Community Dental Health (2004) 21 (Supplement), 121–130
© BASCD 2004
International comparisons of health inequalities in childhood
dental caries
Cynthia M Pine1, Pauline M Adair2, Alison D Nicoll1, Girvan Burnside1,
Poul Erik Petersen3,4, David Beighton5, Angela Gillett1, Ruth Anderson6,
Shahid Anwar7, Susan Brailsford5, Zdenek Broukal8, Ivor G Chestnutt9,
Dominique Declerck10, Feng Xi Ping11, Roberto Ferro12, Ruth Freeman13,
Tshepo Gugushe14, Rebecca Harris1, Brent Lin15, Edward C M Lo16,
Gerardo Maupomé17, Mohamed Hanif Moola18, Sudeshni Naidoo19,
Francisco Ramos-Gomez20, Lakshman P Samaranayake16, Swarngit Shahid21,
Marit Slåttelid Skeie22, Christian Splieth23, Betty King Sutton24, Teo Choo Soo25
and Helen Whelton26
1
WHO Collaborating Centre on Oral Health in Deprived Communities, University of Liverpool Dental School, England; 2Department
of Clinical Psychology, The Royal Hospitals, Belfast, Northern Ireland; 3Faculty of Health Sciences, University of Copenhagen,
Denmark; 4World Health Organization, Oral Health Programme, Geneva, Switzerland; 5Guy’s, King’s and St Thomas’ Dental
Institute, London, England; 6Howard University College of Dentistry, Washington DC, USA; 7Oral Health and Ethnicity Unit, Leeds
Dental Institute, England; 8Institute of Dental Research, Prague, Czech Republic; 9Dental Public Health Unit, Dental School, Cardiff,
Wales; 10School of Dentistry, Oral Pathology and Maxillofacial Surgery, Catholic University Leuven, Belgium; 11School of Stomatology, Shanghai Second Medical University, China; 12Dentistry Unit, Cittadella Hospital, Padova, Italy; 13School of Dentistry,
Queen’s University of Belfast, Northern Ireland; 14Faculty of Dentistry, Medical University of Southern Africa, South Africa;
15
University of Texas Health Science Center, San Antonio, USA; 16Faculty of Dentistry, University of Hong Kong; 17Center for Health
Research, Portland, Oregon, USA; 18University of the Western Cape, Cape Town, South Africa; 19Department of Community
Dentistry, University of Stellenbosch, Tygerberg, South Africa; 20Division of Pediatric Dentistry, University of California San
Francisco, USA; 21Personal Dental Service, Bradford City Primary Care Trust, England; 22Department of Odontology, University
of Bergen, Norway; 23University of Greifswald, Germany; 24Department of Health and Human Services, Raleigh, North Carolina,
USA; 25Faculty of Dentistry, National University of Singapore; 26Oral Health Services Research Centre, University Dental School
Cork, Ireland.
Objective To undertake formative studies investigating how the experience of dental caries in young children living in diverse settings
relates to familial and cultural perceptions and beliefs, oral health-related behaviour and oral microflora. Participants The scientific
consortium came from 27 sites in 17 countries, each site followed a common protocol. Each aimed to recruit 100 families with
children aged 3 or 4 years, half from deprived backgrounds, and within deprived and non-deprived groups, half to be “caries-free”
and half to have at least 3 decayed teeth. Outcome measures Parents completed a questionnaire, developed using psychological
models, on their beliefs, attitudes and behaviours related to their child’s oral health. 10% of children had plaque sampled.
Results 2,822 children and families were recruited. In multivariate analyses, reported toothbrushing behaviours that doubled the odds
of being caries-free were a combination of brushing before age 1, brushing twice a day and adult involvement in brushing. Analyses
combining beliefs, attitudes and behaviours found that parents’ perceived ability to implement regular toothbrushing into their child’s
daily routine was the most important predictor of whether children had caries and this factor persisted in children from disadvantaged
communities. 90% of children with lactobacillus had caries. Conclusions Parental beliefs and attitudes play a key role in moderating
oral health related behaviour in young children and in determining whether they develop caries. Further research is indicated to
determine whether supporting the development of parenting skills would reduce dental caries in children from disadvantaged
communities independent of ethnic origin.
Key words: children, dental caries, international, multi-centre, epidemiology
Introduction
Despite reductions in the prevalence of dental caries in
many countries (Marthaler et al., 1996), children from
disadvantaged communities continue to experience higher
disease levels (Pitts et al., 2003; Vargas et al., 1998).
