| Sign In to gain access to subscriptions and/or personal tools. |
Validity and Reliability of a Questionnaire for Measuring Child Oral-health-related Quality of Life
1 Community Dental Health Services Research Unit, Faculty of Dentistry, University of Toronto, 124 Edward Street, Toronto, ON, Canada M5G 1G6; Correspondence: * corresponding author, aleksandra.jokovic{at}utoronto.ca
Oral-health-related quality of life measures that exist are designed for adults. This study aimed to develop and evaluate the CPQ11-14, a self-report measure of the impact of oral and oro-facial conditions on 11- to 14-year-old children. An item pool was generated with the use of a literature review and interviews with health professionals, parents, and child patients. The 36 items rated the most frequent and bothersome by 83 children were selected for the CPQ11-14. Validity testing involved a new sample of 123 children. Test-retest reliability was assessed in a subgroup of these children (n = 65). Mean CPQ11-14 scores were highest for oro-facial (31.4), lower for orthodontic (24.3), and lowest for pedodontic (23.3) patients. There were significant associations between the CPQ11-14 score and global ratings of oral health (p < 0.05) and overall well-being (p < 0.01). The Cronbach's alpha and intraclass correlation coefficient for the CPQ11-14 were 0.91 and 0.90, respectively. These results suggest that the CPQ11-14 is valid and reliable.
Key Words: child • oral health • quality of life • validity • reliability
Measures which address oral-health-related quality of life (OHRQoL) are being used with increasing frequency in oral health surveys and clinical trials in dentistry. They document the functional and psychosocial outcomes of oral disorders and are intended to supplement clinical indicators to provide a comprehensive account of the health of individuals and populations. To date, several OHRQoL measures have been developed (Cushing et al., 1986; Atchison and Dolan, 1990; Locker and Miller, 1994; Slade and Spencer, 1994; Leao and Sheiham, 1996). However, they all are designed to assess OHRQoL in adult populations. No comprehensive measure has yet been developed for children. Since pediatric oral disorders are numerous (Surgeon General's Report, 2000) and are likely to have a negative effect on the quality of life, there is a need for a measure that documents oral health outcomes in child populations. The absence of pediatric OHRQoL measures is due, in part, to the complex conceptual and methodological issues involved in the construction of self-report health status indicators for children (Pantell and Lewis, 1987; Pal, 1996). For example, the structure of children's self-concept and health cognitions is age-dependent as a result of their continuous cognitive, emotional, social, and language development. Similarly, the content of daily activities, understanding of feeling states, perceptions of relationships, and communication skills evolve with children's age (French and Christie, 1996; Theunissen et al., 1988). These age differences in the cognitive, emotional, functional, and behavioural characteristics must be accommodated within a child health status questionnaire (French and Christie, 1996; Pal, 1996). According to child developmental psychology, the age of 6 marks the beginning of abstract thinking and self-concept (Hetherington et al., 1999). Children start to compare their physical features and personality traits with those of other children or against a norm. Their ability for evaluative judgments of their appearance, the quality of friendships, and other people's thoughts, emotions, and behaviors gradually develops through middle childhood (6-10 yrs) (Bee, 1998; Hetherington et al., 1999). By the age of 11 or 12, children have clear understanding of complex emotions such as worry, shame, and jealousy, and their self-concept acquires sophisticated dimensions such as romantic appeal and popularity with peers (Bee, 1998). These years also correspond to the onset of puberty. Children then enter a period of early adolescence (11-14 yrs) that is characterized by the increasing centrality of peer crowd and clique dynamics in children's lives and their pre-occupation with others' views of self (Bee, 1998; Hetherington et al., 1999). Children develop the concept of time about the age of 8, when their recall period starts to lengthen and their understanding of frequency of events begins to emerge (French and Christie, 1996; Rebok et al., 2001). Given these developmental differences, a single standardized self-report health status measure for children of ages 6 to 14 cannot be developed. Instead, age-specific questionnaires for 6-7-, 8-10-, and 11-14-year-olds are required. These groups are homogeneous in terms of roles and cognitive abilities. Questionnaire biases, such as position bias and acquiescence bias, are commonly encountered with child respondents (Pantell and Lewis, 1987). These problems have given rise to concerns that children's reports of their health may not meet the psychometric standards expected of health outcome measures. However, recently developed generic and disease-specific measures, such as the Child Health Questionnaire (Landgraf et al., 1996) and the Childhood Asthma Questionnaires (French and Christie, 1996), have demonstrated that, with appropriate techniques, it is possible to obtain valid and reliable reports of HRQoL from children. This paper describes the development and validity and reliability testing of the Child Perceptions Questionnaire (CPQ11-14), a measure of the OHRQoL for children aged 11 to 14 yrs. The aim was to produce a measure which conformed to contemporary concepts of child health and had discriminative and evaluative properties, and which is applicable to children with various dental, oral, and oro-facial disorders. It forms one component of the Child Oral Health Quality of Life Questionnaire (COHQOL), which consists of a Parental-Caregiver Perceptions Questionnaire, a Family Impact Scale, and three age-specific questionnaires for children.
