clinical indicators. While several measures have been developed and
validated for use among adults (Cushing et al., 1986; Atchison and Dolan,
1990; Locker and Miller, 1994; Slade and Spencer, 1994; Leão and
Sheiham, 1996), work in the field of child OHRQoL has lagged behind.
However, Jokovic et al. (2002) recently reported on the Child Perceptions
Questionnaire (CPQ11-14), designed to measure OHRQoL among children
aged between 11 and 14 years. The CPQ11-14includes the four domain
subscales of oral symptoms (6 items; e.g., pain), functional limitations (9
items; e.g., difficulty eating), emotional well-being (8 items; e.g., avoiding
smiling or laughing around other children), and social well-being (12 items;
e.g., being asked questions by other children about his/her mouth).
To date, the validity and reliability of the CPQ11-14have been examined
in a clinical convenience sample of 123 children recruited from among
pediatric dentistry, orthodontic, and craniofacial patients in Toronto. These
groups were chosen since they had distinct clinical characteristics that were
expected to have differential effects on the children's quality of life, thus
maximizing variation for validity testing. The discriminative properties (i.e.,
cross-sectional validity and test-retest reliability) of the CPQ11-14were found
to be acceptable.
However, there have been no reports from other populations or settings,
and questions remain about the performance of the CPQ11-14in child
populations which exhibit the full distribution of clinical presentations. It is
important that the discriminative properties of such measures be acceptable
in these populations. Their ability to distinguish between individuals (or
groups) with poor OHRQoL and those with better OHRQoL is a key
characteristic which would enable such instruments to contribute to
improvements in oral health, through identifying those clinical or public
health interventions which produce the greatest improvement in OHRQoL.
The aim of this study was to examine the construct validity of the
CPQ11-14in a probability-based population sample of 12- and 13-year-old
New Zealanders. It was hypothesized that children with more severe
malocclusions would have higher overall (and domain) CPQ11-14scores, and
that this would also apply to those with greater dental caries experience.
he last decade has seen increasing recognition of the value of measures
of oral-health-related quality of life (OHRQoL) in supplementing
MATERIALS & METHODS
Ethical approval was obtained from the Taranaki Ethics Committee. In 2003,
there were approximately 1600 12- to 13-year-olds enrolled in the Taranaki
District Health Board's (TDHB) school dental service, of whom 961 attended
four large intermediate schools. A simple random sample of children in their 8th
year of schooling, and who were enrolled with the TDHB school dental service
(SDS), was selected from the four intermediate schools and invited to
participate. Assuming the prevalence of malocclusion to be 30% in that age
While the use of adult oral-health-related quality-
of-life (OHRQoL) measures in supplementing
clinical indicators has increased, that for children
has lagged behind, because of the difficulties of
developing and validating such measures for
children. This study examined the construct
validity of the Child Perceptions Questionnaire
(CPQ11-14) in a random sample of 12- and 13-year-
old New Zealanders. It was hypothesized that
children with more severe malocclusions or
greater caries experience would have higher
overall (and subscale domain) CPQ11-14scores.
Children (N = 430) completed the CPQ11-14and
were examined for malocclusion (Dental Aesthetic
Index) and dental caries. There was a distinct
gradient in mean CPQ11-14scores by malocclusion
severity, but there were differences across the four
subscales. Children in the worst 25% of the DMFS
distribution had higher CPQ11-14scores overall
and for each of the 4 subscales. The construct
validity of the CPQ11-14appears to be acceptable.
KEY WORDS: child, oral health, quality of life,
validity, malocclusion, dental caries.
Received September 13, 2004; Last revision March 15,
2005; Accepted April 22, 2005
Validation of the Child Perceptions
L.A. Foster Page1, W.M. Thomson1*,
A. Jokovic2, and D. Locker2
1Department of Oral Sciences, School of Dentistry,
University of Otago, PO Box 647, Dunedin, New Zealand;
and 2Community Dental Health Services Research Unit,
Faculty of Dentistry, University of Toronto, 124 King
Edward Street, Toronto, ON, Canada M5G 1G6;
*corresponding author, email@example.com
J Dent Res 84(7):649-652, 2005
650 Foster Page et al.J Dent Res 84(7) 2005
group, it was decided that, for a power of 0.8 and a significance
level of 0.05, a sample size of 325 was needed. Assuming a 60%
response rate, a sample size of 542 was required, and this was
rounded up to 600.
Parents/caregivers of the sampled children were mailed
consent documentation and a questionnaire which sought
information on whether the child had received any orthodontic
advice and/or treatment, and which included the Parental-
Caregiver Perceptions Questionnaire that had been developed,
along with the CPQ11-14(www.cdhsru-uoft.ca/cohqol; Jokovic et
al., 2002). We obtained consent from both parent and child before
proceeding. Each child completed the CPQ11-14in the dental clinic
waiting room just prior to the dental examination; questions asked
about the frequency of events during the previous three months.