Further, children from minority ethnic groups are also
over represented (Pine et al, 2003; Vargas et al., 1998).
However, in some countries it appears that the association between higher caries experience and ethnic diversity may be a minor confounder and that the underlying
relationship relates to poverty, educational attainment
Correspondence to: Professor Cynthia M. Pine, The University of Liverpool, School of Dentistry, Pembroke Place, Liverpool L3
5PS, UK. E-mail: [email protected]
and acculturisation. (Bedi and Uppal, 1995; Sundby and
Petersen, 2003; Vargas et al.,1998; Verrips et al, 1992).
This project brought together an international research
consortium with the overall aim of determining optimum
interventions to reduce dental caries in children in disadvantaged communities and to minimise the effects of
exclusion from health care systems, of ethnic diversity,
and health inequalities. In order to provide a common
platform for the development of evidence-based interventions, a series of initial cross-national studies to a
common protocol have been undertaken (Pine et al., 2004).
This paper describes the overall results from the crossnational child and family studies which were undertaken
following a systematic review of the literature of key
aetiological factors (Harris et al., 2004). Explanatory
models were developed a priori which provided a framework within which appropriate psychological models could
be operationalised to develop a standardized measure
examining attitudes, beliefs and the nature of the oral
health related behaviours (Pine et al., 2004).
Univariate descriptive statistics were calculated. Multivariate analyses were undertaken and the exhaustive
CHAID algorithm within SPSS Answer Tree version 3
was used to model the data. This method fully partitions
the data into groups using a set of possible predictor
variables, to give the best prediction of a target variable,
in this case, dental health, i.e. caries free children. Starting with the whole data set, the data are partitioned
automatically using the predictor that shows the most
significant relationship with the target variable. The
significance of the relationships between predictors and
target variables is assessed using the chi-squared statistic. The categories of the predictor variables are merged
automatically to give the best split for each variable. This
process is then repeated within each of the subgroups,
until there are no more significant results, or the groups
become too small to split further. The groups that cannot
be split further, are called terminal nodes, and each case
is included in one and only one of these nodes. For each
terminal node, the chances of the children in this group
being caries free were calculated using odds ratios.
Methods
Ethical approval, study populations and examinations
Investigators were required to complete the on-line ethics
training and evaluation that is provided by the National
Institutes of Health. At each site, ethics committee
approval was obtained for the conduct of the crossnational studies. Study populations were to be selected
from urban areas and were not designed to be nationally
representative. The research consortium provided study
populations from a wide range of ethnic, racial and
cultural groups.
Each study site was asked to recruit 100 study families with children aged 3½ to 4½ years (sampling details
are provided in Pine et al., 2004). Fifty families were
required to be from a relatively advantaged area and 50
from a relatively deprived area. Two extreme groups of
young children from the point of view of caries experience were selected, namely children with no visually
evident carious lesions (dmft=0) and those with at least
3 untreated carious lesions (dt > 3). These extreme groups
were selected, as variation in the un-calibrated examiner
diagnosis was least likely to affect case definition. Each
of the two groups of 50 children was to comprise 25 who
were caries free (dmft = 0) and 25 who had caries (dt >
3: three or more carious teeth – not filled). In addition to
the dental examination, dental plaque was collected from
a sub-sample of children and the procedure and results
are described in Beighton et al. (2004) and Pine et al.
(2004). Data collection began in September 2001.
Parental questionnaires
The purpose of the questionnaire to parents was to collect
information in relation to child and family oral-health
related behaviours, attitudes and beliefs. Details of the
development, structure and content of the parent’s questionnaire are given in Pine et al. (2004).
Statistical methods
All data were analysed centrally by the study statistician.
Data cleaning and resolution were undertaken with reference back to key researchers at each site as needed.
Results
Demographics and reported behaviours
The demographic profile of the children included in the
study at each study site is presented in Table 1. In total,
the study involved 2,822 children and their families in 27
sites in 17 countries. The overall mean age of the children
was 4.0 years (sd 0.6) and 50% were boys. As described,
locally appropriate methods were used to classify families
as being disadvantaged. In order to standardise this
classification for the international comparisons, the level
at which mothers completed their formal education was
used as a proxy for socio-economic status (Pine et al.,
2004). This single measure was found to correctly identify 78% of families described as low socio-economic
status by the study site and 60% of families described as
middle or high socio-economic status. An example of the
difference in the two classification methods is that of
Norway where most women entered some type of further
education, as only 6% of mothers completed their education at school. Of the 74 identified by the Norwegian
research team as being low socio-economic families, 36
mothers (49%) attended colleges principally offering
vocational courses. When using mother’s education as
the classification method for the international comparisons, these families would be re-classified as middle/high
socio-economic status.
