Salt fluoridation for preventing dental caries
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
The main objective of the review is to assess the effects of salt fluoridation in preventing caries in the population. Specific objectives include:
-
To evaluate the effectiveness and safety of fluoridated salt intervention compared with control (non‐fluoridated salt), as a means of delivering fluoride systemically on a community basis, for preventing dental caries in children, adolescents and adults
-
To examine whether the effectiveness of fluoridated salt on caries prevention is influenced by types of salt fluoridation schemes adopted and other particular intervention features, by initial caries levels and other relevant population characteristics, geographical regions/climates, and by exposure to other non‐systemically delivered fluorides (such as toothpastes)
-
To evaluate the effectiveness of fluoridated salt compared with water or milk fluoridation (differential effect).
Background
Dental caries (tooth decay) is a disease of the hard tissues of the teeth caused by the interactions over time between certain microorganisms found in dental plaque (cariogenic bacteria) and dietary fermentable carbohydrates (principally sugars, such as sucrose). This interaction produces organic acids which dissolve tooth substance (demineralization). Progressive dental caries may result in cavities, pain, and loss of teeth, but given its dynamic nature, once established the carious process may be arrested prior to significant cavitation taking place. Research undertaken on the mechanisms by which fluoride prevents dental caries indicates that the predominant effect of fluoride is direct on the tooth surface (a topical effect), through its interaction with the tooth enamel surface. Fluoride present in saliva and in dental plaque inhibits tooth enamel demineralization, and promotes remineralization. When ingested, fluoride also works through a systemic effect by being built into the enamel during the development of the teeth (pre‐eruptive effect), although this effect is minor compared to the topical one. Fluoridation has been supported by the World Health Organization (WHO 2004).
Dental caries remains an extremely common disease which is strongly related to socio‐economic status ‐ caries is unequally distributed among the population, with caries incidence, prevalence, and severity being greater among economically disadvantaged children than among other groups. Although the disease is largely preventable, and despite substantial improvements in dental health over the last 30 years, it remains a public health problem in most industrialised countries, affecting 60% to 90% of school children and the vast majority of adults (Petersen 2003). In many developing economy countries, such as in the Latin American countries, particularly in Central America, caries prevalence has been notably high, reaching 95% prevalence and a number of decayed, missing and filled teeth (DMFT) over the WHO goal of 3.0 DMFT at 12 years of age (Beltran‐Aguilar 1999). This combined with the lack of access to dental services emphasizes the need for community caries prevention.
Water fluoridation has been the principal approach to community caries prevention for 60 years and is both a primary preventive and a public health measure for caries control. Water fluoridation involves adjusting the fluoride content of the water supply to the 'optimal' level of one part of fluoride per million parts of water (l ppm, or 1 mg/litre), or defined levels in tropical environments (0.5 to 1.0 ppm depending on climate). The advantage of water fluoridation is that it reaches the entire population served by the water distribution system, including the socially deprived groups who constantly have the highest levels of caries. The York review (McDonagh 2000), a systematic review on public water fluoridation conducted to the highest standard, showed that water fluoridation reduces the number of decayed, missing and filled teeth by, on average, two and a quarter teeth per child and on average increases the proportion of children completely free from tooth decay by 15%, but there was substantial heterogeneity between the studies reviewed. The review also showed that water fluoridation increases the risk of dental fluorosis (a form of enamel hypomineralization in the event of excessive ingestion of fluoride during tooth development) ‐ the prevalence of fluorosis at a water fluoride level of 1 ppm was estimated to be 48%, and for fluorosis of aesthetical concern 12.5% of exposed people would be affected. The review concluded that there was little high quality research on fluoride and health, and its findings have often been misinterpreted and have been used to support arguments on both sides of the water fluoridation debate.
Various political, geographical, financial and technical reasons have prevented the availability of water fluoridation to a large proportion of the world's population and alternative methods of fluoride therapy (topical and systemic) have been introduced. Fluoridated toothpaste is the most widely used delivery method of fluoride nowadays, and other methods of community fluoridation including ingested fluorides in milk and salt have been used.
