The distribution of burden of dental caries in schoolchildren: a critique of the high-risk caries prevention strategy for populations
© Batchelor and Sheiham; licensee BioMed Central Ltd. 2006
Received: 16 August 2005
Accepted: 31 January 2006
Published: 31 January 2006
The 'high-risk approach' is a commonly adopted strategy recommended for the prevention of dental caries in populations. The scientific basis for the strategy has been questioned. The objective of this study is to assess the contribution that children identified at 'high-risk' made towards the total of new caries lesions over a 4-year period, by analysing the distribution of new lesions per 100 children.
Data are from the National Preventive Dentistry Demonstration Programme (NPDDP) in the United States. The analyses identified the distribution of new carious lesions over a 4-year period in four groups of 7 year-old children who received differing preventive regimes.
The majority of new lesions occurred in those children classified at lowest caries risk at baseline. Irrespective of the preventive regime adopted and the initial caries levels, children classified as 'highest risk' contributed less than 6% of the total number of new lesions developing over 4 years.
These findings challenge the basis for the adoption of a high-risk strategy.
A commonly adopted approach for the prevention of caries is the 'high-risk' strategy. For example, Messer  concluded that "... the need for, targeted prevention of dental caries for those at high risk has become apparent". The approach is based on three assumptions: first, those individuals with high future caries increments can be identified; second, measures taken to prevent the caries lesions are effective, and third, that those individuals belonging to sub-groups within a population who have previously experienced the highest levels of caries in the past will continue to do so in future. Indeed, any high-risk strategy aims to target those individuals at the greatest risk of future disease based on their current caries status or markers of disease.
From a public health perspective, what is important when deciding upon a preventive approach is what impact the measure adopted would have on the total dental health and disease burdens of the population as a whole. Even if an approach was highly accurate in predicting future caries development and the intervention was relatively successful in reducing caries in that group, the distributive features of new disease may make the proposed approach inappropriate. Even if a high-risk group had a far higher annual increment than the remaining sub-groups of the population, due to the underlying distributive properties of caries within a population, a far greater number of lesions are likely to occur in the low risk individuals because there are more of them. Batchelor and Sheiham  referred to this issue when examining caries distributions within a population. They outlined the limitations of adopting a 'high-risk' approach for the prevention of caries highlighting that any changes in the average caries experience within the population were not limited to specific sub-groups but occurred throughout the population. In addition, they found a mathematical relationship between the mean DMF score for a population and the prevalence of caries within that population. For a given mean DMF score the prevalence within the population could be defined and vice versa. The relationship was independent of both age and fluoride levels .
The problems of a high-risk strategy are increased by the low accuracy of methods used to identify the high-risk children. While the idea that an individual's future caries increment can be predicted from their past caries experience underpins the basis of caries risk assessment, the rigour of these measures to date is poor. Powell  and Zero et al.  reviewing the literature covering the use of indicators of risk found that the predictive validity of the models were heavily dependant upon the prevalence of caries and the characteristics of the population for which they were designed. Zero et al.  found that a single indicator gave as good results as more complex combinations of variables. This finding is in agreement with van Palenstein Helderman et al.  who, using longitudinal study data examining past caries variables, found that the gain in accuracy by including additional predictor variables was limited. Irrespective of their complexity, no predictive model is able to identify those individuals who will get the highest future caries increments. Hausen et al.  and Hausen  have highlighted the limitations of current methods used to identify high-risk individuals. Furthermore, even at a population level Poulsen and Scheutz  also recognised that a high-risk strategy could be challenged on the grounds of effectiveness. Examining the changes in dental caries experience in Danish children and adolescents over a ten-year period, they concluded that, if adopted, a high-risk strategy that was 40% effective would reduce the mean DMFS for the whole population by a mean of only one surface.
A major shortcoming of the high-risk approach is the failure to examine its impact on the overall number of new lesions within a population. Does a strategy targeting the high-risk group prevent more lesions for a population than a whole population strategy? Do 'low risk' sub-groups develop less new lesions than those with high caries levels? To answer the question, an analysis is required of the distribution of new caries lesions within a population as the baseline levels of caries increase.
Our critique of the high-risk strategy which we shall demonstrate in this paper is based on the concept that the largest "... burden of ill health comes more from the many who are exposed to low inconspicuous risk than from the few who face an obvious problem" . The aim of this study is to assess whether 'high risk groups' of children accounted for a high percentage of new caries lesions in children. The objective was to analyse the distribution of new lesions per 100 7-year-old children in four populations with differing initial caries levels over a 4-year period.
The data used were derived from the National Preventive Dentistry Demonstration Programme (NPDDP). The programme was an extensive project aimed at determining the costs and benefits of various types and combinations of school-based preventive dental care procedures. The project ran from 1976 to 1983, with the data used here being collected over the four-year period 1977–81. The project was conducted in 10 communities in the United States, 5 non-fluoridated and 5 fluoridated. The preventive regimes included fissure sealants, topical fluoride rinses, fluoride tablets, and school based oral health education programmes. The subjects in the trial were children in grades 1, 2 and 5 in the participating schools. From those selected as eligible to participate in the study, 82% responded positively. The data used in this study cover only those children. No attempt was made to assess the impact of non-respondents as the aim of the present study is fundamentally different from that of the original project. The background, organisation and results of the NPDDP programme have been reported elsewhere [10–13].
The children were categorised according to age, geographical location and preventive programme. The minimum number in each group exceeded 500 children. The mean 4-year caries increment was recorded for each of the sub-groups created according to the child's water fluoridation and sealant status. Group 1, formed by children who did not receive fluoridated water but who received sealants had an initial DMF-S score of 0.61; Group 2, whose water supply was not fluoridated and did not receive sealants, a DMF-S of 0.85; Group 3, who lived in fluoridated water supply areas but did not receive sealants, a DMF-S = 0.86 and Group 4, fluoridated and sealants, a DMF-S = 0.95. For the purposes of this study the four groups were followed over a 4-year period to assess the distribution patterns of new caries lesions. The data were analysed using SPSS Version 11 for Mac OS X.
