Scholarly article on topic 'Factors related to dental health in 12-year-old children: a cross-sectional study in pupils'

Factors related to dental health in 12-year-old children: a cross-sectional study in pupils Academic research paper on "Economics and business"

CC BY-NC-ND
0
0
Share paper
Academic journal
Gaceta Sanitaria
OECD Field of science
Keywords
{Children / "Dental knowledge" / "Health practice" / "Oral health surwey" / Spain / Niños / "Conocimientos dentales" / "Práctica odontológica" / "Encuesta sobre salud bucal" / España}

Abstract of research paper on Economics and business, author of scientific article — Ernesto Smyth, Francisco Caamaño

Abstract Objective The aim of this study was to identify factors related to the prevalence of caries in 12-year-old schoolchildren. Methods A cross-sectional study was carried out using a representative sample (n=1217) of the population of 12-yearold schoolchildren in Galiza (northwest Spain). Independent variables were measured through a questionnaire, and dependent variables were determined through oral examination. Multiple and logistic regression were applied. Results The decayed, missing and filled permanent teeth/decayed, filled primary teeth (DMFT-dft) value in the sample was 1.83 (95% confidence interval [CI], 1.67-1.98), the DMFT value was 1.53 (95% CI, 1.37-1.67), and the prevalence of caries was 61% (95% CI, 57.7-64.5). The prevalence of caries was directly related to a low frequency of brushing, greater use of toothpaste, and a higher consumption of sweets. The prevalence of caries was higher in rural than in urban areas. In contrast, the higher the mother's level of education and the greater the subject's knowledge of dental health, the lower the prevalence of caries. Conclusions The main goals of dental health programmes should be to achieve quality brushing every day in children, to reduce the consumption of sweets, and to increase knowledge of dental health.

Academic research paper on topic "Factors related to dental health in 12-year-old children: a cross-sectional study in pupils"

ORIGINALES

Factors related to dental health in 12-year-old children: a cross-sectional study in pupils

Ernesto Smyth / Francisco Caamano

Department of Preventive Medicine and Public Health. University of Santiago de Compostela. Santiago de Compostela. A Coruna. Spain.

(Factors related to dental health in 12-year-old children: a cross-sectional study in pupils)

Abstract

Objective: The aim of this study was to identify factors related to the prevalence of caries in 12-year-old schoolchildren.

Methods: A cross-sectional study was carried out using a representative sample (n = 1217) of the population of 12-year-old schoolchildren in Galiza (northwest Spain). Independent variables were measured through a questionnaire, and dependent variables were determined through oral examination. Multiple and logistic regression were applied.

Results: The decayed, missing and filled permanent teeth/decayed, filled primary teeth (DMFT-dft) value in the sample was 1.83 (95% confidence interval [CI], 1.67-1.98), the DMFT value was 1.53 (95% CI, 1.37-1.67), and the prevalence of caries was 61% (95% CI, 57.7-64.5). The prevalence of caries was directly related to a low frequency of brushing, greater use of toothpaste, and a higher consumption of sweets. The prevalence of caries was higher in rural than in urban areas. In contrast, the higher the mother's level of education and the greater the subject's knowledge of dental health, the lower the prevalence of caries.

Conclusions: The main goals of dental health programmes should be to achieve quality brushing every day in children, to reduce the consumption of sweets, and to increase knowledge of dental health.

Key words: Children. Dental knowledge. Health practice. Oral health surwey. Spain.

Resumen

Objetivo: Identificar los factores asociados a la prevalencia de caries en escolares de 12 años.

Métodos: Estudio transversal sobre una muestra (n = 1.217) de escolares de 12 años de Galicia. Las variables independientes se midieron mediante un cuestionario y las dependientes, a través de exploración bucal. En el análisis estadístico se aplicaron regresión logística y regresión lineal múltiple.