Response options and scores were: 'Never' (scoring 0); 'Once or
twice' (1); 'Sometimes' (2); 'Often' (3); and 'Every day or almost
every day' (4). An overall CPQ11-14score was computed by
addition of all of the item scores, and scores for each of the four
domains were also computed. The test-retest reliability of the
CPQ11-14was not examined.
The clinical examinations (by LFP) took place in dental
clinics at the children's schools. A standardized sequence was used,
with a standard dental caries examination (World Health
Organization, 1997) preceding an assessment for malocclusion.
Teeth were not cleaned and were examined wet. The
orthodontic assessment was carried out based on the
Dental Aesthetic Index (Cons et al., 1986), which
assesses the relative social acceptability of dental
appearance by collecting and weighting data on 10
intra-oral measurements. This enables each
individual to be placed on a dental appearance
continuum ranging from 13 (the most socially
acceptable) to 100 (the least acceptable), and
orthodontic treatment need can be prioritized based
on the pre-defined categories of 'minor/none' (scores
13 to 25), 'definite' (26 to 31), 'severe' (32 to 35), or
'handicapping' (36 or more; Estioko et al., 1994).
Where a child presented with a mixed dentition,
he/she was asked directly about the reason for any
missing teeth, since these are allocated the highest
DAI weight. Prior to data capture, the dental
examiner underwent a calibration session with an
experienced dental epidemiologist (WMT), resulting
in inter-examiner intraclass correlation coefficients
(ICC) of 0.98 for the DAI score, and 0.93 for DMFS.
We investigated intra-examiner reliability by
conducting replicate examinations on 19 individuals;
an ICC of 0.94 was obtained for the DAI score, and
0.94 for DMFS.
Clinical data were entered into a laptop
computer by a research assistant. The resulting data
were analyzed with the use of SPSS version 10.1
(SPSS Inc., Chicago, IL, USA). Descriptive statistics
were followed by bivariate analyses, which used
(where appropriate) Chi-square tests for comparison
of proportions, and Mann-Whitney or Kruskal-
Wallis tests (as appropriate) for comparison of the
means of continuous variables. The alpha value was
set at P < 0.05.
A simple random sample of 600 was selected with the use of
SPSS. Due to a clerical error, one child had been entered twice
on a school roll and was sampled twice; this meant that the
final random sample was effectively reduced to 599. Some 19
households could not be contacted by mail, and those
questionnaires were returned to the examiner. This resulted in
an effective sample of 580. The two mail-outs yielded a total of
435 children with parent/caregiver consent to be examined.
Five children were unable to be examined, due to three having
changed schools during the examination period, one child being
absent on each of three visits to the school, and one autistic
child being unable to complete his questionnaire. Based on the
effective sample, the study participation rate was 74.1%
Males slightly outnumbered females, and one in five
participants was Mäori (Table 1). Almost three-quarters of the
sample had had caries experience, and just over one-quarter
had 4+ DMFS. Mean DAI scores were higher among females
than males, but there were no significant sex differences by
treatment category membership.
Scores on the CPQ11-14ranged from 0 to 103 (Table 2), and
were positively skewed, as were the domain scores (requiring
the use of non-parametric tests of statistical significance).
While there was an apparent sex difference in overall
Table 1. Summary Data on Sample Characteristics
Males FemalesAll Combined
Mean age (SD)
Dental caries experience
Mean DMFS (SD)
Number with DMFS > 0 (%)
Number with DMFS > 3 (%)
Mean DAI score (SD)
Treatment need category
12.7 ( 0.5)
12.8 ( 0.4)
12.7 ( 0.5)
88 ( 20.5)
342 ( 79.5)
2.56 ( 4.15)
3.24 ( 4.34)
27.40 (6.73) 29.34 (8.79)a
P < 0.05.