Salt fluoridation began in Switzerland in 1955, and is far more common in Europe than water fluoridation. It is widely used in countries such as Germany, France and Switzerland, where some 30% to 80% of the marketed salt for domestic use is fluoridated (Marthaler 2005), and in a further 30 countries worldwide. The Pan American Health Organization (PAHO), a regional division of the World Health Organization (WHO), has been active in developing strategies to implement caries prevention programmes in the regions of the Americas using both water and salt fluoridation (Gillespie 1986). Salt fluoridation schemes are implemented in more than 10 countries in Central and South America (including Bolivia, Ecuador, Colombia, Peru, Jamaica, Costa Rica, Mexico, Uruguay, and Venezuela), and frequently involve all salt for human consumption (Estupinan‐Day 2005).
The use of salt as a vehicle for providing additional fluoride in a dental public health programme has some important attractions, and the methods and procedures for community introduction of this vehicle appear to be more acceptable and less conflictive than water fluoridation. Salt is an essential component of the diet, reaches all sectors of society, has worldwide distribution, and is not dependent upon a limited distribution or treatment system. Fluoride in salt is also compatible with iodisation. Community‐based caries prevention programmes indicate that salt provides a relatively cost‐effective vehicle for fluoride in the prevention of dental caries (salt fluoridation may be done more cheaply than water fluoridation). Salt is most commonly fluoridated at 250 parts per million (ppm) (range 200 to 250 ppm) which means 250 mg of fluoride per kg of salt, depending on dietary practices. It is expected that the use of table (domestic) salt used at the table and in the kitchen can contribute 1 to 4 g of the daily salt intake. Thus a person would take in 1 mg of fluoride a day at a salt intake of 4 grams a day ‐ 1 mg/day being the 'optimal' dose of fluoride intake (the same concept as in water fluoridation). Properly fluoridated salt produces levels of urinary fluoride excretion similar to those found in communities with fluoridated water.
Challenges with implementation of salt fluoridation can occur when there are multiple sources of drinking water in an area with high fluoride content, requiring that the natural fluoride level of each major source of drinking water must be determined prior to implementation of a salt fluoridation programme. A secondary concern relates to reluctance to implement such a programme, because a high consumption of sodium is a risk factor for hypertension (high blood pressure) and people who must restrict their salt intake may not find salt fluoridation an acceptable method of receiving fluoride. There is also the belief that the availability of fluoridated salt may lead to people consuming more salt. However, water fluoridation did not increase water intake and the same is true with fluoridated salt, they are merely vehicles. Nevertheless, where only domestic salt (as opposed to all salt for human consumption) is fluoridated, it permits people the freedom of choice. Approximately a quarter of salt consumption in many countries comes from domestic salt bought in shops, and any change in the amount purchased, either an increase or decrease, is likely to have little effect on the total consumption. It has been indicated however, that when fluoridation of salt is implemented in a way to reach all consumers of a region, i.e., when both domestic salt and bulk salt is fluoridated (used by commercial bakeries, restaurants, institutions, and industrial food production), the caries‐reducing effect of salt fluoridation may be marginally greater than when only domestic salt is fluoridated (Kunzel 1993; Marthaler 2000).
Nevertheless, the impact of salt fluoridation depends not only on the potential benefits of this intervention (including reduced caries levels), but also on the potential benefits over that offered by the widespread use of common approaches to control caries (e.g. fluoridated toothpaste), and on the potential harms (including dental fluorosis and other possible effects on general health such as bone fractures and bone development problems).
Many studies and reviews are available on salt fluoridation, but no systematic review has been undertaken. In addition, existing reviews do not address the issues of benefit and harm in conjunction and in a systematic manner, as this review aims to do. As the use of fluoridated salt is growing and a clear evaluation of impact has not been made, a systematic review of its effect is therefore essential.