Resultados: El índice CAO-co en la muestra fue 1,83 (intervalo de confianza [IC] del 95%, 1,67-1,98), el índice CAO 1,53 (IC del 95%, 1,37-1,67), mientras que la prevalencia de caries se situó en el 61% (IC del 95%, 57,7-64,5). La prevalencia de caries estuvo directamente asociada a la baja frecuencia del cepillado, al mayor uso de pasta y al elevado consumo de golosinas. La prevalencia de caries determinada en el medio rural fue más alta que en el medio urbano. Por último, el mayor nivel de estudios de la madre y los mayores conocimientos sobre salud oral de los individuos se asocian a una menor prevalencia de caries.

Conclusiones: A pesar de la moderada afectación por caries, los principales objetivos de los programas de salud oral para este grupo de edad deben ser: lograr el cepillado dental diario, reducir el consumo de golosinas e incrementar los conocimientos sobre salud oral. Por último, los programas de salud oral deben estar dirigidos principalmente a la población del medio rural y de menor nivel sociocultural. Palabras clave: Niños. Conocimientos dentales. Práctica odontológica. Encuesta sobre salud bucal. España.

Correspondencia: Ernesto Smyth.

Departamento de Medicina Preventiva y Salud Pública.

Facultad de Medicina.

San Francisco, s/n. 15782 Santiago de Compostela.

A Coruña. España.

Correo electrónico: mrsmyth@usc.es

Recibido: 11 de mayo de 2004. Aceptado: 28 de octubre de 2004.

Introduction

Epidemiological studies about dental caries in schoolchildren are numerous1-3. However, many of these studies only analyse caries prevalence, as the prevalence rate of caries, or through the different caries ratios defined in the bibliography (DMFT, decayed, missing, filling in definitive tooth; dmft, decayed, missing, filling in temporary tooth; dft,

decayed, filling in temporary tooth)12. Dental caries is a disease in which cultural and hygienic habits are decisive, so prevalence found in different habitats and different moments could be strong related with these factors.

On the other hand, determining the factors associated with the appearance of caries is of greater interest, given that these factors present high geographical and temporal stability4. However, the number of articles that analyse these factors is lower, and despite there being studies on the factors associated with caries in 12 year olds5-7, studies that use multivariate methodology including cultural variables are scarce8, a method that allows us to isolate the contribution of each of the risk factors.

The objectives of this study were: to determine caries prevalence in 12-year-old pupils and to identify the factors related to caries prevalence.

Methods

Design and sample

This study has a cross-sectional design, and the study participants were 28 297 children aged 12 years old. The pool of schools was stratified by province (A Coruña, Lugo, Ourense, Pontevedra) and habitat (urban; and rural). Eight clusters were generated. A total of 95 schools were selected proportional to cluster size. Finally, each school had a number of sampled pupils proportional to its size. This sample is comprised of 1217 subjects.

Data collection

For data collection, six teams were created and each one of them were made up of two persons: 1 dentist and 1 assistant who administered out the questionnaire. Diagnosis criteria between the six teams were calibrated by a training of two weeks. The training was made in 4 schools.

The teams visited the schools during the second term of the year 2000. Before the team visited the schools, they were contacted to set dates and determine requirements to carry out the questionnaire and the exploration. Authorisation from the pupils' parents was requested.

Taking previous studies as a starting point, we collected socio-demographic and medical variables, which could be associated with dental caries, through the questionnaire. The questionnaire was designed to be short and easy to fill. Pupils were asked whether they considered they had a healthy mouth, and what pathology they had. Pupils were also queried about their beliefs

on health: whether they believed it was important to look after their teeth, if they thought that with age their teeth would be less healthy, and if they believed that sugar produces caries. Pupils were also queried about sweets' consumption.

In as far as their hygienic habits were concerned, the pupils were asked whether they cleaned their teeth habitually, if they used dental floss and if they used an electric toothbrush. Pupils were also asked at what age they had started cleaning their teeth, when was the last time they had cleaned their teeth, how many times they clean their teeth a day, when was the last time they changed their toothbrush, how much toothpaste they put on the toothbrush, and who showed them how to clean their teeth.

