Introduction
The impact of orthodontic treatment on oral health–related quality of life (OHRQoL) in adolescents being treated in orthodontic practices has not yet been explored longitudinally. The aim of this cohort study was to describe the changes in both malocclusion and OHRQoL with orthodontic treatment.
Methods
One hundred seventy-four patients (ages, 10-17 years; 64.4% girls; 81.6% European) underwent 2-arch, fixed-appliance treatment in a 4-year prospective study conducted across 19 specialist orthodontic practices throughout New Zealand. They were assessed before treatment, at debond (when 87.4% of the baseline sample were reassessed), and at a mean 21 months postdebond (when 59.4% of the baseline sample were reassessed). OHRQoL was measured using the Child Perceptions Questionnaire, and the Dental Aesthetic Index was used to measure occlusion.
Results
Among the 104 patients who took part in all 3 assessments, little change in OHRQoL overall was seen at the end of treatment, despite considerable improvement in malocclusion (with the mean Dental Aesthetic Index score falling from 35.9 at baseline to 21.3 at debond). The mean Child Perceptions Questionnaire 11-14 was slightly greater at debond, and this was most notable in the functional limitations subscale. By the end of the study (21 months postdebond, on average), the decreases in Child Perceptions Questionnaire 11-14 scores were all substantial, especially in the emotional well-being and social well-being subscales.
Conclusions
Malocclusion affects orthodontic patients’ OHRQoL before treatment. A temporary increase in symptomatic impacts seen by the debond stage appears to ameliorate with time, with the benefits of orthodontic treatment for OHRQoL manifesting themselves some months later.
Highlights
- •
Treatment-associated effects on malocclusion were found by the end of treatment.
- •
Improvement in OHRQoL was not observed until 2 years later.
The aims of orthodontic treatment are to improve appearance, to correct the occlusal function of the teeth, and to eliminate occlusion that could damage long-term oral health. In reality, much (if not most) is undertaken for psychosocial reasons rather than to improve oral health. This means that determining the impact of orthodontic treatment on patients’ oral health–related quality of life (OHRQoL) is important. In the nearly 4 decades since the use of sociodental indicators was first mooted, considerable development of appropriate measures has occurred. The most frequently used child OHRQoL instrument has been the 37-item Child Perceptions Questionnaire (CPQ 11-14 ), in which a higher score indicates poorer OHRQoL.
Cross-sectional studies have confirmed an association between malocclusion and poorer OHRQoL, although not all have observed it to be strong. Accordingly, orthodontic treatment would be expected to lead to an improvement in OHRQoL. Documentation of such improvements using longitudinal studies is surprisingly scarce, given the central importance of the psychosocial effects of orthodontics.
Some authors have examined OHRQoL changes in the first few months of treatment only, during treatment itself, or for changes in individual malocclusion traits that may be treated more quickly, such as a midline diastema. The only long-term investigation of orthodontic treatment-associated OHRQoL changes had methodologic deficiencies: this involved a 20-year follow-up in the United Kingdom of 12- to 13-year-olds who had been first assessed in 1981. The authors found few psychological benefits of orthodontic treatment, but the study’s poor follow-up rate and use of a nonvalidated OHRQoL instrument meant that the findings have questionable utility. Three shorter-term longitudinal studies have used more comprehensive pretreatment and posttreatment assessments. All 3 observed short-term improvements in OHRQoL associated with orthodontic treatment, but whether those improvements were sustained in the medium term remained unanswered. Moreover, all 3 studies were conducted with patients in dental school orthodontic clinics, as was a Brazilian study in which significantly better OHRQoL was found in patients in retention than before treatment. An adult study made similar observations. In no adolescent study were the medium-term treatment-associated changes in OHRQoL determined, and none documented changes among patients being treated in practices outside dental schools. There is a need for “real-world” observations of treatment-associated changes in malocclusion and OHRQoL.
The aim of this study was to describe the changes in malocclusion and OHRQoL associated with orthodontic treatment in a 4-year prospective study conducted in 19 specialist orthodontic practices in New Zealand.
Material and methods
Ethical approval for the study was granted by the Multi-regional Ethics Committee (MEC/0611/143) in Wellington, New Zealand. Informed consent was obtained in writing from the participants and their parents.
Longitudinal data were obtained from baseline and follow-up assessments in a 4-year prospective study conducted in 19 private specialist orthodontic practices throughout New Zealand. Participants were assessed at pretreatment, immediately posttreatment, and at an end-of-study point a mean of 21 months after treatment.