In this review we are primarily assessing the likelihood that use of salt fluoridation will reduce caries in children and adults. While randomised controlled trials (RCTs) are the strongest method for proving a causal effect of an intervention, as these provide the least biased estimates of effect, where such RCTs are lacking other study designs such as non‐randomised studies with a concurrent control group (where the intention is experimental) may also provide weaker evidence of effect when the participants and exposure (salt fluoridation) match what is likely to happen in an intervention programme, and when other confounding factors can be excluded (some of these factors may be related to the outcomes under investigation (such as initial caries level) and so will confound any observed relationship and thus should be controlled for in the analysis, and in the case of salt fluoridation these are also likely to include age, gender, ethnicity, social class and other sources of fluoride exposure (e.g. toothpaste)).
This review aims therefore to assess the evidence on the efficacy and safety of salt fluoridation to prevent caries in children, adolescents and adults focusing on randomised controlled trials and quasi‐randomised trials, and, should no such controlled trials be identified, the inclusion of non‐randomised studies with a concurrent control group would be considered. To achieve this aim three main questions have been identified:
1) Is fluoridated salt effective in preventing dental caries (any positive and negative effects)?
2) Is any caries preventive effect of fluoridated salt influenced by:
(a) salt fluoridation scheme (domestic or universal)
(b) levels of fluoride in salt (in ppm F)
(c) type of fluoride in salt (sodium fluoride or potassium fluoride)
(d) frequency of fluoride salt consumption
(e) baseline caries prevalence levels in the population
(f) different age groups, ethnic/socio‐economic groups
(g) geographical regions, climates
(h) exposure to other non‐systemically delivered fluoride such as fluoridated toothpastes?
3) How does the effect of salt fluoridation compare with water and milk fluoridation?
Objectives
The main objective of the review is to assess the effects of salt fluoridation in preventing caries in the population. Specific objectives include:
-
To evaluate the effectiveness and safety of fluoridated salt intervention compared with control (non‐fluoridated salt), as a means of delivering fluoride systemically on a community basis, for preventing dental caries in children, adolescents and adults
-
To examine whether the effectiveness of fluoridated salt on caries prevention is influenced by types of salt fluoridation schemes adopted and other particular intervention features, by initial caries levels and other relevant population characteristics, geographical regions/climates, and by exposure to other non‐systemically delivered fluorides (such as toothpastes)
-
To evaluate the effectiveness of fluoridated salt compared with water or milk fluoridation (differential effect).
Methods
Criteria for considering studies for this review
Types of studies
Randomised (or quasi‐randomised) controlled trials (RCTs) which may be designed with randomisation at the level of the individual, or at the group, school or community level (cluster), and non‐randomised studies with a concurrent control group only; i.e., any prospective study comparing at least two groups/communities concurrently, one receiving fluoridated the other non‐fluoridated salt, with at least two points in time evaluated (e.g. non‐randomised trials, controlled before‐and‐after studies, cohort studies).
Eligible studies should also have an intervention and follow‐up time of a minimum of 3 years.
Study designs other than the ones above (e.g. single group pre/post‐ (before/after) designs, historical control studies, interrupted time series) with an intervention or follow‐up period of less than 3 years will be excluded.
See 'Additional Table 1' and 'Additional Table 2.'