Participants Participants in the study were children aged 11 to 14 yrs with dental diseases (primarily caries), orthodontic disorders, and oro-facial conditions (primarily cleft lip and/or palate). They were recruited from pediatric dentistry and orthodontic clinics at the Faculty of Dentistry, University of Toronto, the Craniofacial Clinic at The Hospital for Sick Children, Toronto, and Toronto Public Health dental clinics. The study was approved by the Human Subjects Certification Committee, University of Toronto, the Research Ethics Board, The Hospital for Sick Children, Toronto, and the Research Office, Toronto Public Health. A parental written consent was obtained before a child's verbal assent was sought. A child's dissent superseded the parental consent.
Development Process Items for the final questionnaire were selected by means of an item impact study (Guyatt et al., 1986; Juniper et al., 1996). This identifies items on the basis of their frequency and importance to the target population. Participants were children from the three clinical groups defined above. Data were collected with the use of an item impact questionnaire administered in face-to-face interviews. The children were asked if in the past three months they had experienced the problem described by each item. Those responding positively then rated the importance of the problem on a 4-point Likert scale ranging from 0 ("Does not bother me at all") to 4 ("Bothers me very much"). For each item, we calculated an impact score by multiplying the percent of children with positive responses by the item's mean importance rating. Items were ranked within the four health domains (symptoms, functional limitations, emotional well-being, social well-being) according to their impact scores. Separate rankings were created for each clinical group. Items that were above the median in each ranking were selected for the final questionnaire: the CPQ11-14. In the CPQ11-14, questions asked about the frequency of events in the previous three months. The response options were: "Never" = 0; "Once/twice" = 1; "Sometimes" = 2; "Often" = 3; and "Every day/almost every day" = 4. Global ratings of the child's oral health and the extent to which the condition affected his/her overall well-being were also obtained. The questions were worded as follows: "Would you say that the health of your teeth, lips, jaws and mouth is...?" and "How much does the condition of your teeth, lips, jaws or mouth affect your life overall?" These global ratings had a five-point response format ranging from "Excellent" = 0 to "Poor" = 5 for oral health, and from "Not at all" = 0 to "Very much" = 5 for well-being.
Evaluation of the CPQ11-14 We generated an overall CPQ11-14 score by summing the response codes for the questionnaire items. Scores for each health domain were also computed. We assessed discriminant validity by comparing scores overall and domain scores for the three clinical groups. Because of the small within-group sample size and asymmetrical score distributions, the Kruskal-Wallis one-way analysis of variance was used. The variation in scores according to the severity of the child's condition was also examined insofar as this was feasible, given the clinical data that were collected. In the pedodontic group, the correlation between the overall score and the number of decayed tooth surfaces (DS) at the time of recruitment was assessed. In the oro-facial group, differences in mean scores between the oral cleft diagnostic categories were tested. To examine construct validity, we determined the associations between questionnaire scores and the two global indicators. Internal consistency reliability was tested by means of Cronbach's alphas, and test-retest reliability by means of intraclass correlation coefficients calculated by the one-way analysis of variance random-effects parallel model (Bartko, 1966; Shrout and Fleiss, 1979).
Characteristics of Participants Eighty-three children were interviewed for the item impact study, and 123 participated in the validity and internal consistency reliability study. Seventy of the latter completed the CPQ11-14 two times and provided data for the assessment of test-retest reliability. The characteristics of children in terms of clinical group, gender, and age are shown in Table 1  .
Item Impact Scores Appendix 1 (www.dentalresearch.org) shows, for each item in the modified item pool, its prevalence, mean importance rating, and the impact score in the sample as a whole. The 36 items that were selected for the final questionnaire are indicated with superscript a. They consist of those items that were ranked above the median within each of the four health domains in at least one of the clinical groups. In general, they are the highest prevalence items.
