Abstract
Objectives:
To develop and evaluate the psychometric properties of child- and parent-proxy measures of oral health aesthetics.
Methods:
Items that describe children’s perceptions of their oral attractiveness and its impact on social, emotional, and behavioral functioning were developed based on a systematic review of existing measures, clinician feedback (n = 13), and child semi-structured interviews (n = 27). The tools’ content validity was assessed in cognitive interviews with 21 children. Items were administered to socio-demographically diverse samples of 998 children ages 8–17 years and 626 parents of children ages 5–17 years. Psychometric methods were used to finalize and calibrate item banks, generate short questionnaire forms, and evaluate the tools’ reliability, precision, and validity.
Results:
The item banks and their short forms provide precise measurement across a wide range of oral health aesthetic states. They measure relevant and meaningful positive and negative experiences using terminology that most children as young as 8 years of age can understand. Known-group comparisons and convergence with existing measures of oral health-related quality of life, global health, and body image provide evidence of construct validity. The scores are interpretable relative to the US general population.
Conclusions:
The oral health aesthetic item banks and short forms provide precise and valid assessments of children’s satisfaction with their oral appearance. They may be useful for targeting and evaluating pediatric dental and orthodontic care in clinical practice and research settings.
Aesthetic features of the mouth, teeth, lips, and gums significantly influence overall facial appearance and others’ judgements of one’s physical and social attractiveness.1,2 In children, self-perceptions of oral attractiveness influence overall self-concept, social acceptance, peer relationships, and school performance.1–3 Oral health aesthetics is a stronger predictor of self-concept than objectively defined malocclusion.3 Patient-reported outcome measures (PROMs) are increasingly used in oral health research and clinical care, a trend that reflects a shift from biomedical to broader biopsychosocial perspectives on oral health and an increased emphasis oral health-related quality of life (OHRQoL).4
The most commonly used oral PROMs for children are the Child Perceptions Questionnaire (CPQ),5,6 Child Oral Impacts on Daily Performances Index (C-OIDP),7,8 Child Oral Health Impact Profile (COHIP),9,10 and the PedsQL Oral Health Scale.11 These tools assess multiple OHRQoL domains, but only the COHIP and the PedsQL-Oral Health assess aesthetics using items that are embedded in scales with other social-emotional aspects of OHRQoL. Only one measure (the COHIP) includes positively-valanced items (e.g., felt attractive), which may be important for detecting meaningful improvements and identifying ways to optimize OHRQoL. Moreover, child-report OHRQoL measures lack sufficient evidence of content validity, because they were developed without substantial input from children.12 Important oral health aesthetics experiences may be omitted and items may be irrelevant or difficult for children to understand.12 Lastly, none of the tools were developed using item-response theory (IRT)-based measurement approaches. IRT-based methods characterize a scale’s precision at varying levels of the measured concept.
This manuscript describes the development of parallel child- and parent/proxy-report measures of oral health aesthetics. In contrast to the global multidimensional PROMs described above, these tools provide in-depth information about a single, high-priority, and clinically actionable dimension of OHRQoL. The instruments were developed using a rigorous, mixed-method approach that complies with scientific standards.13 We emphasized content validity by developing and iteratively refining the tools based on qualitative information obtained directly from children and pediatric dentists. Thereafter, we used a combination of classical and contemporary methods to refine, calibrate, and evaluate the tools’ psychometric properties in a large US-based sample.
Methods
All procedures involving human subjects were reviewed and approved by the institutional review board at the Children’s Hospital of Philadelphia (IRB protocol #13-010096).
Item pool development
Literature review.
Existing oral health PROMs were identified through a systematic literature review.1 Search terms captured oral health experiences (e.g., dentistry), self-report instruments (e.g., questionnaire), and measurement (e.g. reliability). Using MEDLINE, the query was limited to studies that included children (< 18 years old) and articles written in English. Oral health concepts reflected in each PROM item were derived from the 25 most commonly cited PROMs. Four investigators sorted oral aesthetics concepts into smaller conceptual sub-categories called facets. This process yielded a preliminary concept map, which provided an organizing structure for the PROM.
Clinician and child semi-structured interviews.
The oral health aesthetics concept map was refined based on interviews conducted with 13 pediatric dentists and 27 children aged 8–17 years (Table 1). Dentists were members of the American Dental Association Dental Quality Alliance.15 They provided feedback on concept labels, definitions, and organization. Children were recruited from a large dental clinic affiliated with the University of Pennsylvania School of Dental Medicine. They were asked to describe past problems with their teeth or mouth and the causes and consequences of those problems, as well as experiences of having a healthy mouth and teeth. We derived oral aesthetic experience concepts through thematic analysis of interview transcripts until concept saturation was achieved, and refined the concept map to reflect the identified experiences (Table 2).16
Table 1.