| Study design | Description |
| Randomised/Quasi‐randomised controlled trial | A prospective study where participants are allocated to intervention or control groups and followed up over time to assess any differences in outcome rates (the outcomes of interest are compared between groups at the end of the follow‐up time). In a randomised trial appropriate methods are used for randomisation and concealment of allocation (e.g. use of a telephone or computer system to conceal randomisation, centrally controlled randomisation, pre‐numbered or coded identical containers administered serially to participants, sequentially numbered sealed opaque envelopes). It avoids selection bias because it ensures that on average both known and unknown determinants of outcome (prognostic factors) are randomly distributed between groups. In a quasi‐randomised trial the method of allocation falls short of genuine randomisation and allocation concealment, i.e. the allocation procedure is entirely transparent before assignment (e.g. allocation by date of birth, hospital record number, alternation, odd‐even numbers, patient social security numbers, days of the week, etc). |
| Non‐randomised trial | A prospective study where the investigator has control over the allocation of participants to groups, but does not attempt randomisation or quasi‐randomisation to allocate participants (e.g. patient or physician preference, patient characteristics and clinical history). |
| Controlled before‐and‐after study | A non‐randomised prospective study where a control population of similar characteristics and performance as the intervention group is identified. Data are collected in both the control and intervention groups before (at baseline) and after the intervention, and either final values (if the groups are comparable at baseline) or change from baseline values are compared. The before and after measurements may be made in the same participants or in different samples. An example of the before‐after design: a study in which 2 groups of 12‐year olds from 2 similar populations were examined for prevalence of caries prior to initiating salt fluoridation in 1 of the groups, and 5 years after starting fluoridation 12 year olds are examined in the 2 areas (1 fluoridated, the other not) and caries in the 2 groups/areas are compared. It is important to note that here the children are different in the before and after periods. |
| (Concurrent) Cohort study | A follow‐up study that compares outcomes of participants in subsets of the cohort ‐ to examine people who were exposed or not exposed (or exposed at different levels) to a particular intervention (or other factor of interest). Participants are studied during the same (concurrent) period prospectively or retrospectively. A prospective cohort study assembles participants and follows them into the future. A retrospective cohort study identifies subjects from past records and follows them from the time of those records to the present. An example of a prospective cohort study: 2 groups of children (subsets of a cohort) are examined at age 12 prior to the initiation of fluoridation of salt for 1 group, and these same children are followed up with annual examinations for caries for 4 years. |
| Study design | Description |
| Historical control study | A variation on the traditional cohort study where the outcome from a new intervention/exposure is established for participants studied in 1 period and compared to those who did not received the intervention in a previous period, i.e. participants are not studied concurrently as 1 of the groups is taken from the past. |
| Case‐control study | Participants with and without a given outcome are identified (cases and controls respectively) and exposure to a given factor(s)/intervention(s) are compared between the 2 groups (it is investigated how exposure to suspect agents differed between the 2 groups). |
| Before‐and‐after study (single group pre/post‐design) | In this type of study there is a comparison of outcomes from study participants before and after they have received an intervention. The before and after measurements may be made in the same participants or in different samples. There is no control group for comparison. |
| Interrupted time series | Interrupted time series (ITS) studies collect observations at multiple time points before and after an intervention (interruption). The design attempts to detect whether the intervention has had an effect significantly greater than the underlying trend. The intervention effect is measured against the pre‐intervention trend. There is no way to assess the impact of any concurrent events on the outcomes of interest. |
| Cross‐sectional study | A study examining the prevalence of a condition/exposure and/or the relationship between a condition and other variables of interest/other characteristics as they exist in a defined population at a particular point in time. |
Types of participants
General population ‐ Children, adolescents and adults, irrespective of initial level of dental caries, background exposure to fluorides, dental treatment level, nationality, etc.
Studies where participants/populations were selected on the basis of special (general or oral) health conditions will be excluded.
Types of interventions
Active intervention: fluoridated salt only (domestic salt or all salt available for human consumption), of any type (F agent), at any level (ppm F), or amount used, provided the change in the level of fluoride in the salt of at least one of the study groups/communities took place within 3 years of the baseline survey.
Control: non‐fluoridated salt (placebo or no treatment), or fluoridated water or milk.
Studies will be excluded where the active intervention consisted of any other systemically delivered fluoride (e.g. water, milk, tablets) provided in addition to fluoridated salt.
Types of outcome measures
Primary: For permanent and deciduous dentition, changes in caries experience and caries increment, as measured by changes from baseline in decayed, filled, missing figures at the surface (dmfs/DMFS) and tooth (dmft/DMFT) level, and final caries values (if the groups are comparable at baseline).
Caries will be assessed clinically at the dentine level of diagnosis. However, if a combined clinical and radiological assessment is available it will be considered.
Secondary: Any other measures of dental caries such as proportion of children developing new caries, changes in caries‐free subjects at specific ages, tooth loss, dental pain.
Adverse effects: Dental fluorosis (where reporting of enamel fluorosis will be assessed according to established specific indices used (e.g. Dean, Thylstrup, Horowitz)), skeletal fluorosis, incompatibility/adverse reactions (possibly due to other halogens added to salt or adverse impact on disease prevention by other salt additive), bone effects (fractures or dyscrasias), cancer, high blood pressure, heart disease, mortality and any other possible negative effects. We anticipate that without the inclusion of observational study designs, such as case‐control studies, a full investigation of adverse effects may not be possible, and further consideration shall be given to this aspect in the first update of this review.