Overall, 50% of children were judged to be caries free.
In some sites, it proved difficult to find children with at
least 3 decayed teeth in the higher socio-economic group.
The aim was to identify two distinct groups of children,
those who were apparently caries free and those with
at least 3 decayed teeth. This relatively high level of
caries experience was chosen to ensure those most at risk
were found and to minimise the potential for case
misclassification. In fact, the mean dmft of the children
with caries was found to be 6.1, and ranged from 3.8 to
7.2 between sites (Table 1).
Tables 2 and 3 present some key aspects of the children’s oral-health related behaviours in relation to
toothbrushing and sugar consumption as reported by
122
Table 1. Demographic profile of children by site, country and ethnic group
Site
n
Mean age
(s.d.)
% male
% mothers % of children
with school
caries free
education only
Mean dmft of
children with
caries
Europe
Scotland White
Northern Ireland White
Wales White
Ireland White
Belgium White
Denmark White
Italy White
Norway White
Germany White
Czech Republic White
England Pakistani
England Chinese
108
100
91
99
97
77
145
160
100
98
113
42
4.1
4.1
4.2
3.9
4.3
4.1
4.2
4.0
4.0
4.0
4.1
4.0
(0.5)
(0.3)
(0.4)
(0.5)
(0.6)
(0.4)
(0.6)
(0.8)
(0.5)
(0.7)
(0.3)
(0.7)
46
43
55
36
56
57
56
59
54
59
50
52
54
55
54
60
44
37
49
6
69
59
74
52
62
51
55
52
51
60
72
63
50
49
47
52
5.8
6.2
5.3
4.6
5.2
7.1
3.9
5.6
5.2
4.0
6.6
6.8
(3.4)
(2.8)
(2.3)
(1.5)
(2.1)
(4.0)
(2.6)
(3.0)
(2.6)
(3.5)
(3.9)
(4.0)
North America
USA White
USA Mexican-American
USA Chinese
USA African-American
Mexico Mexican
61
82
151
53
110
3.7
3.7
4.2
3.8
4.1
(0.6)
(0.8)
(0.9)
(0.6)
(0.6)
54
55
44
51
46
11
57
81
36
63
49
35
21
81
51
5.0
7.0
6.4
3.8
7.1
(3.3)
(4.3)
(4.0)
(2.3)
(2.4)
Africa
South Africa: Capetown Cape Coloured 214
South Africa Black African
80
Madagascar African
105
Tanzania African
100
3.9
4.3
3.9
4.0
(0.5)
(0.8)
(0.8)
(0.7)
54
47
42
39
62
76
66
76
43
50
42
41
6.9
6.2
6.1
5.6
(3.5)
(2.8)
(2.7)
(3.9)
Asia
China: Shanghai Chinese
China: Hong Kong Chinese
China: Guangzhou Chinese
Singapore Chinese
Thailand Thai
Pakistan Pakistani
4.0
3.8
4.1
3.9
4.0
3.8
(0.4)
(0.3)
(0.4)
(0.6)
(0.5)
(0.6)
49
59
54
42
44
54
82
88
56
55
59
49
52
55
50
61
22
50
6.4
5.9
6.5
3.8
7.2
6.8
(3.8)
(2.1)
(3.3)
(3.3)
(4.9)
(3.5)
4.0 (0.6)
50
58
50
6.1 (3.5)
Total (from 27 sites in 17 countries)
100
93
100
117
125
101
2822
their parents. These data are presented for the whole
study sample and give an overall picture, although there
were differences in behaviour according to the caries
status of the child and mothers’ education level. Brushing twice a day was reported by at least 50% of parents
in all but 4 sites (Table 2). One of the latter was Belgium
where the advice given to parents is to brush once a day.
Only around a third of the families in China reported that
their children’s teeth were brushed twice daily, however
this figure was doubled for Chinese families in the USA
and UK. Brushing from a young age was a frequently
reported habit in Northern Europe and all USA ethnic
groups except Chinese families. The majority of parents
are still involved in brushing their 4 year olds’ teeth, the
exceptions being families in Tanzania, Madagascar, China,
and English children of Pakistani parents.
Most parents reported that their children had sweets
or chocolate everyday or most days in Germany, the UK,
Belgium and South Africa but this was rare for Norwegian families (7%) and reported by less than one in 3
families from Thailand, China and Tanzania (Table 3).