The pupils were also asked about their use of fluorinated toothpaste, fluorine tablets, fluorine drops, and fluorinated mouthwashes. Pupils were also queried about whether they had been to the dentist, and if the dentist had advised them to wash their teeth. The schoolchil-dren's knowledge was also measured on the usefulness of fluorine and on prevention of caries and gingivitis.

Dependent variables were obtained through mouth exploration of the schoolchildren carried out by the dentist. This information was registered using a modified exploration form9.

Independent variables

Pupils were also queried about their knowledge on dental health through 6 questions, giving 0 (incorrect) or 1 (correct) point to each answer. A variable with values between 0 and 10 was generated, given that some questions had a multiple answer. The questions considered were: a) sugar provokes caries, 0 = none, 0 = little, 1 = quite a lot or a lot; b) fluorine is good so that, 1 = teeth are more resistant, 0 = teeth are whiter, 0 = teeth are sparkler; c) caries is a disease, 1 = that destroys teeth, 0 = that makes your gums bleed, 0 = that gives a bad smell, 0 = in which your teeth get whiter; d) gingivitis is a disease, 0 = that destroys your teeth, 1 = that makes your gums bleed, 0 = that gives a bad smell, 0 = in which your teeth get whiter; e) I can avoid having caries, 1 = with hygiene, 1 = eating few sweets, 1 = using fluorine, 1 = going for check ups, and f) I can avoid gingivitis, 1 = with hygiene, 0 = eating few sweets, 0 = using fluorine, 1 = going for check ups.

In as far as their hygienic habits, the pupils were asked when they had cleaned their teeth last (today, yesterday, day before yesterday, or more than two days ago); and how much toothpaste they had put on the toothbrush (a third, two thirds, or complete).

Pupils were queried about their consumption of sweets and where they eat them habitually (doesn't eat them, at school, at home, with friends, in other situations). This

variable is part of the models as the number of situations in which they eat sweets (doesn't eat them, in 1 situation, in 2, in 3, or in 4). Pupils were also queried about visiting a dentist: when was the last time they had visited the dentist (never visited, more than 1 year ago, between 3 months and one year, less than 3 months ago). Finally, pupils were also queried about orthodontic treatment (yes/no).

One socio-economic variable was considered: mother's education (no education, primary, secondary, university). In addition, an ecological variable which measures the socio-economic habitat was considered: municipality (rural, urban).

Dependent variables

We defined 3 variables to measure caries affectation: 1. Presence of decayed tooth, missing pieces or with fillings due to caries, dichotomous variable (0 = no; 1 = yes); 2. DMFT-value; 3. DMFT-dft-value, average decayed teeth surface in temporary and definite pieces.

Data analysis

The weighted Cohen's Kappa was used to evaluate the concordance between gold standard (evaluation of specialist-professor) and the teams in four schools. A univariate (prevalence) analysis was performed. To analyse the factors related to DMFT-value and DMFT-dft-value (continuum variable) linear multiple regression was used. To study the factors related to caries presence we carried out a logistic regression analyses. Taking into account that our study is a cross-sectional study, odds ratios calculated are really prevalence odds ratio (POR).

According to the hypothesis, maximum models were generated. We excluded from the logistic models variables that had no effect and were not cofounders of the other independent variables (change in coefficients > 10%). We used the Hosmer-Lemeshow test to determine the goodness of fit of the models to the data10.

Results

Of the 1217 pupils in our sample, 1105 pupils participated in the study (90.8%). The caries prevalence in the studied population was 61.1% (95% confidence interval [CI], 57.7-64.5). The DMFT-dft value in the sample was 1.82 (95% CI, 1.67-1.98), and the DMFT value 1.52 (95% CI, 1.37-1.67). The kappa statistics for concordance ranged from 0.75 to 0.95.