CPQ11-14 Descriptive Statistics
CPQ11-14 Discriminant and Construct Validity
Analysis within clinical groups was limited by the fact that clinical data were not available for all participants. As hypothesized, within the pedodontic group there was a significant positive correlation between overall scale scores and the number of decayed tooth surfaces (DS) (r = 0.64; p = 0.007). Within the oro-facial group, the mean score for children with either isolated cleft of the lip or isolated cleft of the palate (n = 11) was 27.1 compared with 33.1 for children with either unilateral or bilateral cleft of the lip and palate (n = 18), as expected. The difference was not statistically significant, which could be due to the small numbers in each group. Testing of discriminant validity for the orthodontic group could not be undertaken, because the diagnostic categories were in the form of a nominal rather than an ordinal scale.
Hypotheses concerning construct validity were confirmed—that is, there were significant positive correlations between CPQ11-14 scale scores and global ratings of oral health (p = 0.013) and overall well-being (p < 0.001). The rank correlation coefficient was higher for the rating of overall well-being (r = 0.40) than the rating of oral health (r = 0.23). Significant correlations were also observed between the scores for all health domains and the global ratings of oral health and overall well-being, except between the functional limitations score and the oral health rating (Table 3
CPQ11-14 Reliability Cronbach's alpha for the sample as a whole was 0.91 (Table 4 ). For the domains, the coefficient ranged from 0.64 for oral symptoms to 0.86 for emotional well-being, indicating acceptable to good internal consistency reliability (Streiner and Norman, 1994). Alphas for the CPQ11-14 were 0.93 for the pedodontic group, 0.89 for the orthodontic group, and 0.91 for the oro-facial group.
The reliability of the questionnaire was based on data from 65 children. The remaining five cases were excluded because they reported that their oral health and/or overall well-being had changed between the two administrations of the questionnaire. The intraclass correlation coefficient (ICC) for the CPQ11-14 was 0.90, indicating almost perfect agreement (Landis and Koch, 1977), while for the domains it ranged from 0.79 to 0.88, indicating substantial to almost perfect agreement (Table 4  ). The ICCs for the pedodontic, orthodontic, and oro-facial groups were 0.89, 0.83, and 0.84, respectively.
Early attempts at measuring the HRQoL of children used parents as proxies. This was intended to avoid the methodological problems involved in obtaining self-reports concerning health and quality of life from children. However, it is now recognized that parents' knowledge of their children may be limited, so that parental and child reports of health and well-being may not correspond (Levi and Drotar, 1999; Sawyer et al., 1999). While parents are useful as informants, their reports are now regarded as complementary to and not substitutes for child reports (Guyatt et al., 1997; Theunissen et al., 1998; Parsons et al., 1999; Le Coq et al., 2000). Since recent research has demonstrated that children's reports of their health-related quality of life are valid and reliable, child questionnaires should always be used in the documentation of outcomes of specific clinical conditions. Accordingly, this paper describes the development and evaluation of a questionnaire to measure the OHRQoL of children aged 11 to 14 yrs who have a variety of dental, orthodontic, and oro-facial conditions. This questionnaire consists of 36 items organized into four health domains. The development process used ensured that these items concerned problems that were not only the most frequent but also created the most bother for the children concerned. This means that the questionnaire conforms to a central requirement of HRQoL instruments as defined by Guyatt et al. (1991) and Guyatt and Cook (1994)—that is, it addresses aspects of daily life that patients value. Analysis of the data presented here shows that the questionnaire is valid and has excellent reliability. It also indicates that the impact of child oral and oro-facial conditions on functional and psychosocial well-being is substantial, and that children are able to give psychometrically acceptable accounts of that impact. In this regard, the questionnaire performs well as a discriminant measure. However, since we intended the questionnaire to be used as an outcome measure in clinical trials, its evaluative properties need to be assessed. Longitudinal studies are then necessary to determine its longitudinal construct validity, responsiveness, and minimal clinically important difference.
The research on which this paper was based was supported by The Hospital for Sick Children Foundation, Grant No. XG99-085. Received for publication November 13, 2001. Revision received May 6, 2002. Accepted for publication May 15, 2002.
Journal of Dental Research, Vol. 81, No. 7,
459-463 (2002) This article has been cited by other articles:
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
TOP
ABSTRACT
INTRODUCTION