Adding sugar to drinks is a cultural norm for Pakistani
(91%), Mexican (88%), Czech (79%) and Italian (73%)
families but much less common for Chinese families (14%).
Czech and Mexican families were also the most likely to
report that their child drank sugary drinks in bed. In order
to examine the relative importance of the differences in
123
reported behaviour, multivariate analyses were undertaken.
Multivariate analyses
Figure 1 presents the results of the CHAID analysis into
which all brushing parameters were included and show
which aspects of reported child toothbrushing behaviour
are significant in predicting which children are caries free.
The single most important brushing-related behaviour
predicting whether children are caries free was found to
be the age parents report commencing brushing their
children’s teeth. Those who began toothbrushing after
2 years of age were significantly more likely to have
caries (p < 0.001). Additional analyses were conducted
(not shown) which found that the relationship between
age of commencement of toothbrushing and caries holds
for sites with and without water fluoridation and for
deprived and non-deprived groups. The most significant
combination of beneficial toothbrushing behaviours was
found for children in Group 1. They were as follows:
parents began brushing during their child’s first year,
brushed twice a day and an adult was involved with the
child’s toothbrushing. This combination of behaviours
reduced the odds of children having caries by 50%.
Figure 2 presents the multivariate analysis undertaken
to identify which were the significant aspects of reported
child sugar consumption that predicted which children
were likely to be caries free. The single most important
Table 2. Child oral-health related toothbrushing behaviour as reported by families by country, site and racial/ethnic group
ordered by frequency of each behaviour
Percentage brush at least twice a day
Percentage start brushing before age 2
Percentage adult involved in brushing
Scotland White
Wales White
England Chinese
Thailand Thai
Denmark White
Northern Ireland White
Tanzania African
Norway White
South Africa Black African
85
83
83
83
81
77
74
71
68
99
99
96
94
94
90
86
83
82
Denmark White
Norway White
USA White
USA Mexican American
Czech Republic White
Scotland White
Thailand Thai
Germany White
Wales White
97
97
90
86
86
84
83
80
79
Germany White
Czech Republic White
66
65
Norway White
Scotland White
Denmark White
Wales White
Germany White
Northern Ireland White
USA Mexican American
USA African American
South Africa: Capetown
Cape Coloured
USA White
Ireland White
82
76
79
77
USA African American
England Pakistani
USA Chinese
Pakistan Pakistani
USA Mexican American
China: Hong Kong Chinese
South Africa: Capetown
Cape Coloured
Madagascar African
Ireland White
Singapore Chinese
Mexico - Mexican
USA White
Italy White
China: Guangzhou Chinese
China: Shanghai Chinese
Belgium White
64
64
61
60
59
59
59
Thailand Thai
England Chinese
Belgium White
Italy White
England Pakistani
Czech Republic White
Singapore Chinese
75
71
65
63
62
53
50
Italy White
South Africa: Capetown
Cape Coloured
USA African American
Ireland White
Belgium White
China: Hong Kong Chinese
Northern Ireland White
England Chinese
Mexico – Mexican
53
52
52
51
50
43
36
32
32
Tanzania African
China: Hong Kong Chinese
USA Chinese
Mexico - Mexican
Pakistan Pakistani
China: Guangzhou Chinese
South Africa Black African
China: Shanghai Chinese
Madagascar African
50
46
43
40
30
29
27
26
24
South Africa Black African
Pakistan Pakistani
Singapore Chinese
USA Chinese
Tanzania African
China: Guangzhou Chinese
England Pakistani
Madagascar African
China: Shanghai Chinese
63
60
60
59
48
48
42
29
25
75
74
74
73
69
69
66
Table 3. Child oral-health related sugar behaviour as reported by families for children by country, site and racial/ethnic group
ordered by frequency of each behaviour
Percentage eat sweets or chocolate
every day or most days
Percentage add sugar to drinks
Percentage drink sugary drinks
in bed
Germany White
Scotland White
England Pakistani
88
81
80
Madagascar African
Pakistan Pakistani
Mexico - Mexican
94
91
88
61
37
27
82
79
73
59
Czech Republic White
Mexico - Mexican
South Africa: Capetown
Cape Coloured
Wales White
Ireland White
Thailand Thai
Germany White
Italy White
Ireland White
Northern Ireland White
Wales White
76
76
75
75
South Africa: Capetown
Cape Coloured
Belgium White
England Chinese
Pakistan Pakistani
China: Shanghai Chinese
Mexico - Mexican