Costs/cost‐effectiveness of intervention is also considered a relevant outcome, as well as changes in salt consumption.
Studies reporting only on fluoride or salt/sodium intake, changes in plaque/calculus formation, plaque regrowth/vitality, plaque/salivary bacterial counts, or gingival bleeding/gingivitis, dentine hypersensitivity or fluoride physiological outcome measures (fluoride uptake by enamel or dentine, salivary secretion levels, etc.) will be excluded.
Search methods for identification of studies
A comprehensive search will be undertaken to locate all relevant studies, and articles in any language will be considered. It involves searching a wide range of databases and other sources for the identification of relevant studies.
Electronic database search
A preliminary literature search was undertaken to provide background information on reviews/possible systematic reviews of salt fluoridation and on the scope of the literature in the field. Checks for existing and ongoing reviews were made with the Cochrane Database of Systematic Reviews (CDSR), the Database of Abstracts of Reviews of Effects (DARE), the NHS Economic Evaluation Database (NHS EED), MEDLINE and EMBASE. These searches were performed without any date limits, and were supplemented by contact with key researchers in the field. The preliminary MEDLINE search strategy (performed via OVID, and adapted to search EMBASE) combined a subject search with phases 1 to 3 of the Cochrane Sensitive Search Strategy for Randomised Controlled Trials (RCTs) (Appendix 5b in the Cochrane Handbook for Systematic Reviews of Interventions 4.2.6 (updated September 2006)) with relevant amendments.
For the actual identification of studies considered for this review only a refined subject search strategy is used, without methodological filters. The subject search, to be further developed for EMBASE, will also be adapted for use in the Cochrane Oral Health Group Trials Register, in the Cochrane Central Register of Controlled Trials (CENTRAL), and other databases (revised appropriately for each database to take account of differences in controlled vocabulary and syntax rules).
The revised MEDLINE/OLDMEDLINE subject search, performed via OVID, is published in full (seeAppendix 1).
Details of search strategies to be applied to other databases are available from the contact review author. All databases are searched from date of inception, and the full list of databases considered for this review follows:
Cochrane Oral Health Group Trials Register (to present)
Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library, current issue)
MEDLINE (1966 to present)
OLDMEDLINE (1950 to 1965)
EMBASE (1980 to present)
BIOSIS (1985 to present)
Dissertation Abstracts (1981 to present)
SIGLE (1980 to present)
LILACS/BBO (1982 to present)
PAHO/WHOLIS/MedCarib (1982 to present).
Language
The search will attempt to identify all relevant studies irrespective of language. Non‐English papers will be translated by members of the review group (members of the group have capability in French, German, Portuguese, Spanish) or relevant data extracted by members of The Cochrane Collaboration or specific translators for any other languages (such as Hungarian).
Unpublished studies
Organisations, researchers and experts known to be involved in the field will be contacted in an effort to trace unpublished or ongoing studies.
Handsearching
The following journals and publications have been identified as being important to be handsearched for this review for the period 1945 (or journal start date if later) to 2006:
Acta Odontologica Scandinavica
British Dental Journal
Caries Research
Community Dental Health
Community Dentistry and Oral Epidemiology
Journal of the American Dental Association
Journal of Dental Research
Journal of Public Health Dentistry
Boletin de la Oficina Sanitaria Panamericana
Fluoruracion al Dia
Documents of Ministries of Health (Jamaica, Costa Rica, Mexico, Peru, Venezuela, Uruguay, Belize, Bolivia)
Revista de la Federacion Odontologica Colombiana
Revista Facultad de Odontologia Universidad de Antioquia
International Dental Journal
Jamaican Dental Journal
Swiss Dental Journal
WHO and PAHO resolutions and documentation
Hungarian Dental Journal
Documents of World Salt Federation, EUROSALT, American Salt Institute
Revista Cubana de Estomatologia
Nutrition Guidelines
Archives Oral Biology
Health Conditions in the Americas
PAHO and WHO publications
WHO Bulletin and International Journal of Public Health
FDI World Dental Federation publications
Specific author books and monographs
Schweiz Monatsschrift Zahnmed
Helvetica Odontologica Acta.
