- to long-term trends in S-FMR program dissemination. The unit of analysis was the prefecture because the exposure of interest, such as prefectural dental health ordinances and their policy language, operates at the prefectural level and is expected to influence the dissemination of S-FMR programs through prefecture-to-municipality policy channels. Moreover, for policymakers, decisions regarding ordinance enactment or revision are typically based on aggregated population-level outcomes rather than on individual-level effects (20). The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines were followed.

The exclusion criteria were as follows: 1) prefectures without dental health ordinances as of April 1, 2018 (n = 4), 2) prefecture that experienced change in the ordinance category during the study period (n = 1), and 3) prefectures with only one data point before the enactment of dental health ordinances, that is, ordinances enacted in 2008 or 2009 (n = 3) (Figure 1). Prefectures without dental health ordinances were excluded because of their small sample sizes and because this study focused on the policy language of such ordinances. Prefectures with only one pre-ordinance data point were excluded because the pre-intervention trends could not be adequately evaluated.

Information on policy language in dental health ordinances from all prefectures in Japan as of April 1, 2018, was obtained from official prefectural websites and the 8020 Promotion Foundation website (21). Prefectural dental health ordinances are discretionary local policy regulations that are not mandated by national law (22). First enacted in 2008, they have since expanded nationwide (22), with 45 of 47 prefectures in Japan having established such ordinances as of March 2023 (21). Although the Act Concerning the Promotion of Dental and Oral Health in Japan was promulgated in 2011 (23, 24), the widespread enactment of these ordinances preceded national legislation, reportedly accelerating its establishment (22). These prefectural ordinances play an important role in reducing regional inequalities in oral health and ensuring coordination with other policies (22). In this study, prefectural dental health ordinances enacted between 2010 and 2014 were categorized into three groups based on the presence of explicit fluoride-related policy language: 1) explicit mention of “fluoride mouth-rinse” (FMR group); 2) explicit mention of “fluoride application” (FA group); and 3) no explicit fluoride-related policy language (NF group).

The outcome, defined as the dissemination of S-FMR programs, was the annual proportion of children aged 4–15 years participating in S-FMR programs in each prefecture, calculated by dividing the number of children participating in S-FMR programs by the total number of eligible children aged 4–15 years.

Data on the number of children participating in S-FMR programs were collected on a biennial basis. Data for 2007, 2009, and 2011 were derived from joint surveys conducted by the Nonprofit Japanese Conference on the Promotion of the Use of Fluoride in Caries Prevention, the 8020 Promotion Foundation, and the WHO Collaborating Centre for Translation of Oral Health Science (8, 25, 26). Data for 2013 and 2015 additionally included collaborations with the Japan Association of School Dentists (27, 28). The 2018 data originated from the Ministry of Health, Labour, and Welfare (MHLW) (29). The total number of eligible children aged 4–15 years in each prefecture each year, including nursery schools, kindergartens, elementary schools, junior high schools, secondary education schools, and special support schools, was obtained from government statistics [nursery schools: statistics of MHLW (30); others: statistics of Ministry of Education, Culture, Sports, Science, and Technology (31)].

The total prefectural income per year (32), prefectural mean age (33), and prefectural mean number of decayed, missing, or filled primary teeth (dmft) among 3-year-old children (34) were included as covariates. These variables were selected based on a previous study (14) and theoretical considerations because they reflect socioeconomic conditions, demographic structure, and baseline risk of dental caries at the prefectural level. In addition, in our empirical analyses, this set of covariates was the one for which the conditional parallel trends assumption in the Callaway and Sant'Anna Difference-in-Differences (CSDID) method was satisfied. Data for total prefectural income per year and the prefectural mean dmft of 3-year-old children were obtained for the same year as the outcome. Because the prefectural mean age was collected at 5-year intervals, the nearest preceding survey data were used when data corresponding to the outcome year were not available.

Four analytical steps were undertaken. First, we described the trends in outcomes and covariates by survey year. Second, we summarized the timing of dental health ordinance enactment and compared the proportions of children participating in S-FMR programs before and after ordinance enactment within prefectures. Group-level pre- and post-intervention comparisons were performed using the Wilcoxon signed-rank test. Third, we graphed the time trends in the proportion of children participating in S-FMR programs from 2007 to 2018 for each of the three groups.

Fourth, we estimated the impact of prefectural dental health ordinances on the proportion of children participating in S-FMR programs using the CSDID method (35). Difference-in-Differences is a widely used method for evaluating the causal effects of policy interventions. However, traditional two-way fixed-effects regression models have been criticized for producing biased estimates when the timing of interventions varies across units (35). As an alternative, the CSDID method, which offers a more robust statistical framework for settings with staggered treatment adoption, was developed (35). In this study, because the timing of the enactment of dental health ordinances differed among prefectures, the CSDID was considered the most suitable approach. Using the NF group as the control, we estimated the coefficients representing the average treatment effect on the treated (ATT) and 95% confidence intervals (CIs) for the FMR and FA groups.

The parallel trends assumption, requiring that pre-intervention outcome trajectories for the treatment and control groups are similar, was formally assessed using event studies with CSDID. The parallel trends assumption was violated in the preliminary model without covariates. The assumption was satisfied after covariate adjustment, indicating that these variables accounted for the observed differences in the pre-treatment trends. Consequently, analyses were conducted under the conditional parallel trends assumption; that is, after adjusting for covariates, the untreated potential outcomes for the treated and control groups were assumed to follow similar trajectories in the absence of treatment, as permitted by the CSDID framework (35).