China: Hong Kong Chinese
Singapore Chinese
USA White
Madagascar African
Czech Republic White
USA Chinese
USA African American
Denmark White
USA Mexican American
South Africa Black African
Thailand Thai
China: Guangzhou Chinese
Tanzania African
Norway White
71
South Africa Black African
Czech Republic White
Italy White
South Africa: Capetown
Cape Coloured
Tanzania African
53
Scotland White
20
68
67
65
62
61
55
47
46
45
40
36
35
35
34
33
30
29
18
7
USA Mexican American
Germany White
USA African American
China: Hong Kong Chinese
Ireland White
England Chinese
Denmark White
Wales White
USA White
Norway White
Scotland White
China: Guangzhou Chinese
China: Shanghai Chinese
England Pakistani
USA Chinese
Singapore Chinese
Thailand Thai
Northern Ireland White
Belgium White
46
36
35
26
25
23
22
21
20
17
14
14
12
12
11
11
9
5
4
Belgium White
China: Hong Kong Chinese
Madagascar African
Singapore Chinese
China: Shanghai Chinese
Italy White
Northern Ireland White
China: Guangzhou Chinese
USA Mexican American
Pakistan Pakistani
USA Chinese
Denmark White
USA White
Norway White
USA African American
England Chinese
South Africa Black African
Tanzania African
England Pakistani
19
18
18
17
17
17
15
14
13
13
13
12
11
10
8
7
5
4
3
26
22
22
21
124
Figure 1. Multivariate analysis to identify which aspects of reported child toothbrushing behaviour are significant in predicting
which children are likely to be caries free
Figure 2. Multivariate analysis to identify which aspects of reported child sugar consumption behaviour are significant in
predicting which children are likely to be caries free
sugar-related behaviour predicting whether children are
caries free or not was found to be night-time drinking
behaviour. Children whose parents reported that their
child has the habit of drinking sugar-containing drinks in
bed were significantly less likely to be caries free (p <
0.001). The most favourable combination of behaviours
that led to the best reduction in the odds of children
having caries were those in which a wide variety of sugar
consumption was restricted. This combination was not
drinking sugary drinks in bed, restricting the frequency
of sugary drinks, not adding sugar to drinks and restricting the frequency of sweets and chocolates (Group 2 in
Figure 2).
Figures 3 to 5 presents the multivariate analysis
describing which children are most and least likely to be
caries free according to the full range of parameters
measured, i.e. sugar and toothbrushing behaviour and
parents’ attitudes to these behaviours (the measurement
of the latter is described in Adair et al., 2004). The
125
outcome of the modeling undertaken within the CHAID
analysis comprise unique terminal nodes, i.e. each group
of children is separated and represent the end of the
analysis. These groups are presented as the columns on
the far right. In the multivariate analyses presented in
Figures 3 to 5, all the groups of children who had been
to the dentist had higher levels of caries experience.
Unexpectedly, the most significant variable for the
whole study population predicting whether children would
be caries free was not the children’s oral health-related
behaviour, but a parents’ attitude to their perceived ability to deliver the behaviour (in this case, regular and
effective toothbrushing; Figure 3). This factor was termed
the Brushing Parental Efficacy Factor and was a combination of beliefs and attitudes (Adair et al., 2004). The
Brushing Parental Efficacy Factor (split in the analysis at
mean scores < 4; 4 to 4.3; >4.3) comprises the average
response (from categories 1= strongly agree to 5= strongly
disagree) to the following statements:
Figure 3. Multivariate analysis to identify which aspects of reported child toothbrushing behaviour, sugar consumption
behaviour and parental attitudes are significant in predicting which children are likely to be caries free (for the whole sample).
•
If our child does not want to brush his/her teeth
every day we don’t feel we should make them
• I don’t know how to brush my child’s teeth properly
• It would not make any difference to our child getting
tooth decay, if we helped him/her brush every day
• We don’t have time to help brush our child’s teeth
twice a day
• We cannot make our child brush his/her teeth twice
a day
• It is not worth it to battle with our child to brush his/
her teeth twice a day
Low mean factors are in families where parents
believe they cannot effectively ensure toothbrushing
behaviour in their child.