For Group 3 (baseline DMF-S = 0.86), 67% had an initial DMF-S of 0, 19% 1 or 2, 8% 3 or 4, 4% 5 or 6 and 7% an initial DMF-S of 7 or more. For Group 4 (baseline mean DMF-S = 0.95) 65% had a DMF-S of 0, 20% a DMF-S of 1 or 2, 7% 3 or 4, 5% 5 or 6 and 3% a DMF-S greater than 7. No children in any of the groups had a DMF-S score of 9 or more.
Figures 1, 2, 3, 4 also show the mean caries increment for each of the sub-groups according to initial caries levels. For example, for those children with a DMF-S of 0 at the start of the 4-year programme in Group 1, the 4-year increment was 0.81, while for Group 4, the increment was 1.08. The highest increments were in Groups 2 and 3. For those children in Group 3 with an initial DMF-S of 5 or 6, the mean 4-year increment was 4.54
These findings challenge the fundamental arguments used to justify the adoption of a high-risk strategy from a public health perspective. Namely, the main strategy should be directed at the smallish group with high baseline caries levels because they are at highest risk of future caries. As stated by Rose  when applied to general health and by Batchelor and Sheiham  for caries, the largest "... burden of ill health comes more from the many who are exposed to low inconspicuous risk than from the few who face an obvious problem". Burt , reviewing the concepts of risk as applied to dental public health, reinforced this approach arguing that the geographic targeting of caries preventive programmes should supplement population based approaches. Geographic targeting is one form of a directed population approach: an approach that uses socio-demographic or epidemiologic data to identify groups as opposed to screening for individuals who may benefit from the intervention .
Irrespective of the preventive regimes adopted, for all sub-groups, the majority of the new lesions were in those children who would have been classified as at lowest risk. Indeed, with the exception of those individuals in non-fluoridated areas and receiving sealants, over 50% of the total number of new lesions occurred in individuals with an initial DMF-S score of 0. For all preventive regimes, the contribution made by individuals with the highest initial grouped DMF-S score, those with 7 or more lesions, was less than 6% of the total new caries over a 4 year period. For example, for those individuals receiving fluoridated water and sealants, the contribution to the total number of lesions made by those with an initial DMF-S of 7 or more was less than 2%.
The present study has shown that from a public health perspective, policy for caries preventive strategies should be based on a 'population' or 'directed population' approach.
- Messer LB: Assessing caries risk in children. Aust Dent J. 2000, 45 (1): 10-16.View ArticlePubMedGoogle Scholar
- Batchelor PA, Sheiham A: The limitations of a 'high-risk' approach for the prevention of dental caries. Community Dent Oral Epidemiol. 2002, 30 (4): 302-312. 10.1034/j.1600-0528.2002.00057.x.View ArticlePubMedGoogle Scholar
- Powell LV: Caries prediction: a review of the literature. Community Dent Oral Epidemiol. 1998, 26 (6): 361-371.View ArticlePubMedGoogle Scholar
- Zero D, Fontana M, Lennon AM: Clinical applications and outcomes of using indicators of risk in caries management. J Dent Educ. 2001, 65 (10): 1126-1132.PubMedGoogle Scholar
- van Palenstein Helderman WH, van't Hof MA, van Loveren C: Prognosis of caries increment with past caries experience variables. Caries Res. 2001, 35 (3): 186-192. 10.1159/000047454.View ArticlePubMedGoogle Scholar
- Hausen H: Caries prediction. Dental Caries. The disease and its clinical management. Edited by: Fejerskov O, Kidd E. 2003, Oxford: Blackwell Munksgaard, 393-411.Google Scholar
- Hausen H: Caries prediction-state of the art. Community Dent Oral Epidemiol. 1997, 25 (1): 87-96.View ArticlePubMedGoogle Scholar
- Poulsen S, Scheutz F: Dental caries in Danish children and adolescents 1988–1997. Community Dent Health. 1999, 16 (3): 166-170.PubMedGoogle Scholar
- Rose G: The Strategy of Preventive Medicine. 1993, Oxford: Oxford University Press, SecondGoogle Scholar
- Bell RM, Klein SP, Bohannan HM, Graves RC, Disney JA: Results of the baseline dental exams in the National Preventive Dentistry Demonstration Program. 1982, Santa Monica: Rand CorporationGoogle Scholar
- Robert Wood Johnson Foundation: National Preventive Dentistry Demonstration Program; special report. 1983, Princeton, NJ: Robert Wood Johnson FoundationGoogle Scholar
- Klein SP, Bohannan HM, Bell RM, Foch CB, Disney JA, Graves RC: The cost and effectiveness of school-based preventive dental care. Am J Public Health. 1985, 75: 382-391.View ArticlePubMedPubMed CentralGoogle Scholar
- Disney JA, Bohannan HM, Klein SP, Bell RM: A case study in contesting the conventional wisdom: school based fluoride mouthrinse programs in the USA. Community Dent Oral Epidemiol. 1990, 18: 46-56.View ArticlePubMedGoogle Scholar
- Burt B: Concepts of risk in dental public health. Community Dent Oral Epidemiol. 2005, 33 (4): 240-247. 10.1111/j.1600-0528.2005.00231.x.View ArticlePubMedGoogle Scholar
- Watt RG: Strategies and approaches in oral disease prevention and health promotion. Bull World Health Org. 2005, 83: 711-718.PubMedPubMed CentralGoogle Scholar
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1472-6831/6/3/prepub