The review authors will undertake to search these journals and documents. If journals from this list are being handsearched as part of the Cochrane Oral Health Group's handsearching programme, the review authors will search as necessary to complete the search period. A full list of journals being handsearched by the Cochrane Oral Health Group can be accessed from their website at http://www.ohg.cochrane.org/.
Reference lists searching
The reference lists of all eligible trials and review articles will be checked for additional studies. Reference lists of relevant books/book chapters will also be checked.
Data collection and analysis
Management of records
The set of records identified from each database, will be imported to the bibliographic software package Reference Manager and merged into one core database to remove duplicate records and to facilitate retrieval of relevant articles. References that cannot be downloaded in this way from databases and other searches will be entered manually.
Study selection
The titles and abstracts (when available) of all reports identified through the searches will be scanned independently by two review authors. Records that are obviously irrelevant will be discarded.
Records will be rejected on initial screen (considered irrelevant) according to study design/duration, or interventions/comparisons/outcome: when it can be determined that the article is not a report of a controlled trial or of a prospective intervention study/trial with a concurrent control group (non‐randomised follow‐up study where a control population of similar characteristics and performance as the intervention group is identified and data are collected before (at baseline) and after the intervention in the control and intervention groups); or the trial/intervention is clearly of less than 3 years duration; or did not address salt fluoridation for caries prevention; or the trial did not compare salt fluoridation to placebo, no treatment, or to water or milk fluoridation.
Full reports will be obtained for trials appearing to meet the inclusion criteria or for which there is insufficient information in the title and abstract to make a clear decision. The full reports obtained from all the electronic and other methods of searching will be assessed independently, by two review authors to establish whether the trials meet the inclusion criteria or not. Disagreements will be resolved by discussion. Where resolution is not possible, a third review author will be consulted.
For reports thought to be potentially relevant in languages not known by the review authors, translation will be attempted and the inclusion form will be completed with reference to the translator. Attempts will also be made to contact authors of articles that cannot be classified in order to ascertain whether inclusion criteria are met. It is considered essential to identify and check all reports related to the same study; and in case of any discrepancy, authors will be contacted.
All studies meeting the inclusion criteria will then undergo a validity assessment and data extraction. Studies rejected at this or subsequent stages will be recorded as excluded, and reasons for exclusion will be noted.
Quality assessment
The quality assessment of included trials/studies will be undertaken independently and in duplicate by two review authors as part of the data extraction process. A standardised quality assessment form will be developed specifically for this review. The form will include two separate sections ‐ one for randomised controlled trials (RCTs)/quasi‐randomised trials and one for non‐randomised concurrent control studies.
Included trials will be assessed on the following key criteria:
method of randomisation (A = adequate, B = unclear, C = inadequate); concealed allocation of fluoride addition to salt (A = adequate, B = unclear, C = inadequate); blinding of patients (0 = no, 1 = yes, 2 = unclear), providers, and outcome assessors (0 = no, 1 = yes, 2 = indicated/unclear); completeness of information on drop outs by trial group (0 = no, 1 = yes), and on any reasons for withdrawal by trial group (0 = no, 1 = yes).
The agreement between the review authors will be assessed by calculating the Kappa score.
The following key methodological aspects will be considered if/when assessing non‐randomised studies for inclusion.
-
Systematic differences in measurement of intervention/exposure (performance bias) ‐ how (reliably) was salt fluoridation status ascertained (fluoride level reliably measured).
-
Systematic differences in outcome assessment (detection bias) ‐ how (reliably) were relevant outcomes (e.g. caries, fluorosis) ascertained, blindness of examiners to salt fluoridation status.
-
Systematic differences in follow up (attrition bias) ‐ groups followed up for the same/adequate time, proportion of participants in all groups included in the final analysis, description of those not included not suggestive of bias.