Significant factors found in the second level of the
CHAID analysis were different according to the Brushing
Parental Efficacy Factor (BPE factor). For parents with
low mean factor scores (<4), the second factor was whether
their child had visited a dentist. For parents with moderate BPE factor scores, age of commencement of child’s
toothbrushing was next in level of significance. For
parents who were confident about implementing regular
toothbrushing, i.e. with high BPE factor scores (over 4.3),
the most important predictor was whether their children
drank sugary drinks regularly. For several groups, the
fourth or fifth level, where present, also mainly comprised
measures of parental attitudes, i.e. how serious they
regard caries as a disease, Perceived Seriousness Factor
(items given in Adair et al., 2004) and how important
parents perceive twice a day toothbrushing to be,
126
Figure 4. Multivariate analysis to identify which aspects of reported child toothbrushing behaviour, sugar consumption
behaviour and parental attitudes are significant in predicting which children are likely to be caries free (for those children
whose mothers had further education)
Figure 5. Multivariate analysis to identify which aspects of reported child toothbrushing behaviour, sugar consumption
behaviour and parental attitudes are significant in predicting which children are likely to be caries free (for those children
whose mothers had no further education).
127
Brushing Importance/Intention Factor (items given in
Adair et al., 2004).
In Figure 3, children in Groups A and B had the
lowest number that were caries free at 10% and 24% and
those in Groups C and D have the highest number at
65%. Starting from the left of Figure 3, it can be seen that
parents of children in Group A have the lowest mean
Brushing Parental Efficacy factor scores (<4). The next
most significant variable is that they have taken their
child to a dentist. This is linked to higher caries experience as 39% of parents in Group A reported that their
child had toothache in the previous year, compared to
only 14% of the whole dataset. The families in Group A
are less likely to consider dental caries as a serious childhood disease (Perceived seriousness factor <4) and they
perceive themselves to be least able to establish and
maintain regular toothbrushing behaviour for their child
(BPE factor <3.3). This combination of attitudes and
behaviours resulted in the children of these families
having an over 8-fold increase in the chances of having
dental caries (odds ratio=8.4). For Group B with only 24%
caries free, the first two variables are the same as Group
A, but these families do rate caries as a serious childhood
disease. However, two behaviours mitigate against their
children being caries free in that their teeth are not brushed
twice a day and they eat sweets and chocolates regularly.
This combination of attitudes and behaviours increase
the chances of caries developing over 3-fold.
Families in Group C and D represent the most favourable outcome with the chances of their children being
caries free as 50% above the average. In these families,
there are high levels of parents’ perceived ability to
deliver regular toothbrushing for their children (BPE
factor scores >4.3). In Group C, the children are reported
to have restricted consumption of sugary drinks and
these two factors alone are sufficient to halve the odds
of having children with caries. In Group D, although
parents report that their child does drink sugary drinks
regularly, the odds of having caries are reduced as
parents began brushing their children’s teeth during their
first year.
A further analysis (not shown) was undertaken for
children whose dental plaque was cultured (n=249). A
multivariate analysis to explain the presence of caries was
undertaken including all possible parameters, i.e. child’s
plaque composition, reported toothbrushing and sugar
behaviour, parental attitudes to brushing, sugar and
caries and the child’s reported dental visiting. The most
significant of all these factors was found to be whether
the child had lactobacilli present, 60 of whom did. Of
these 54, or 90% had caries. There was only one further
significant factor (as the numbers were small), that was
whether they ate cakes and biscuits frequently. Of those
without lactobacilli and who did not eat cakes and
biscuits frequently, 69% were caries free, compared to
42% of those without lactobacillus who did eat cakes and
biscuits frequently.
Separate multivariate analyses were undertaken for
families in which mothers had further education (Figure
4) and, for those in which mothers had no further
education (Figure 5). In advantaged families, the most
significant factor predicting whether children were caries
free was whether children were reported as brushing twice
a day (Figure 4). It is noted (not shown in the figure) that
the Brushing Parental Efficacy factor was very close
behind twice daily brushing as the first level significant
predictor. Of the 5 groups of children, who did not brush
twice daily, 4 groups had increased chances of having
caries. In the group of children with reduced odds of
having caries (OR=0.5), parents reported high levels of
control of the children’s sugar consumption (Sugar
Parental Efficacy factor>3.5; for item contents in this factor,
see Adair et al. (2004)). Children in Group E had the least
favourable combination of factors leading to an almost
4-fold increase in the chances of having caries. Children
in Group F had the most favourable combination, with
90% being caries free. These children were significantly
more likely to brush twice a day, their parents were regular dental attenders, they had not taken their child to the
dentist and, they had a high level of belief in the seriousness of caries as a childhood disease.