Table 1. Description of subjects studied for the main dependent variables

95% CI

Decayed tooth by subject (temporary and definite, D-d) 1.03 (0.90-1.18)

DMFT-value (decayed, missing, filling in definitive tooth) 1.52 (1.37-1.67)

dft-value (decayed, filling in temporary tooth) 0.31 (0.26-0.35)

DMFT-dft 1.82 (1.67-1.98)

DMFM (first permanent molar) 1.23 (1.12-1.34)

Proportion of subjects with decayed pieces (D-d) (%) 42.6 (38.9-46.4)

Proportion subjects with missing pieces (M) (%) 3.09 (1.91-4.27)

Proportion of subjects with pieces with fillings (F-f) (%) 33.5 (29.9-37.1)

Proportion of subjects with caries 61.1 (57.7-64.5)

(prevalence caries, DMFT-dft) (%)

Table 1 shows the characteristics of the participants; in as far as the main dependent variables are concerned. Table 2 shows the distribution of subjects according to the different categories of the variables of knowledge.

Table 3 shows multiple regression models including the variables chosen for DMFT value and DMFT-dft value as outcomes. The DMFT-dft value and the DMFT value reduce with knowledge on dental health, and with educational level. On the other hand, low frequency in brushing teeth, the use of a lot of toothpaste, and the consumption of sweets are related to higher ratios. The ecological habitat variable (rural/urban) is significant, given that a rural habitat is associated with higher ratios.

Logistic regression model is shown in table 4. The model includes all the variables comprised in the table. This table also includes the description of the sample through independent variables and caries prevalence in different groups.

Discussion

The results of this study show that caries is directly related to a low frequency in brushing, the use of more toothpaste, and a higher consumption of sweets. The study has also shown that there is higher caries prevalence in rural habitats compared to urban habitats. On the other hand, the higher the mother's level of education is and the more knowledge on dental health the subjects have, the lower the caries prevalence.

According to the results of the study, subjects with low knowledge of dental health show more caries (POR = 1.32; 95% CI, 1.20-1.61) than those subjects with higher knowledge. Different studies have found that health education could reduce caries affectation, concluding that higher knowledge generates more positive attitudes that in turn generate healthier habits. On the other hand, the results of our study show that knowledge has

Table 2. Distribution of subjects according to the different categories of the variables of knowledge. The numbers shows the percentage of subjects that agree with the statement

A lot Quite a lot Little No DK/DAb

Sugar provokes caries 47.0a 43.2 8.4 1.0 0.5

Destroys tooth Makes your gums bleed Gives a bad smell Whitens your tooth DK/DA

Caries is a disease thatc 87.7a 13.5 14.6 1.0 9.3

Gingivitis is a disease thatc 2.7 30.6a 3.1 0.7 63.0

With hygiene Eating less candies Using fluor Going to reviews DK/DA

I can avoid having cariesc 79.8a 66.1a 39.8a 35.4a 2.7

I can avoid gingivitisc 20.5a 12.1 17.9 18.2a 61.0

Resistant White Shiny DK/DA

Fluor is for teeth to bec 75.3a 35.6 11.5 12.7

aAnswers that are considered as correct in the evaluation of knowledge. bDoesn't know, doesn't answer.

cQuestions with multiple answers. The percentages do not add up to 100%.

Table 3. Related factors with DMFT-dft value and DMFT-value in 12 year old. Linear multiple regression coefficients (Coef), confidence intervals (95% CI) and statistical significance (p-value)

DMFT-dft-value DMFT-value

Coef* 95% CI* p-value Coef* 95% CI* p-value

Dental health knowledgea -0.068 (-0.136 to -0.000) 0.049 -0.047 (-0.109 to 0.015) 0.135

Last time brushed teethb 0.245 (0.093-0.397) 0.002 0,261 (0.123-0.399) < 0.001

How much toothpaste usedc 0.251 (0.053-0.449) 0.013 0.201 (0.021-0.381) 0.029

Sweet consumptiond 0.142 (0.017-0.267) 0.026 0.164 (0.050-0.277) 0.005

Visits to the dentist5 -0.309 (-0.440 to -0.177) < 0.001 -0.271 (-0.391 to -0.152) < 0.001

Orthodontics1 -0.537 (-0.950 to -0.123) 0.011 -0.377 (-0.752to -0.002) 0.049

Mother's education5 -0.259 (-0.417 to -0.100) < 0.001 -0.207 (-0.351 to -0.063) 0.005

Habitath 0.664 (0.406-0.922) < 0.001 0.559 (0.325-0.793) < 0.001

•Adjusted for the other independent variables included in this table.