Sample characteristics
| Semi-structured interviews |
Cognitive interviews |
Psychometric Analyses |
||
|---|---|---|---|---|
| Child n=27 |
Child n=21 |
Child n=998 |
Parent n=626 |
|
| Child age | ||||
| 5–7 years | 2 (7.4) | -- | -- | 130 (20.8) |
| 8–11 years | 14 (51.9) | 16 (76.2) | 400 (40.1) | 199 (31.8) |
| 12–17 years | 11 (40.7) | 5 (23.8) | 598 (59.9) | 297 (47.4) |
| Child gender | ||||
| Male | 11 (40.7) | 11 (52.4) | 509 (51.0) | 306 (48.9) |
| Female | 16 (59.3) | 10 (47.6) | 489 (49.0) | 320 (51.1) |
| Race/ethnicity | ||||
| White/Non-Hispanic | 4 (14.8) | 3 (14.3) | 627 (62.8) | 399 (63.7) |
| Black/Non-Hispanic | 16 (59.3) | 13 (61.9) | 88 (8.8) | 53 (8.5) |
| Other/Non-Hispanic | 4 (14.8) | 3 (14.3) | 42 (4.2) | 22 (3.5) |
| 2+ races, Non-Hispanic | 1 (3.7) | 1 (4.8) | 57 (5.7) | 39 (6.2) |
| Hispanic | 2 (7.4) | 1 (4.8) | 184 (18.4) | 113 (18.1) |
| Annual household income | ||||
| ≤$29,999 | -- | -- | 178 (17.8) | 112 (17.9) |
| $30,000-$59,999 | -- | -- | 258 (25.9) | 163 (26.0) |
| $60,000-$99,999 | -- | -- | 327 (32.8) | 198 (31.6) |
| >$100,000 | -- | -- | 235 (23.6) | 153 (24.4) |
| Parent education | ||||
| Less than high school | -- | -- | 29 (2.9) | 17 (2.7) |
| High school diploma/GED | -- | -- | 150 (15.0) | 92 (14.7) |
| Some college | -- | -- | 358 (35.9) | 211 (33.7) |
| Bachelor’s degree or more | -- | -- | 461 (46.2) | 306 (48.9) |
| Family below poverty line | -- | -- | 158 (15.8) | 101 (16.1) |
| Child has special healthcare need | -- | -- | 260 (26.1) | 153 (24.4) |
| Dental insurance status | ||||
| Private insurance | -- | -- | 663 (66.4) | 409 (65.3) |
| Public insurance | -- | -- | 253 (25.4) | 160 (25.6) |
| Uninsured | -- | -- | 76 (7.6) | 55 (8.8) |
| Delayed or unable to get dental care (past 12 months) | -- | -- | 144 (14.4) | 84 (13.4) |
Table 2.
Oral health aesthetics concept saturation and representation in existing PRO measures, initial item pools, and final item banks
| Child Concept Elicitation Interviews |
PROM items |
# of items | |||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | n | % | Initial pool | Final bank | |
|
|
|||||||||||||||||||||
| Impact: positive emotions | x | x | x | x | x | x | 3 | 2% | 2 | 2 | |||||||||||
| Impact: social interactions/activities | x | x | x | 27 | 20% | 3 | 2 | ||||||||||||||
| Impact: bully victimization | x | x | x | 10 | 7% | 5 | 3 | ||||||||||||||
| Smiling/laughinga | x | x | 0 | 0% | 4 | 2 | |||||||||||||||
| Satisfaction: oral structures | x | x | x | x | x | x | x | x | x | x | x | 21 | 15% | 14 | 10 | ||||||
| Impact: body image | x | x | x | 9 | 7% | 5 | 3 | ||||||||||||||
| Impact: negative emotions | x | x | x | 27 | 20% | 4 | 4 | ||||||||||||||
| Hiding mouth | x | 19 | 14% | 2 | 1 | ||||||||||||||||
| Bad breathb | x | 12 | 9% | 1 | 0 | ||||||||||||||||
| Satisfaction: reflectionsc | 8 | 6% | 2 | 2 | |||||||||||||||||
Notes: PROMs = number and percentage of 136 patient reported outcome (PRO) items that assess each oral health aesthetics concept; item pools were subjected to qualitative and psychometric testing to yield final item banks
concepts not included in existing PRO measures,
concept excluded from final item pools,
concept not identified by children.
Item pool development.
We transformed each concept into item expressions that met the following criteria: items could stand alone without reference to any other item; item context was “In the past 4 weeks…”; items asked about the magnitude of aesthetics experiences using 5-point Likert scale response options: Not at all (1) - A little bit (2) – Somewhat (3) - Quite a bit (4) – Very much (5); and items were as concise and simply worded as possible. Child-report items were transformed into parallel parent/proxy-report items.
Cognitive interviews.
We tested item applicability and understandability (content validity) through cognitive interviews with 21 children aged 8–17 years and 17 parents of children aged 5–17 years (Table 1).17,18 Participants from a large urban dental clinic completed about 20 oral health aesthetics items. Then, using standardized probes, interviewers asked participants to read each question aloud, state the item’s meaning in their own words, and explain their response. Interviewees’ understanding of each item was coded as demonstrating poor (1), partial (2), or full (3) understanding. Each item was tested with at least 5 children and 5 parents. Items with average ratings of less than 2 were eliminated.