Table 2. Descriptive Data on the CPQ and Subscales
Number of ItemsMean score (SD) Range of Observed Scores
35 17.3 (14.2) 0 to 103
0 to 19
0 to 27
0 to 32
0 to 28 12
J Dent Res 84(7) 2005 Validation of the CPQ651
CPQ11-14score (Table 3), it did not quite reach statistical
significance (P = 0.06); however, the mean emotional well-
being domain score for females was higher than that for
males. Children in the top quartile of the DMFS distribution
had higher CPQ11-14scores overall, as well as higher scores
for each of the four domains. There was a distinct gradient in
mean CPQ11-14scores across the categories of malocclusion
severity, whereby those in the 'Handicapping' category had
the highest and those in the 'Minor/none' category had the
lowest CPQ11-14score, on average. Such a gradient was also
observed with respect to the emotional well-being and social
well-being domain scores, but not with the other two
This cross-sectional study of the construct validity of the
CPQ11-14in a population-based sample of 12- to 13-year-olds
has found that it appears to be acceptable, showing higher
overall scores among children with more severe
malocclusions or with greater dental caries experience. In any
study which claims to have generalizable findings, the sample
representativeness should be closely examined. Comparison
of the sample's characteristics with the Taranaki child
population with the use of three markers (Census estimates
for gender and the Mäori population, and school dental
service caries data) suggested that the sample was indeed
This study set out to examine the construct validity of the
CPQ11-14. It was hypothesized that children with more severe
malocclusion would have higher scores. This was certainly the
case with the overall CPQ11-14scores, with a clear ascending
gradient demonstrated across ascending categories of
orthodontic treatment need. However, the domain scores
showed some noteworthy differences, with no clear,
statistically significant gradient observed for oral symptoms or
functional limitations; the emotional and social well-being
domain scores did show clear gradients, though. Concerning
dental caries experience, there were distinct differences in both
the overall and the domain scores between those who were in
the highest quartile for DMFS and the remainder. These
findings are not counter-intuitive: Other factors being equal,
children in the most severe disease quartile are likely (for
example) to have experienced more oral pain, had difficulties
in chewing, to have worried or been upset about their mouths,
or to have missed school due to their cumulative disease
experience. However, malocclusion is as much a social
phenomenon as an anatomical one, and the DAI was designed
specifically to assess the relative social acceptability of dental
appearance based upon public perceptions of dental aesthetics.
Thus, it is not surprising that clear gradients were observed
(across the ascending DAI treatment-need categories) for two
of the domains, since 'being teased' or 'avoiding smiling or
laughing' (social well-being) and 'being upset' or 'worrying
about being different' (emotional well-being) are known to be
associated with malocclusion, and are important motivating
factors in the uptake of orthodontic treatment (Plunkett, 1997).
That no clear gradients were observed with the oral symptoms
or functional limitations domains is also unsurprising, perhaps,
since only the most severe malocclusion might be expected to
produce effects in those domains.
Validation of measures such as the CPQ11-14at the
population level is important, since clinical samples may give a
misleading picture of their utility, because of the biased nature
of the sample (Locker, 2000). Further research should (a)
examine the validity of the CPQ11-14in other populations and
settings, and (b) investigate its evaluative properties to
determine its usefulness as a clinical outcome measure in dental
health services research.
We thank the NZ Dental Association Research Foundation and
Taranaki District Health Board for their support.
Table 3. CPQ Scores by Sex and Categories of Nalocclusion and Caries Severity (SD)
CPQ Overall ScoreCPQ Domain Scores
Oral Symptoms Functional Limitations Emotional Well-being Social Well-being
Dental caries categoryb
DMFS = 0
DMFS = 1
DMFS = 2 or 3
DMFS = 4+
P < 0.05.
The DMFS = 4+ category differed significantly from each of the other three DMFS categories for all except the Functional limitations and
Emotional well-being domain scores (one-way ANOVA).
652Foster Page et al.J Dent Res 84(7) 2005
Atchison KA, Dolan TA (1990). Development of the Geriatric Oral
Health Assessment Index. J Dent Educ 54:680-687.
Cons NC, Jenny J, Kohout FJ (1986). DAI: the Dental Aesthetic Index.
Iowa City, IA: College of Dentistry, The University of Iowa.
Cushing AM, Sheiham A, Maizels J (1986). Developing socio-dental
indicators—the social impact of dental disease. Community Dent
Estioko LJ, Wright FA, Morgan MV (1994). Orthodontic treatment
need of secondary school children in Heidelberg, Victoria: an
epidemiologic study using the Dental Aesthetic Index. Community
Dent Health 11:147-151.
Jokovic A, Locker D, Stephens M, Kenny D, Tompson B, Guyatt G
(2002). Validity and reliability of a questionnaire for measuring
child oral-health-related quality of life. J Dent Res 81:459-463.
Leão A, Sheiham A (1996). The development of a socio-dental
measure of dental impacts on daily living. Community Dent
Locker D (2000). Response and nonresponse bias in oral health
surveys. J Public Health Dent 60:72-81.
Locker D, Miller AM (1994). Subjectively reported oral health status
in an adult population. Community Dent Oral Epidemiol 22:425-
Plunkett DJ (1997). The provision of orthodontic treatment: some
ethical considerations. NZ Dent J 93:17-20.
Slade GD, Spencer AJ (1994). Development and evaluation of the Oral
Health Impact Profile. Community Dent Health 11:3-11.
World Health Organization (1997). Oral health surveys. Basic
methods. 4th ed. Geneva: World Health Organization.