"Measurements on four question (0 = all questions wrongly answered or doesn't answer... 10 = all questions correctly answered).

b1 = today, 2 = yesterday, 3 = day before yesterday, 4 = more than two days ago.

c1 = a third of the toothbrush, 2 = two thirds, 3 = the toothpaste covers the toothbrush completely.

dFive possibilities don't have any, at school, with friends, in my house, others (0 = doesn't consume. 4 = consumes in all situations). e1 = less than 3 months, 2 = less than three months and less than a year, 3 = more than a year, 4 = I've never been to the dentist. f0 = no, 1 = yes.

g1 = without education, 2 = primary education, 3 = secondary education, 4 = university education. h0 = urban, 1 = rural.

an effect on its own, independently of being able to modify attitudes and habits.

The fact that the models have been adjusted for follow brushing guidelines, visits to the dentist and consumption of sweets indicates that with the same habits, the subjects with more knowledge on oral health show less caries. There are 2 possible explanations for this result: first, better knowledge is related to better brushing techniques, although when adjusting by the quantity of toothpaste used, part of this effect should be con-

trolled, and second, a certain amount of residual confounding cannot be ignored due to the misclassification introduced in the variables that measure the habits11.

In as far as the effect of brushing on caries is concerned, the results of our study are consistent with those found by different authors7. Therefore, considering the subjects who cleaned their teeth today as a reference category, among those who cleaned their teeth yesterday, we found more caries (POR = 1.48; 95% CI, 1.22-1.78), (POR = 1.57; 95% CI, 0.91-2.33) among those who brus-

Table 4. Factors related with carles presence

n* %* Prevalence odds ratio (POR)a,b(95% CI)

Knowledge on dental health

More knowledge (5-10 points) 340 (30.8%) 59.0 1c

Less knowledge (0-4 points) 761 (69.2%) 65.6 1.32 (1.20-1.61)

Last time you brushed your teeth

Today 687 (62.6%) 55.2 1c

Yesterday 294 (26.8%) 68.0 1.48 (1.22-1.78)

Day before yesterday 47 (4.3%) 68.2 1.57 (0.91-2.36)

More than two days ago 70 (6.4%) 81.4 1.60 (1.20- 2.28)

How much toothpaste do you use

A third 80 (7.3%) 53.8 1c

Two thirds 323 (29.4%) 59.1 1.45 (0.87-2.09)

The whole toothbrush 694 (63.3%) 62.7 1.52 (1.10-2.16)

Sweet consumption

Never 62 (5.7%) 37.1 1c

In 1 situation 419 (38.2%) 60.2 1.39 (0.88-2.22)

In 2 situations 377 (34.4%) 61.3 1.44 (0.95-2.30)

In 3 situations 171 (15.6%) 66.7 1.46 (1.02-2.41)

In 4 situations 76 (6.9%) 72.4 1.68 (1.06-2.86)

Visits to the dentist

Never 129 (11.7%) 49.6 1c

> 1 year 229 (20.7%) 59.0 1.74 (1.27-5.83)

Between 3 months and 1 year 335 (30.3%) 63.1 1.50 (1.12-2.13)

< 3 months 411 (37.2%) 64.2 1.48 (1.05-2.06)

Do you have orthodontic treatment

Yes 126 (11.4%) 53.2 1c

No 979 (89.6%) 62.1 1.73 (1.24-2.10)

Mother's education

University 161 (16.4%) 49.7 1c

Secondary 253 (25.8%) 58.9 1.60 (1.32-2.92)