Psychometric evaluation
The item pools (42 items each) were administered to 998 children ages 8–17 years and 626 parents of children aged 5–17 years. Participants were recruited from GfK Knowledge panel, a dual frame (random-digit dial and address-based) online panel of a representative random sample of the U.S. population.19 Adult panelists who were known to have a child aged 5– 17 years were notified by e-mail of their eligibility to participate in the study. For panelists with more than one child within the target age range, one child was randomly selected for possible inclusion. To ensure adequate representation across developmental levels, we enrolled participants until child age quotas were met: 100 children per child year of age (8–17 years) and 50 parents per child age year (5–17 years). Children were excluded if their parent reported that they had a cognitive limitation that prevented them from accurately responding to survey questions.
Children completed the COHIP Short Form-19, a measure of oral symptoms (e.g., pain) and their impact on functioning (e.g., chewing) and socio-emotional wellbeing (e.g., mood, friendships, self-image).9,10,20 Children and parents completed the Healthy Pathways Body Image Scale21,22 and the PROMIS® Pediatric Global Health scale, a measure of children’s overall physical, mental, and social wellbeing.23 Parents completed the Children with Special Health Care Needs Screener, a measure of chronic health problems that require health services or cause functional limitations.24 Parents reported dental insurance status (private insurance, public insurance, or uninsured) and responded to two questions about difficulties or delays in obtaining dental care in the past 12 months. The questions were modified items from the Household Survey, Access to Care Section of the Medical Expenditure Panel Survey.25 Children reported their age and gender and parents provided information on race/ethnicity, household composition, annual household income, and parental educational attainment. Households were categorized as above or below the poverty threshold using 2015 federal poverty guidelines.26 The data were weighted, so that the weighted sample’s distributions of gender, age, race/ethnicity, education, U.S. Census region, metropolitan area, household internet access, and language (English/Spanish) matched those in the most recent Current Population Survey (CPS).27
Item bank reduction and calibration.
Item pools included both positively (e.g., happy) and negatively (e.g., worried) worded items. Given the known challenge of operationalizing a single dimension with items that have opposing valances,28 we evaluated the dimensionality of the 42 items using a bifactor model that specified a general factor (all items, G) and two specific factors: S1 (positive items) and S2 (negative items). Separate models were fit to the child- and parent/proxy-report data using Mplus version 8.1.29 Item responses were treated as ordinal and we used the weighted least squares, mean- and variance-adjusted (WLSMV) estimator. We evaluated model fit according to widely used model fit index criteria (RMSEA (Root Mean Square Error of Approximation) ≤ 0.08, TLI (Tucker-Lewis Index) ≥ 0.95, CFI (Confirmatory Fit Index) ≥ 0.95.30 Items were considered locally dependent if constraining the residual correlation between the items to 0 resulted in poor fit, as identified through modification indices.31 When observed, one of the two locally dependent items was removed.
After arriving at the final factor model, to be consistent with the larger IRT literature, we estimated the final model using a logistic link and full information maximum likelihood estimation. Given that we expect that end users’ primary interests would be in the general factor and given that most end users will likely not have access to the statistical software needed to estimate scores from a bifactor IRT model, we followed procedures described by Stucky et al to compute the marginal IRT parameters for the general dimension using the bifactor IRT model’s results.32 Unlike the multidimensional bifactor parameters, which require conditional interpretation, one can interpret the marginal parameters similarly to those from a unidimensional IRT model.32 The marginal discrimination parameter for the general dimension gives the relationship between an item response and the general dimension, given that the specific factors are fixed at zero (their mean). The marginal location parameter for the general dimension is the level of the general dimension where the probability of endorsing a particular response category is 0.5, given the specific factors are fixed at zero (their mean).
Fixed length short forms.
Highly discriminating items that provided precision across most of the full range of oral health aesthetics were selected for inclusion on 8 item (SF-8) and 4 item (SF-4) short forms.
Scoring.
Full bank, SF-8, and SF-4 theta (Θ) scores based on both the bifactor IRT parameters and the marginal IRT parameters were calculated for each respondent using the Bayesian Expected A Posteriori (EAP) estimation. Theta scores were transformed into T-scores: T=(Θ*10) + 50. A score of 50 theoretically corresponds to the average level of oral health aesthetics for children based on a national sample.
Reliability and precision.
The internal consistency reliability of item banks and SFs were evaluated using Cronbach’s alpha statistic. We evaluated reliability by computing marginal reliability curves as suggested by Green, et al.33 These curves present the test information (a measure of reliability and precision) in the metric of Cronbach’s alpha.
Construct validity.
We used Pearson Product Moment correlation to estimate associations between scores on the oral health aesthetics measure’s general factor and other PROMs. We hypothesized that the general oral health aesthetics would be positively associated with COHIP socio-emotional wellbeing, and to a lesser degree with the more general measures of body image and global health. We conducted separate regression analyses using child- and parent/proxy-report data to test hypothesized associations between child age, special healthcare need status, and access to oral healthcare (predictors) and the general oral health aesthetics (criteria). We expected older children, those with special healthcare needs, and those without access to oral healthcare to have poorer oral health aesthetics. Because the construct validity results differed little across scores for the general oral aesthetic satisfaction based on the full bi-factor model and those based on the marginal IRT parameters, we report results based on the marginal IRT parameters.