Although the interval between the 2015 and 2018 data points was three years, whereas all other intervals were 2 years apart, we coded the 2018 observation as 2017 when constructing the event time for the CSDID analyses to maintain a consistent spacing of periods. All analyses were conducted using Stata SE version 17 (Stata Corp., College Station, TX, USA), with “csdid” package used for CSDID analyses. Statistical significance level was set at α = 0.05, and P-values of <0.05 were considered statistically significant.

This longitudinal, prefecture-level, quasi-experimental study found that prefectural dental health ordinances including explicit fluoride-related language were associated with a greater increase in the proportion of children participating in S-FMR programs. In particular, ordinances that explicitly specified “fluoride mouth-rinse” showed a larger and progressively strengthening impact on S-FMR participation compared with ordinances that referred more broadly to “fluoride application” or contained no explicit fluoride-related language.

These findings are partially consistent with implementation research demonstrating that the greater specificity in policy instruments can facilitate effective implementation by reducing ambiguity among local actors (18, 36). Ambiguous policy language allows discretion in interpretation, which may be advantageous during policy formulation but can place an interpretive burden on local implementers and delay the translation of prefectural policy into concrete, on-the-ground implementation (18, 37).

In the present study, ordinances referring broadly to “fluoride application” likely allowed municipalities to interpret the policy in multiple ways, such as emphasizing individual-level measures (e.g., promoting the use of fluoridated toothpaste) rather than implementing resource-intensive, population-based programs like S-FMR. Ordinances without explicit fluoride-related language may have been even more ambiguous. Consequently, these ordinances may have exerted a weaker influence on the dissemination of S-FMR programs.

In contrast, ordinances explicitly specifying “fluoride mouth-rinse” clearly delineates the intended programmatic modality, minimizing the need for local authorities to expend effort on interpretation. This enhanced policy clarity may have enabled local implementers to move directly toward planning and budgeting for the most cost-effective large-scale interventions suitable for school settings, thereby accelerating implementation. Furthermore, explicitly specifying the optimal delivery modality (i.e., S-FMR programs) in the ordinance may have facilitated intersectoral alignment between the health and education departments, thereby streamlining the governance processes required to achieve higher coverage.

Our results differ from those of a previous Japanese cross-sectional study that examined the correlation between the policy language of prefectural dental health ordinances and the proportion of municipalities implementing S-FMR programs in 2011 (38). In that study, no significant association was observed. This lack of association is likely explained by the fact that only 23 prefectures had enacted dental health ordinances at that time and, as suggested by the present results, the effects of such ordinances strengthen over time, making it difficult to detect a significant relationship in an early cross-sectional snapshot.

These findings have important implications for the implementation of oral health policies. S-FMR programs are supported by strong evidence of their effectiveness, cost-effectiveness, and equity; however, their dissemination remains heterogeneous across regions. Our results suggest that, when prefectural officials draft dental health ordinances, they can strategically use clearer and more explicit fluoride-related wording as an implementation tool to facilitate the uptake of S-FMR programs. Given that only about 13% of children and adolescents participated in S-FMR programs nationally in Japan in 2018 (29), an increase of 8% in the FMR group and 5% in the FA group represents a meaningful public health gain. Thus, for policymakers seeking to expand S-FMR coverage while navigating negotiations with stakeholders, this study provides empirical support for using specific, explicit references to “fluoride mouth-rinse” in ordinance language as a practical lever to promote broader dissemination of the program.

This study has several strengths. First, to our knowledge, this is the first study to use long-term panel data to longitudinally examine the impact of prefectural dental health ordinances on the dissemination of S-FMR programs at the prefectural level. Second, it is the first study to focus explicitly on the policy language of these ordinances and assess their effects over time, thereby providing a more nuanced understanding of how ordinance specificity is associated with implementation. Third, the use of a quasi-experimental CSDID method strengthened the causal interpretation of the effects of different categories of policy language on dental health ordinances.

This study has some limitations. First, as an ecological study based on prefecture-level data, the associations observed at the prefectural level cannot be directly interpreted at the individual-level. Furthermore, the use of aggregate data limited the ability to account for key individual-level determinants of children's oral health, including family and parenting practices and socioeconomic circumstances. However, this design was appropriate given that the objective was to assess the impact of prefectural dental health ordinances at the prefectural level, and because policymakers typically make decisions about ordinance enactment or revision on the basis of aggregated population-level outcomes rather than individual-level effects (20). Second, prefectures without dental health ordinances were excluded because of their small sample size; therefore, the study could not assess the impact of the presence or absence of an ordinance on S-FMR dissemination. Third, one prefecture that changed its ordinance category during the study period was excluded because of the small number of such cases. Therefore, it remains unclear whether revising existing ordinances would produce effects comparable to those observed in this study. Some prefectures revised their ordinance categories only after the end of the study period. Therefore, future research should examine whether revisions to the ordinance language led to subsequent improvements in S-FMR coverage. Fourth, covariates were selected to satisfy the conditional parallel trends assumption. However, unmeasured confounding factors may remain, such as the involvement of dentists in prefectural policymaking and the number of dentists in each prefecture, which could bias the estimated effects. Future studies that comprehensively consider these covariates are warranted. Finally, the generalizability of these findings to other countries and health policies beyond S-FMR programs is limited as this study focuses specifically on Japanese prefectures and S-FMR programs. Specifically, differences in legal systems, governance structures, and oral health care delivery across countries may limit the direct transferability of our findings to other settings. Nevertheless, the findings highlight a potentially generalizable mechanism through which policy design, particularly the use of explicit and actionable language, can facilitate the implementation of preventive health programs. To advance the implementation of evidence-based public health interventions that, like S-FMR programs, are supported by robust evidence yet display substantial regional heterogeneity in uptake, further implementation research explicitly investigating the role of legal frameworks in shaping dissemination and adoption is warranted, and the present study may provide a foundation for such work.