Primary 514 (52.4%) 63.2 1.40 (1.01-1.93)

Without education 52 (5.3%) 78.8 1.40 (1.06-1.86)

Habitat

Urban 617 (55.8%) 55.4 1c

Rural 488 (44.2%) 68.2 1.78 (1.42-2.05)

*Number of subjects and percentage (n), percentage of subjets with caries (DMFT-dft) (%). "Adjusted for the effects of the other independent variables included in the table. bHosmer-Lemeshow test. p-value > 0.05. 'Reference category.

hed their teeth before yesterday and among those who brushed their teeth more than two days ago (POR = 1.60; 95% CI, 1.20-2.28). These results are of statistical significance.

The final models also considered the variable of quantity of toothpaste on the toothbrush. Taking as a reference category those who use a third of the toothbrush with toothpaste, the subjects who use two thirds show more caries (POR = 1.32; 95% CI, 1.10-2.16), and also those who use the whole brush (POR = 1.52; 95% CI, 1.20-1.61). These results show a clear significant tendency. This association between quantity of toothpaste and caries could be due to 3 reasons: a) the variable quantity of toothpaste could be a proxy of the quality of brushing. Therefore, the subjects who use less toothpaste

are probably those who have more knowledge on the more adequate way to brush their teeth; b) the quantity of toothpaste used is related to the frequency of brushing (p = 0.02). Therefore, it is reasonable to think that the subjects who brush their teeth more frequently use less toothpaste, and c) the use of great quantities of toothpaste could generate a false sensation of cleanliness (production of great amounts of foam and a pleasant sensation), which reduces brushing time.

Consistently with previous studies, a high consumption of sweets is related to higher caries prevalence, these results are consistent with the bibliography12-15, even though the consumption between meals could be a more suitable measure. Therefore, taking the subjects who never consume as a reference category, we found

increasing values of caries prevalence as the situations in which they are consumed increase: in one situation (POR = 1.39); in two situations (POR = 1.44); in three situations (POR = 1.46), and in four situations (POR = 1.68). The fact that the models have been adjusted according to brushing, quantity of toothpaste, or visits to the dentist, allows us to confirm the negative effect that the number of times we eat sweets has on caries, independently of the subject maintaining adequate hygienic habits.

In as far as the variable of education of the mother is concerned, a univariate analysis shows how caries prevalence increases as the level of education decreases, we go from a prevalence of 49.7% in children with mothers with higher education, to a prevalence of 78.8% in children with mothers who have no education. These results are consistent with those in the bibliography616, which generally show worse dental health in the lower economic strata17-19.

However, when analysing prevalence odds ratio we observe different data: taking higher education as a reference category, caries prevalence odds ratio in secondary education is POR = 1.62, while in the categories of primary and no education this increase is lower POR = 1.40. These results show an important confusing effect of the remaining variables, an effect that has not been described in the bibliography.

The lower caries prevalence in subjects belonging to the lower cultural strata compared with the middle cultural strata is not easy to explain. Perhaps the subjects belonging to a lower socio-economic strata consume less refined sugar products20, although we cannot ignore that this difference in prevalence may be a sample effect, in fact, the confidence intervals are not significant.

We also considered one aggregate variable of the municipality where the pupils live, rural municipalities and urban municipalities, finding that there is a lower level of dental health in rural habitats. Traditionally, there has always been a lower economical and cultural level in rural habitats, and less possibility of access to a dentist. And despite including the parents' level of education and check ups in the models, we cannot ignore a certain degree of residual confounding. Moreover, the level of studies is a proxy of socio-cultural level, but both are not equal. In any case, these interpretations must be made with caution, due to the possibility of «ecological fallacy» in the measurement of these variables11.

Moreover, the models include variables of orthodontics and visits to the dentist. These variables have been included solely to adjust the models, and their coefficients have no direct interpretation on the models.

Since the current analysis is based on cross-sectional data, the validity of the conclusions could be li-

mited by the difficulty in differentiating between cause and effect. In this case however, the factors associated with caries (i.e. habitat, socio-economical level, oral hygiene, sweet consumption) are variables that are unlikely to change during the period of time in which the dependent variable is measured11.