Results
Literature review.
The literature search yielded 268 citations for articles that described 148 unique pediatric oral health PROMs. Each PROM was described in an average of 2.6 articles (SD = 5.6, range: 1–48 articles). A total of 470 item concepts were derived from the 25 most commonly identified PROMs. Of these, 136 concepts (28.9%) described satisfaction with the appearance of one’s teeth/mouth. Item concepts represented 9 unique oral aesthetics facets: (1) satisfaction with specific oral structures (e.g., teeth, mouth, lips); (2) hiding mouth; (3) satisfaction with appearance in reflections (e.g., photos, mirrors), (4) bad breath; and the impact of oral aesthetics on (5) positive emotions, (6) social interactions and activities, (7) bully victimization, (8) body image, and (9) negative emotions. There was significant redundancy in item content within most facets. For example, 19 items described hiding one’s mouth or teeth because of dissatisfaction with appearance.
Semi-structured interviews.
The concept maps comprised of the 9 oral health aesthetics facets were reviewed during semi-structured interviews with pediatric dentists. Dentists recognized the critical need for reliable and valid PROMs of children’s oral health aesthetics. They thought the clinical utility of oral heath PROMs would be enhanced if items specified the root causes of children’s PRO experiences. Many existing PROMs include items that reflect general concerns about one’s teeth or mouth (e.g., “worried about my teeth”), but dentists indicated that measures would be more useful if they attributed these general experiences to specific aspects of oral health (e.g., “I worried about the way my teeth looked”). Dentists indicated that physical and socio-emotional consequences are important components of children’s oral health aesthetics.
The child semi-structured interview saturation matrix is shown in Table 2. Seventeen children described satisfaction or concerns with the way their teeth or mouth looked and the impacts of these perceptions on other aspects of their health and functioning. Children most commonly described oral health aesthetics in terms of satisfaction with certain characteristics of specific oral structures (e.g., “having straight teeth”). Some children indicated a desire for cosmetic interventions such as tooth whitening and orthodontia, concepts that were missing from existing PROMs. Children who were dissatisfied with their oral appearance commonly described covering their teeth/mouth or refusing to smile. They reported feeling embarrassed, sad, and being made fun by peers because of oral aesthetics. Other children identified novel positive socio-emotional consequences oral appearance, specifically wanting to “show off” their teeth by smiling or laughing. We added this indicator of oral health aesthetics to the concept map.
Item pool development.
Upon completion of semi-structured interviews, the oral health aesthetics concept map included 10 facets and 42 item concepts, which were transformed into item expressions.
Cognitive interviews.
All 42 oral aesthetics items were adequately understood in cognitive interviews and were retained without revisions.
Psychometric analyses.
Item-level statistics are shown in Table 3 (child-report) and Supplemental Table S2 (parent/proxy-report). Child-report item-level means ranged from 1.07 (“I had a hard time making friends because of how my teeth looked”) to 3.79 (“I was happy with how my lips looked”). Two items (“I wanted to make my teeth whiter” and “I had bad breath”) were removed due to low factor loadings on both the general (< 0.4) and specific (negative) factors (< 0.2). Eleven pairs of similarly-worded items were locally dependent. We eliminated one item from each pair, retaining the item that most specifically attributed the experience to oral aesthetics. For example, we retained “I liked to smile because of my teeth” and eliminated “I liked to smile because of how my teeth looked.”
Table 3.
Child-report item descriptive statistics, factor loadings, and marginal IRT parameters
| Descriptive Statistics |
CFA factor loadings |
Marginal IRT parameters |
||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Item. | M | SD | % Min (Not at all) | % Max (Very much) | G | S1 | S2 | a | b1 | b2 | b3 | b4 |
| I felt proud because of my teeth.1,SF4,SF8 | 3.06 | 1.42 | 19% | 22% | 0.68 | 0.48 | 1.71 | −1.27 | −0.68 | 0.13 | 0.89 | |