For disadvantaged families (Figure 5), the first variable to enter the model was whether parents had taken
the child to the dentist, for those that had done so, the
next significant caries predictor was the level of Brushing
Parental Efficacy factor as had been found for the overall
analysis. The least favourable combination of variables
were found in Group G and resulted in an over 6-fold
increase in the chances of children having caries. These
variables were as follows. Children were more likely to
have been taken to the dentist; parents had low BPE;
parents’ own dental attendance was symptomatic; and, it
was more likely to be the parents’ first child. In contrast,
children in Group H were 50% less likely to have caries.
Parents had high levels of belief in their ability to ensure
children’s toothbrushing, they began brushing their children’s teeth before the age of 2 years and their children
were less likely to eat cakes and biscuits frequently.
Therefore, favourable beliefs and behaviours mitigated
the increased chance of caries that is associated with low
socio-economic status.
Discussion
The general direction of the results from these data is in
line with associations previously reported in the literature
(Harris et al., 2004). This provides a measure of the
robustness of the results and is important as these crossnational studies had some limitations in their execution.
The classification of caries and caries free status was
dependent on un-calibrated examinations. To reduce
misclassifications, children designated as having caries
were required to have at least 3 decayed teeth. Further,
all analyses were conducted on the basis of children
as having caries or being caries free rather than on
the absolute level of disease recorded. The fact that
significant associations were found between caries,
unfavourable sugar consumption and brushing behaviour provides further support for the accuracy of the
classification of caries and caries free derived from the
dental examinations. A second potential limitation was
the accuracy of parents in completing the questionnaire
data. The data are self reported beliefs and behaviours,
so there remains the possibility that parents can report
what they perceive as the correct answer rather than
what they actually believe or do. However, the presence
128
or absence of the disease can assist in validation. For
example, children in Group B were three times more likely
than their peers to have caries and their parents were
significantly more likely to report that their children
brushed less than twice daily and consumed sweets or
chocolates regularly.
The results of these studies provide important new
data in showing how parents’ attitudes and beliefs about
oral-health related behaviour are associated with the likelihood of their children developing caries. This study
also found clear variations in parental beliefs and behaviours for different ethnic and socio-economic groups that
separately impact on their children’s oral health related
behaviour.
This formative international research provides the basis
for the development of a number of hypotheses that
should be tested in longitudinal studies and further
research. The first area of research recommended relates
to parenting and child oral health related behaviours.
Parental efficacy can be measured as a general concept
relating to parents overall beliefs about how effective
they are as parents (Coleman and Karraker, 1997). These
measures often include key known difficult skills such
as ability to discipline, to provide adequate nutrition
or comfort one’s child (Woodruff and Cashman, 1993).
Whilst general parental efficacy tools (validated questionnaires) exist (Bandura, 1982), no measure relating
to oral health has been developed. Research is required
to develop such a new standardised measure that could
be used for identifying families at increased risk and
as an outcome measure in longitudinal intervention
studies.
A central hypothesis arising from this study is that a
behavioural intervention for parents designed to improve
reported parental efficacy and develop key skills should
lead to child behaviours that will prevent caries from an
early age. This hypothesis should be evaluated within a
randomised controlled trial. Taking a holistic view, further
research to examine the parenting skills required to establish the routine of regular toothbrushing behaviour for
young children could provide a valuable template for
understanding how parenting skills can be developed for
other (non-dental) child behaviours.
In summary, the results of these initial studies undertaken by this international consortium of scientists from
diverse communities around the world have provided
fresh insight into childhood dental caries and given a
new research framework to develop holistic interventions
for evaluation in prospective longitudinal studies.
Acknowledgements
With thanks to the children and families who participated
and to the colleagues named below who gave their time
to support this research. I Asselman, Vera Beecher, S
Aida Borges-Yáñez, B Breistein, John P Brown, B Bünger,
G Bunting, Anne Byres, Usuf Chikte, Eu Oy Chu, M
Croon, Julia Csikar, Morag Curnow, Catherine S Dang, J
De Ceulaer, E De Maeyer, Nick DelaCruz, Deirdre Devine,
J B du Plessis, Andrew Forgie, Jill Fortuin, Soraya
Harnekar, Garnett Henley, C Hölzel, Stephen Hsu, Maria
Irigoyen-Camacho, Bonnie Jue, J Kroon, George Lee, Marie
Lee, Erika Lencova, R Lesolang, H C Lin, Teresa Loh-Lee,
Y Luo, Herman Ma, Helen MacDiarmid, Betty Mok, David
Motloba, Nursery school staff and pupils and parents in
Dundee, Margaret Paterson, A Sweeney, Josephine Tang,
K Vermeir, Adeline Wong, Margaret Wu, He Yan, David
W Young, and Phillip Young.