This study may give health educators, planners and other health professionals' information that will help to reduce dental caries. The main goals of dental health programmes should be to achieve quality brushing every day in children, to reduce the consumption of sweets, and to increase knowledge on dental health.

Acknowledgements

Our thanks to Dirección Xeral de Saúde Pública da Xunta de Galicia.

References

1. Álvarez-Arenal A, Álvarez-Riesgo JA, Pena-López JM, Fernández-Vázquez JP. DMFT, dmft and treatment requirements of schollchildren in Asturias, Spain. Community Dent Oral Epidemiol. 1998;26:166-9.

2. Llodra JC, Bravo M, Cortés J. Encuesta de salud oral en España (2000). Rev Col Odont Esp. 2002;7:19-63.

3. Noguerol B, Llodra JC, Sicilia A, Follana M. La salud buco-dental en España 1994: antedecentes y prespectiva de futuro. Av Odontoestomatol. 1990;6:323-30.

4. Diehnelt DE, Kiyak HA. Socioeconomic factors that affect international caries levels. Community Dent Oral Epidemiol. 2001;29:226-33.

5. Mascarenhas AK. Determinants of caries prevalence and severity in higher SES Indian children. Community Dent Health. 1999;16:107-13.

6. Irigoyen ME, Sánchez-Hinojosa G. Changes in dental caries prevalence in 12-year-old students in the State of Mexico after 9 years of salt fluoridation. Caries Res. 2000; 34:303-7.

7. Harris R, Nicoll AD, Adair Pm, Pine CM. Risk factors for dental caries in young children: a systematic review of the literature. Community Dent Health. 2004;21:71-85.

8. Impact of caries and dental fluorosis on 12-year-old school-children's self-perception of appareance and chewing. Ca-dernos Saúde Pública. 2003;19:323-30.

9. WHO. Oral health surveys. Basic methods. 3rd ed. Geneva: WHO; 1987.

10. Hosmer DW, Taber S, Lemeshow S. The importance of assessing the fit of logistic regression models: a case study. Am J Public Health. 1991;81:1630-5.

11. Rothman KJ, Greenland S. Types of epidemiology study. In: Rothman KJ, Greenland S, editors. Modern epidemiology. Philadelphia: Lippincott and Raven; 1998. p. 75-6.

12. Moore WJ. The role of sugar in the aetiology of dental caries (I). Sugar and the antiquity of dental caries. J Dent. 1983;11:189-90.

13. Rugg-Gunn AJ, Murray JJ. The role of sugar in the aetiology of dental caries (II). The epidemiologi-cal evidence. J Dent. 1983;11:190-9.

14. Woodward M, Walker ARP. Sugar consumption and dental caries: evidence from 90 countries. Br Dent J. 1994;176:297-302.

15. Bedos C, Brodeur JM. Determinants of dental caries in Haitian schoolchildren and implications for public health. Sante. 2000;10:161-8.

16. Flinck A, Kallestal C, Holm AK, Allebeck P, Wall S. Distribution of caries in 12-year-old children in Sweden. Social and oral health-related behavioural patterns. Community Dent Health. 1999;16:160-5.

17. Dummer PM, Addy M, Hicks R, Kingdon A, Shaw WC. The effect of social class on the prevalence of caries, plaque, gin-

givitis and pocketing in 11-12-year-old children in South Wales. J Dent. 1987;15:185-90.

18. Hjern A, Grindefjord M, Sundberg H, Rosen M. Social inequality in oral health and use of dental care in Sweden. Community Dent Oral Epidemiol. 2001;29:167-74.

19. Pitts NB. Inequalities in children's caries experience: the nature and size of the UK problem. Community Dent Health. 1998;15:296-300.

20. Burt BA, Pai S. Sugar consumption and caries risk: a systematic review. J Dent Educ. 2001;65:1017-23.