| I felt happy because of my teeth.1 | 3.34 | 1.37 | 12% | 26% | 0.85 | 0.40 | 2.96 | −1.32 | 0.79 | 0.14 | 0.55 | |
| I had a hard time making friends because of how my teeth looked.2 | 1.07 | 0.42 | 95% | 0% | 0.55 | 0.72 | 1.24 | 3.82 | 3.27 | 2.65 | -- | |
| I felt shy because of how my teeth looked.2 | 1.27 | 0.73 | 82% | 1% | 0.70 | 0.57 | 1.72 | −2.57 | −1.92 | −1.25 | -- | |
| People made fun of me because of how my teeth looked.3 | 1.18 | 0.60 | 87% | 1% | 0.57 | 0.67 | 1.25 | −3.27 | −2.77 | −1.85 | -- | |
| Kids teased me because of how my teeth looked.3,SF8 | 1.14 | 0.55 | 91% | 1% | 0.60 | 0.66 | 1.31 | −3.13 | −1.99 | -- | -- | |
| People stared at me because my teeth looked bad.3 | 1.12 | 0.53 | 92% | 1% | 0.58 | 0.69 | 1.32 | −3.37 | −2.81 | −2.17 | -- | |
| I liked to smile because of how my teeth looked.4,SF4,SF8 | 3.27 | 1.38 | 13% | 24% | 0.81 | 0.49 | 2.52 | −1.30 | −0.72 | −0.09 | 0.66 | |
| I liked to laugh because of how my teeth looked.4 | 2.96 | 1.50 | 24% | 21% | 0.74 | 0.39 | 2.00 | −1.01 | −0.65 | 0.07 | 0.81 | |
| My mouth looked good.5 | 3.72 | 1.23 | 6% | 33% | 0.83 | 0.26 | 2.87 | −1.71 | −1.11 | −0.45 | 0.34 | |
| I was happy with how my teethlooked.5,SF4,SF8 | 3.57 | 1.31 | 9% | 31% | 0.91 | 0.18 | 4.30 | −1.43 | −0.94 | −0.33 | 0.35 | |
| I was happy with how my mouth looked.5 | 3.62 | 1.28 | 7% | 32% | 0.94 | 0.14 | 5.47 | −1.46 | −0.94 | −0.31 | 0.34 | |
| I was happy with how my lips looked.5 | 3.79 | 1.24 | 6% | 36% | 0.71 | 0.08 | 1.90 | −2.11 | −1.48 | −0.64 | 0.34 | |
| I wanted to make my teeth straighter.5,SF8 | 2.09 | 1.39 | 50% | 11% | 0.56 | .45 | 1.16 | −2.16 | −1.57 | −0.81 | −0.01 | |
| I was happy because my teeth are straight.5 | 3.29 | 1.47 | 17% | 29% | 0.75 | 0.37 | 2.03 | −1.19 | −0.74 | −0.13 | 0.54 | |
| I was happy with the color of my teeth.5 | 3.51 | 1.27 | 8% | 28% | 0.75 | 0.17 | 2.10 | −1.75 | −1.03 | −0.22 | 0.64 | |
| I was happy with the amount of space between my teeth.5 | 3.48 | 1.38 | 12% | 31% | 0.80 | 0.15 | 2.41 | −1.43 | −0.91 | −0.26 | 0.53 | |
| I was happy with the size of my teeth.5 | 3.76 | 1.25 | 5% | 36% | 0.84 | 0.04 | 2.89 | −1.77 | −1.20 | −0.47 | 0.30 | |
| I wanted to change the size of my teeth.5 | 1.27 | 0.74 | 81% | 1% | 0.51 | 0.49 | 1.13 | −3.93 | −3.15 | −2.36 | −1.56 | |
| I looked good because of my teeth.6 | 3.28 | 1.34 | 12% | 23% | 0.79 | 0.42 | 2.31 | −1.40 | −0.79 | −0.09 | 0.75 | |
| I felt handsome or pretty because of how my teeth looked.SF8 | 3.20 | 1.43 | 16% | 24% | 0.84 | 0.44 | 2.74 | −1.10 | −0.68 | −0.07 | 0.60 | |
| I felt ugly because of how my teeth looked.6,SF4,SF8 | 1.25 | 0.75 | 84% | 2% | 0.76 | 0.51 | 2.14 | −2.75 | −2.16 | −1.81 | −1.31 | |
| I felt stressed because of how my teeth looked.7 | 1.31 | 0.78 | 81% | 1% | 0.63 | 0.52 | 1.51 | −2.15 | −1.39 | -- | -- | |
| I felt sad because of how my teeth looked.7 | 1.21 | 0.68 | 85% | 1% | 0.70 | 0.64 | 1.69 | −2.58 | −2.00 | −1.48 | -- | |
| I worried about how my teeth looked.7 | 1.43 | 0.89 | 72% | 2% | 0.67 | 0.58 | 1.56 | −1.78 | −0.87 | -- | -- | |
| I was embarrassed because of how my teeth looked.7 | 1.29 | 0.80 | 81% | 2% | 0.77 | 0.58 | 2.08 | −2.07 | −1.78 | −1.16 | -- | |
| I hid my mouth because of how my teeth looked.8,SF8 | 1.22 | 0.73 | 86% | 1% | 0.72 | 0.57 | 1.86 | −2.72 | −2.36 | −1.88 | −1.42 | |
| I liked how my teeth looked in the mirror.9 | 3.50 | 1.29 | 8% | 28% | 0.90 | 0.23 | 3.94 | −1.42 | −0.78 | −0.23 | 0.49 | |
| I liked how my teeth looked in pictures.9 | 3.52 | 1.34 | 10% | 30% | 0.91 | 0.21 | 4.14 | −1.31 | −0.89 | −0.30 | 0.37 | |
Impact: positive emotions
Impact: social interactions/activities
Impact: bully victimization
Smiling/laughing
Satisfaction: oral components
Impact: body image
Impact: negative emotions
Hiding mouth
Satisfaction: reflections
items included on 8-item short form
items included on 4-item short form
All items begin with, “In the past 7 days…”
The final bifactor model adequately fit the data for the remaining 29 items by child-report (CFI = 0.99; TLI = 0.99; RMSEA = 0.03) and parent/proxy-report (CFI = 0.99; TLI = 0.99; RMSEA = 0.03). We also fit the final model using a logistic link and full information maximum likelihood estimation to generate bifactor IRT parameters (Tables S1 and S3). We then computed the general dimension’s marginal IRT parameters using the bifactor IRT model’s results (Tables 3 and S2).