-
Systematic differences in comparison groups (selection bias/control of/adjustment for confounding) ‐ similarity/comparability of groups at baseline, pre‐specified confounding factors addressed (measured)/adjusted for in the analysis (a list of potential confounding factors will be pre‐specified by the review team, and it will be noted whether they are demonstrated to be similar in both groups, or adjusted for in the analysis).
The criteria above include appraisal of internal validity and biases relevant to non‐randomised studies in general, and specific to fluoridated salt caries prevention studies.
Data extraction
Data will be extracted by two review authors independently using a specially designed data extraction form, to be piloted on several papers and refined as required for use. Data presented only in graphs and figures will be extracted whenever possible, but will be included only if two review authors independently have the same result. Attempts will be made to contact authors in order to obtain missing information or for clarification whenever necessary. Papers in languages not known by the review authors will be data extracted with help from appropriate translators. Review authors will not be blinded to the names of the authors, institutions, journal of publication or results of the studies.
From each included study, data will be extracted on details of study (including author(s), published or unpublished, year of publication, number of reports on the study; study design, year in which study began, years when it was conducted/study duration, time of baseline survey in relation to initiation of salt fluoridation, place/region/climate where study was conducted (country), reasons for termination or suspension of salt fluoridation, sponsoring institutions/organisations/manufacturers involved); methodological quality (as described above, with additional information on any co‐intervention or contamination or both); details of participants (including number, age (mean/range), and caries severity at start, education, SES, setting where participants were recruited, exposure to other non‐systemic fluoride interventions); characteristics of intervention (including, methods/schemes to deliver the intervention, fluoride levels, fluoride agents used); details of primary and secondary outcomes* (caries/fluorosis data, etc.); details of analysis (including reporting of measures of effect, confidence intervals, crude/adjusted results, etc.).
*Being aware that caries (and fluorosis) data can be reported differently in different trials we will develop a set of a priori rules to choose relevant outcome data for extraction/analysis from each study. We will also report any adverse events associated with salt fluoridation if recorded in the studies.
The details of the included studies' methods, participants, interventions, outcomes, and analyses/results, as well as all assessments of study quality will be presented in tables. All such assessments will be coded for possible use in metaregression or sensitivity analyses.
Data synthesis and statistical methods
For an outcome such as caries increment (treated as continuous data) the chosen measure of treatment effect is the prevented fraction (PF), that is mean increment in the controls minus mean increment (Higgins 2003) in the treated group divided by mean increment in the controls. This measure is considered more appropriate than the mean difference or standardised mean difference, since it allows combination of different ways of measuring caries increment and a meaningful investigation of heterogeneity between trials. It is also simple to interpret. The meta‐analyses are to be conducted as inverse variance weighted averages and random‐effects meta‐analyses will be performed using RevMan software.
For outcomes other than caries increment, continuous data are to be analysed according to differences in mean treatment effects and their standard deviations (e.g. mean difference of change in percentage caries free and mean difference in (final) DMFS/T). Dichotomous outcome data (e.g. proportion of children without caries, dichotomization of fluorosis indices at a given threshold) are to be analysed by calculating risk ratios (RR) or, especially for adverse effects of fluoride treatment, risk differences (RD). RevMan will be used for estimation of overall treatment effects. Again, a random‐effects model is to be used to calculate a pooled estimate of effect together with the 95% confidence intervals.
Formal tests for heterogeneity will be used to provide an indication of between‐study heterogeneity. In addition, the degree of heterogeneity observed in the results will be quantified using the I2 statistic (Higgins 2003), which can be interpreted as the percentage of variation observed between the studies caused by between‐study differences rather than chance.
Studies will be presented separately by study design; deciduous and permanent teeth will be also analysed separately.
As a general rule only (relevant) outcomes with useable data will be shown in the analyses tables.
Metaregression or subgroup analysis or both will be used to explore the influence of pre‐specified study characteristics (see relevant questions/objectives) on outcome in an attempt to try to explain any heterogeneity between studies (if sufficient number of studies is included). Stata version 9.0 (Stata Corporation, US) will be used for metaregression analyses. Should sufficient data/number of data points be recorded, further investigation regarding the association between salt fluoride concentration and the proportion of the population with dental fluorosis will be considered. Sensitivity analyses may also be undertaken to examine the effect of key methodological quality aspects on the overall estimates of effect. Any such 'post hoc' analyses will be clearly identified.