This study was funded by the National Institutes of
Health, USA, NIH grant number DE13703-02. Additional
funding was obtained from the Ministry of Health, Czech
Republic, Grant No. 002377901, the Welsh Office for
Research and Development, and the Norwegian Foundation for Health and Rehabilitation.
References
Adair, P.M., Pine, C.M., Burnside, G., Nicoll, A.D., Gillett,
A., Anwar, S., Broukal, Z., Chestnutt, I.G., Declerck, D.,
Feng, X.P., Ferro, R., Freeman, R., Grant-Mills, D.,
Gugushe, T., Hunsrisakhun, J., Irigoyen-Camacho, M., Lo,
E.C.M., Moola, M.H., Naidoo, S., Nyandindi, U., Poulsen,
V.J., Ramos-Gomez, F., Razanamihaja, N., Shahid, S., Skeie,
M.S., Skur, O.P., Splieth, C., Teo, C.S., Whelton, H. and
Young, D.W. (2004): Familial and cultural perceptions and
beliefs of oral hygiene and dietary practices among ethnically and socio-economically diverse groups. Community
Dental Health 21 (Suppl.), 102–111.
Bandura, A. (1982): Self-efficacy in human agency. American
Psychologist 37, 122–147.
Bedi, R. and Uppal, R.D. (1995): The oral health of minority
ethnic communities in the United Kingdom. British Dental
Journal, 179, 421–425.
Beighton, D., Brailsford, S., Samaranayake, L.P., Brown, J.P.,
Feng, X.P., Grant-Mills, D., Harris, R.V., Lo, E.C.M.,
Naidoo, S., Ramos-Gomez, F., Teo, C.S., Burnside, G. and
Pine, C.M. (2004): A multi-country comparison of
caries-associated microflora in demographically diverse
children. Community Dental Health 21 (Suppl.), 96–
101.
Coleman, P.K. and Karraker, K.H. (1997): Self-efficacy and
parenting quality: findings and future applications. Devel-
129
opmental Review 18, 47–85.
Harris, R.V., Nicoll, A.D., Adair, P.M. and Pine, C.M. (2004):
Risk factors for dental caries in young children: a systematic review of the literature. Community Dental Health 21
(Suppl.), 71–85.
Marthaler, T.M., O’Mullane, D.M. and Vrbic, V. (1996): The
prevalence of dental caries in Europe 1990–1995. ORCA
Saturday afternoon symposium 1995. Caries Research 30,
237–255.
Pine, C., Burnside, G. and Craven, R. (2003): Inequalities in
dental health in the north-west of England. Community
Dental Health 20, 55–56.
Pine, C.M., Adair, P.M., Petersen, P.E., Douglass, C.W.,
Burnside, G., Nicoll, A.D., Anderson, R., Beighton, D.,
Bian, J.Y., Broukal, Z., Brown, J., Chestnutt, I.G., Declerck,
D., Devine, D., Espelid, I., Falcolini, G., Feng, X.P., Freeman, R., Gibbons, D., Gugushe, T., Harris, R.V., Kirkham,
J., Lo, E.C.M., Marsh, P., Maupomé, G., Naidoo, S.,
Ramos-Gomez, F., Sutton, B.K. and Williams, S. (2004):
Developing explanatory models of health inequalities in
childhood dental caries. Community Dental Health 21
(Suppl.), 86–95.
Pitts, N.B., Boyles, J., Nugent, Z.J., Thomas, N. and Pine,
C.M. (2003): The dental caries experience of 5-year-old
children in England and Wales. Surveys co-ordinated by the
British Association for the Study of Community Dentistry
in 2001/2002. Community Dental Health 20, 45–54.
Sundby, A. and Petersen, P.E. (2003): Oral Health status in
relation to ethnicity in the Municipality of Copenhagen,
Denmark. International Journal of Pediatric Dentistry 13,
150–157.
Vargas, C.M., Crall, J.J. and Schneider, D.A. (1998): Sociodemographic distribution of pediatric dental caries: NHANES
III, 1988–1994. Journal of the American Dental Association
129, 1229–1238.
Verrips, G.H., Frencken, J.E., Kalsbeek, H., ter Horst, G. and
Filedt Kok-Weimar, T.L. (1992): Risk indicators and
potential risk factors for caries in 5-year-olds of different
ethnic groups in Amsterdam. Community Dentistry and
Oral Epidemiology 20, 256–260.
Woodruff, S. and Cashman, J.F. (1993): Task, domain and
general self-efficacy: a reexamination of the self-efficacy
scale. Psychological Reports 72, 423–432.
130