Item category probability curves (available upon request) showed that for 9 of the 13 negatively-oriented child-report items, response categories indicative of the greatest dissatisfaction measured the same level of oral health aesthetics. We combined the two highest response categories for 6 items and the three highest response categories for 3 items. The item “I was happy with how my mouth looked” best differentiated children with varying oral health aesthetics (a = 5.47). The range of threshold parameters was −3.82 (very much: I had a hard time making friends because of how my teeth looked) to 0.89 (very much: I felt proud because of my teeth). Items selected for the SF-8 and SF-4 are shown in Tables 3 and S2. The marginal reliability curves for the full child-report item bank and short forms illustrate acceptable levels of precision across a wide range of the latent variable: 4.5 standard deviations (SDs) for the full bank, 3.5 SDs for the SF-8, and 2.5 SDs for the SF-4 (Figure 1). By parent/proxy-report, the measures have adequate precision across 4 SDs for the full bank and 2.5 SDs for both the SF-8 and SF-4 (Supplemental Figure S1).
Figure 1.
Child-report oral health aesthetics marginal reliability curves
As expected, oral health aesthetics scores were positively associated with scores on the COHIP, PROMIS PGH-7, and the Healthy Pathways Body Image scale. Associations were strongest for the COHIP social-emotional wellbeing scale, which includes items about satisfaction with the appearance of one’s teeth, mouth, and face.9,10,20 Children with special healthcare needs and older adolescents reported poorer oral health aesthetics. Children who were delayed or unable to access dental care in the past 12 months had poorer child- and parent/proxy-reported oral health aesthetics (Table 4 and Supplemental Table S4).
Table 4.
Child-report oral health aesthetics full bank and short form descriptive statistics, internal consistency reliability, and validity
| Full bank | SF-8 | SF-4 | |
|---|---|---|---|
|
|
|||
| Descriptive Statistics | |||
| Mean | 49.6 | 49.4 | 49.3 |
| SD | 9.9 | 9.4 | 9.1 |
| Minimum | 15.0 | 20.3 | 26.5 |
| Maximum | 69.9 | 65.6 | 63.9 |
| Skew | 0.16 | −0.01 | −0.03 |
| Internal consistency reliability (α) | 0.96 | 0.86 | 0.81 |
| Convergent validity (r)a | |||
| Child-report scales | |||
| Oral healthb | 0.51 | 0.50 | 0.47 |
| Functional wellbeingb | 0.42 | 0.39 | 0.35 |
| Social-emotional wellbeingb | 0.69 | 0.66 | 0.62 |
| Global healthc | 0.49 | 0.44 | 0.43 |
| Body imaged | 0.43 | 0.37 | 0.35 |
| Parent/proxy-report scales | |||
| Global healthc | 0.40 | 0.36 | 0.36 |
| Body imaged | 0.35 | 0.31 | 0.29 |
| Construct validity (B)e | |||
| (Intercept) | 54.11 | 53.18 | 52.86 |
| Child age | −0.22* | −0.17+ | −0.18+ |
| Special healthcare need | −1.59* | −1.40* | −1.27* |
| Delayed or unable to get dental care in past 12 months | −2.47*** | −2.26*** | −1.84** |
all correlations are statistically significant, p < 0.0001
COHIP
PROMIS® PGH-7
Healthy Pathways Body Image Scale
unstandardized regression coefficients; Full bank: F(3,965) = 8.93***, R2 = 0.03; SF-8: F(3,965) = 7.81***, R2 = 0.02; SF-4: F(3,965) = 6.19***, R2 = 0.02
p < 0.10
p<0.05
p < 0.01
p < 0.001.