A funnel plot (plots of effect estimates versus the inverse of their standard errors) is to be drawn. Asymmetry of the funnel plot may indicate publication bias and other biases related to sample size, though may also represent a true relationship between trial sizes and effect size.
| Study design | Description |
| Randomised/Quasi‐randomised controlled trial | A prospective study where participants are allocated to intervention or control groups and followed up over time to assess any differences in outcome rates (the outcomes of interest are compared between groups at the end of the follow‐up time). In a randomised trial appropriate methods are used for randomisation and concealment of allocation (e.g. use of a telephone or computer system to conceal randomisation, centrally controlled randomisation, pre‐numbered or coded identical containers administered serially to participants, sequentially numbered sealed opaque envelopes). It avoids selection bias because it ensures that on average both known and unknown determinants of outcome (prognostic factors) are randomly distributed between groups. In a quasi‐randomised trial the method of allocation falls short of genuine randomisation and allocation concealment, i.e. the allocation procedure is entirely transparent before assignment (e.g. allocation by date of birth, hospital record number, alternation, odd‐even numbers, patient social security numbers, days of the week, etc). |
| Non‐randomised trial | A prospective study where the investigator has control over the allocation of participants to groups, but does not attempt randomisation or quasi‐randomisation to allocate participants (e.g. patient or physician preference, patient characteristics and clinical history). |
| Controlled before‐and‐after study | A non‐randomised prospective study where a control population of similar characteristics and performance as the intervention group is identified. Data are collected in both the control and intervention groups before (at baseline) and after the intervention, and either final values (if the groups are comparable at baseline) or change from baseline values are compared. The before and after measurements may be made in the same participants or in different samples. An example of the before‐after design: a study in which 2 groups of 12‐year olds from 2 similar populations were examined for prevalence of caries prior to initiating salt fluoridation in 1 of the groups, and 5 years after starting fluoridation 12 year olds are examined in the 2 areas (1 fluoridated, the other not) and caries in the 2 groups/areas are compared. It is important to note that here the children are different in the before and after periods. |
| (Concurrent) Cohort study | A follow‐up study that compares outcomes of participants in subsets of the cohort ‐ to examine people who were exposed or not exposed (or exposed at different levels) to a particular intervention (or other factor of interest). Participants are studied during the same (concurrent) period prospectively or retrospectively. A prospective cohort study assembles participants and follows them into the future. A retrospective cohort study identifies subjects from past records and follows them from the time of those records to the present. An example of a prospective cohort study: 2 groups of children (subsets of a cohort) are examined at age 12 prior to the initiation of fluoridation of salt for 1 group, and these same children are followed up with annual examinations for caries for 4 years. |
| Study design | Description |
| Historical control study | A variation on the traditional cohort study where the outcome from a new intervention/exposure is established for participants studied in 1 period and compared to those who did not received the intervention in a previous period, i.e. participants are not studied concurrently as 1 of the groups is taken from the past. |
| Case‐control study | Participants with and without a given outcome are identified (cases and controls respectively) and exposure to a given factor(s)/intervention(s) are compared between the 2 groups (it is investigated how exposure to suspect agents differed between the 2 groups). |
| Before‐and‐after study (single group pre/post‐design) | In this type of study there is a comparison of outcomes from study participants before and after they have received an intervention. The before and after measurements may be made in the same participants or in different samples. There is no control group for comparison. |
| Interrupted time series | Interrupted time series (ITS) studies collect observations at multiple time points before and after an intervention (interruption). The design attempts to detect whether the intervention has had an effect significantly greater than the underlying trend. The intervention effect is measured against the pre‐intervention trend. There is no way to assess the impact of any concurrent events on the outcomes of interest. |
| Cross‐sectional study | A study examining the prevalence of a condition/exposure and/or the relationship between a condition and other variables of interest/other characteristics as they exist in a defined population at a particular point in time. |