Discussion
Many of the most commonly-used oral health PROMs for children assess broad dimensions of OHRQoL without a specific focus on aesthetics or its impact on other aspects of health and functioning. We developed parallel child- and parent/proxy-report oral health aesthetics item banks and short forms to meet this need. Item bank content was informed by a systematic review of existing pediatric oral health PROMs, input from pediatric dentists, and children’s first-hand descriptions of their oral aesthetic experiences. About half of the items from existing PROMs characterized specific oral structures (e.g., happy with color of teeth) or thoughts and behaviors indicative of dissatisfaction with oral appearance (e.g., hiding mouth). Remaining items from existing PROMs assessed the perceived impact of oral appearance on social and psychological functions. Nearly all existing impact items were negatively constructed. They assessed social disruptions, bullying, and negative emotional responses to perceived oral unattractiveness. In contrast, our child interviewees commonly identified positive emotional responses such as feeling happy and proud of the way their teeth or mouth looked. A few children described wanting to smile and laugh to “show off their teeth,” an experience that no existing pediatric PROM assesses. Including positive indicators of oral health aesthetics improves the tools’ meaningfulness and relevance for children. It also supports assessment of the full range of outcomes targeted by dental and orthodontic care, which focus on both ameliorating symptoms and maximizing positive health. Cognitive interview results supported the tools’ content validity, an essential measurement property that has not been rigorously evaluated for most other pediatric oral health PROMs.12
We evaluated the psychometric properties of the child- and parent/proxy-report item pools in a large national sample. We eliminated items that attributed social or emotional experiences to oral health without reference to aesthetic qualities. For example, we removed “I was happy with my teeth,” but retained “I was happy with how my teeth looked.” The final measures comply with pediatric dentists’ request for tools that specify the underlying cause of children’s oral health experiences and thus, provide clinically actionable information. The final item banks include 29 items that measure the full continuum of oral health aesthetics levels with a high degree of precision. Short form items were selected to maximize measurement precision across a wide range of oral health aesthetic states, from very poor to excellent. The recommended short forms could be used to identify children with oral health aesthetic concerns and to assess change in these concerns over time (e.g., in response to dental/orthodontic interventions). Alternative short forms composed of a different set of items from the calibrated item banks may be preferred for certain children (e.g., with a clinical condition) or applications (e.g., to screen for poor oral health aesthetics). A major advantage of IRT-calibrated measurement tools is that alternative forms that include different items from the same calibrated item bank can be scored on the same metric and compared.
Convergence with existing measures of OHRQoL, global health, and body image provide evidence of the tools’ construct validity. As expected, children with insufficient access to dental care had poorer oral health aesthetics by child- and parent/proxy-report. Child age was inversely associated with child-reported aesthetics, but unrelated to parent-reported aesthetics. This difference may reflect youths’ tendency to become increasingly concerned about their physical appearance during adolescence. It highlights the importance of asking children (especially teenagers) directly about their oral health aesthetics, as opposed to relying on parent/proxy-report.
The validity of oral health aesthetics measures should be further evaluated based on comparisons to objectively-defined and clinician-assessed aesthetic measures (e.g., anterior-posterior and vertical discrepancies). Longitudinal data, especially repeated administrations of the measures to children who undergo procedures to improve oral appearance are needed to evaluate the tools’ predictive validity and responsiveness, and to establish score cut-points that indicate meaningful within-person change in oral aesthetics. We calibrated the item banks using data collected from children and parents who were purposively sampled to maximize sociodemographic representation of the general US population. Therefore, scores can be interpreted relative to the general population of children in the US. The tools’ should be further evaluated for use in special populations such as children with craniofacial conditions or those who sustain dental trauma.
Supplementary Material
Acknowledgments
This project was supported by grants from the Agency for Healthcare Research and Quality (U18HS020508) and the National Institutes of Health (U01AR057956-02).
References
- 1.Shaw W. The influence of children’s dentofacial appearance on their social attractiveness as judged by peers and lay adults. Am J Orthod. 1981;79:399–415. [DOI] [PubMed] [Google Scholar]
- 2.Shaw W, Rees G, Sawe M, Charles C. The influence of dentofacial appearance on the social attractiveness of young adults. Am J Orthod. 1985;87:21–26. [DOI] [PubMed] [Google Scholar]
- 3.Phillips C, Beal KNE. Self-Concept and the Perception of Facial Appearance in Children and Adolescents Seeking Orthodontic Treatment. The Angle Orthodontist. 2009;79:12–16. doi: 10.2319/071307-328.1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Baiju R, Peter E, Varghese N, Sivaram R. Oral Health and Quality of Life: Current Concepts. Journal of Clinical and Diagnostic Research. 2017;11:ZE21–ZE26. doi: 10.7860/JCDR/2017/25866.10110 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Jokovic A, Locker D, Tompson B, Guyatt G. Questionnaire for Measuring Oral Health-related Quality of Life in Eight- to Ten-year-old Children. Pediatric Dentistry. 2004;26:512–518. [PubMed] [Google Scholar]
- 6.Jokovic A, Locker D, Stephens M, Kenny D, Tompson B, Guyatt G. Validity and Reliability of a Questionnaire for Measuring Child Oral-health-related Quality of Life. Journal of Dental Research. 2002;81:459–463. doi: 10.1177/154405910208100705 [DOI] [PubMed] [Google Scholar]
- 7.Gherunpong S, Tsakos G, Sheiham A. Developing and evaluating an oral health-related quality of life index for children; The CHILD-OIDP. Community Dental Health. 2004;21:161–169. [PubMed] [Google Scholar]
- 8.Yusuf H, Gherunpong S, Sheiham A, Tsakos G. Validation of an English version of the Child-OIDP index, an oral health related quality of life measure for children. Health Qual Life Outcomes. 2006;4:38–45. doi: 10.1186/1477-7525-4-38 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Broder HL, Wilson-Genderson M, Sischo L. Reliability and validity testing for the Child Oral Health Impact Profile-Reduced (COHIP-SF 19). Journal of Public Health Dentistry. 2012;72:302–312. doi: 10.1111/j.1752-7325.2012.00338.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Broder HL, McGrath C, Cisneros GJ. Questionnaire development: face validity and item impact testing of the Child Oral Health Impact Profile. Community Dentistry and Oral Epidemiology. 2007;35:8–19. doi: 10.1111/j.1600-0528.2007.00401.x [DOI] [PubMed] [Google Scholar]
- 11.Bendo CB, Paiva SM, Viegas CM, Vale MP, Varni JW. The PedsQLTM Oral Health Scale: feasibility, reliability and validity of the Brazilian Portuguese version. Health and Quality of Life Outcomes. 2012;10:42. doi: 10.1186/1477-7525-10-42 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Gilchrist F, Rodd H, Deery C, Marshman Z. Assessment of the quality of measures of child oral health-related quality of life. BMC Oral Health. 2014;14:40. doi: 10.1186/1472-6831-14-40 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Mokkink LB, Prinsen CAC, Bouter LM, Vet HCW,de Terwee CB. The COnsensus-based Standards for the selection of health Measurement INstruments (COSMIN) and how to select an outcome measurement instrument. Brazilian Journal of Physical Therapy. 2016;20:105–113. doi: 10.1590/bjpt-rbf.2014.0143 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Klem M, Saghafi E, Abromitis R, Stover A, Dew MA, Pilkonis P. Building PROMIS item banks: librarians as co-investigators. 2009;18. doi: 10.1007/s11136-009-9498-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.American Dental Association. Dental Quality Alliance. http://www.ada.org/en/science-research/dental-quality-alliance/. Accessed October 5, 2016.
- 16.Lasch KE, Marquis P, Vigneux M, et al. PRO development: rigorous qualitative research as the crucial foundation. Quality of life research : an international journal of quality of life aspects of treatment, care and rehabilitation. 2010;19:1087–1096. doi: 10.1007/s11136-010-9677-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Fortune-Greeley AK, Flynn KE, Jeffery DD, et al. Using cognitive interviews to evaluate items for measuring sexual functioning across cancer populations: improvements and remaining challenges. Quality of Life Research. 2009;18:1085–1093. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Willis GB. Cognitive Interviewing: A Tool for Improving Questionnaire Design. Thousand Oaks, CA: Sage; 2005. [Google Scholar]
- 19.GfK. Knowlegde Panel®: A Methodological Overview. https://www.gfk.com/fileadmin/user_upload/dyna_content/US/documents/KnowledgePanel_-_A_Methodological_Overview.pdf. Accessed March 20, 2020.
- 20.Broder H, Wilson-Genderson M. Reliability and convergent and disciminant validity of the Child Oral Health Impact Profile (COHIP Child’s version). Community Dentistry and Oral Epidemiology. 2007;35:20–31. [DOI] [PubMed] [Google Scholar]
- 21.Bevans KB, Riley AW, Forrest CB. Development of the healthy pathways parent-report scales. Quality of Life Research. 2012;21:1755–1770. doi: 10.1007/s11136-012-0111-0 [DOI] [PubMed] [Google Scholar]
- 22.Bevans KB, Riley AW, Forrest CB. Development of the healthy pathways child-report scales. Quality of life research. 2010;19:1195–1214. doi: 10.1007/s11136-010-9687-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Forrest CB, Bevans KB, Pratiwadi R, et al. Development of the PROMIS ® pediatric global health (PGH-7) measure. Quality of Life Research. 2014;23:1221–1231. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Bethell C, Read D, Stein R, Blumberg S, Wells N, Newacheck P. Identifying children with special health care needs: Development and evaluation of a short screening instrument. Ambulatory Pediatrics. 2002;2:38–48. [DOI] [PubMed] [Google Scholar]
- 25.Agency for Healthcare Research and Quality. Medical Expenditure Panel Survey. https://meps.ahrq.gov/mepsweb/. Accessed October 5, 2016. [PubMed]
- 26.U.S. Department of Health and Human Services. 2015. poverty guidelines. https://aspe.hhs.gov/2015-poverty-guidelines. Accessed October 5, 2016.
- 27.Lohr S. Sampling: Design and Analysis. Toronto, ON: Nelson Education; 2009. [Google Scholar]
- 28.Vautier S, Callahan S, Moncany D, Sztulman H. A bistable view of single constructs measured using balanced questionnaires: Application to trait anxiety. Structural Equation Modeling. 2004;11:261–271. [Google Scholar]
- 29.Muthén L, Muthén B. Mplus User’s Guide. Seventh Edition. Los Angeles, CA: Muthén & Muthén; 1998. [Google Scholar]
- 30.Hu L, Bentler PM. Cuttoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling. 1999;6:1–55. [Google Scholar]
- 31.Bollen K. Structural Equations with Latent Variables. New York, NY: Wiley; 1980. [Google Scholar]
- 32.Stucky BD, Thissen D, Orlando Edelen M. Using Logistic Approximations of Marginal Trace Lines to Develop Short Assessments. Applied Psychological Measurement. 2013;37:41–57. doi: 10.1177/0146621612462759 [DOI] [Google Scholar]
- 33.Green BF, Bock D, Humphres RL, Linn MD. Technical Guidelines for Assessing Computerized Adaptive Tests. Journal of Educational Measurement. 1984;21:347–360. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.