All 75 orthodontists in New Zealand were invited to participate. Patients were randomly selected from those awaiting treatment with the 19 orthodontists (from all over the country) who accepted the invitation. They had a mean age of 49 years, and 26% were women; among the specialist orthodontists in practice in New Zealand at the time, the mean age was 52 years, and 20% were women. Inclusion in the study did not alter the provided orthodontic treatment in any way. Included in the study were patients between the ages of 10 and 17 years who were to undergo full maxillary and mandibular fixed orthodontic treatment (not involving headgear or other extraoral appliances) and who consented to take part. Specifically excluded were patients in 2-phase orthodontic treatment (with a separate identifiable skeletal change phase before the placement of appliances), those about to leave the area or practice, surgical patients, and those with major craniofacial abnormalities (eg, cleft lip and palate). We used only records routinely taken as part of a best-practice protocol in orthodontic treatment, and the STROBE guidelines were adhered to. No control group was used because withholding treatment is unethical.
Sociodemographic information was obtained from participants’ parents or guardians, and included age, sex, and self-identified ethnic group. Socioeconomic status (SES) data were collected using an occupationally based measure, the New Zealand Socio-Economic Index of Occupational Status, which allocates a score from 10 to 90, with 10 representing the lowest and 90 the highest SES. Since occupational information was obtained for both parents, the household SES was determined to be the higher of the 2 ratings. Occupations with scores of 10 to 39 were categorized as “low SES,” and scores of 40 to 59 were “medium SES”; those scoring higher were “high SES”.
OHRQoL was measured using the CPQ 11-14 , with the treating orthodontist giving each participant the questionnaire to complete at the office appointment. The scale collects information relating to the previous 3 months. Response options are “never” (scoring 0), “once or twice” (1), “sometimes” (2), “often” (3), and “every day or nearly every day” (4). The overall score is the sum of the individual item scores, and 4 subscale scores can be computed for the domains of oral symptoms, functional limitations, emotional well-being, and social well-being. A high score indicates poorer OHRQoL. The measure has been shown to be valid and responsive, with the minimally important difference for the overall CPQ 11-14 score determined to be 4 scale points. As a concurrent validity check, participants responded to the question “how much does the condition of your teeth, lips, jaws or mouth affect your life overall?” (scored on a 5-point ordinal scale ranging from “excellent” to “poor”), with validity deemed to be acceptable if there was an ascending gradient in mean CPQ 11-14 scores across those response categories.
Clinical measurements were made from duplicate records sent by each treating orthodontist. Malocclusion was assessed from models and photographs using the Dental Aesthetic Index (DAI), which assigned a malocclusion score using 10 clinician-measured occlusal components that were weighted, summed, and added to a constant (13) to give 1 DAI score. Higher scores represented poorer dental esthetics. Participants were then allocated to orthodontic treatment need categories, with scores of 25 or lower indicating “minor or no” treatment need, 26 to 31 indicating a definite malocclusion, 32 to 35 a severe malocclusion, and 36 or more a handicapping malocclusion. The clinical assessor (D.L.H.) underwent training and calibration in a pilot study, with repeated assessments of 20 sets of casts and photographs on 2 occasions 8 weeks apart. The intraclass correlation coefficient for the DAI score was 0.98, indicating very strong agreement.
Statistical analysis
Statistical analysis was undertaken using SPSS software for Macintosh (version 20; IBM, Armonk, NY). The alpha level for all tests was set at P <0.05. Internal consistency for the CPQ 11-14 scale scores was assessed using the Cronbach alpha.
Observed differences in means or proportions at 1 time were tested for statistical significance using (respectively) analysis of variance (ANOVA) and chi-square tests; changes in means over time were examined using paired t tests. Effect sizes were determined by dividing the mean change score by the standard deviation of the pretreatment scores to give a dimensionless measure of effect. Effect size statistics of less than 0.2 indicated a small clinically meaningful magnitude of change; 0.2 to 0.7, a moderate change, and greater than 0.7, a large change.
Results
Almost two-thirds of the 174 participants were girls ( Table I ), and fewer than 1 in 5 were non-European (comprising 15 Maori, 4 Chinese, 3 Indians, and 10 from various smaller groups). The mean age was 13.5 years (SD, 1.3; range, 10-17 years). The mean DAI score in the low-SES group was significantly higher than in the other 2 groups. Girls had a lower mean DAI score than did boys, and Europeans had a higher one than non-Europeans ( P >0.05 for both). About 1 in 10 subjects were categorized as having a “minor” or “no” need for treatment; almost half were in the most severe category.
Number (col %) | Mean DAI (SD) | DAI treatment need category ∗ | ||||
---|---|---|---|---|---|---|
Minor/none | Definite | Severe | Very severe/handicapping | |||
All combined | 174 (100.0) | 35.8 (8.4) | 17 (9.8) | 44 (25.3) | 34 (19.5) | 79 (45.4) |
Sex | ||||||
Female | 112 (64.4) | 35.2 (8.3) | 15 (13.4) | 27 (24.1) | 21 (18.8) | 49 (43.8) |
Male | 62 (35.6) | 36.8 (8.5) | 2 (3.2) | 17 (27.4) | 13 (21.0) | 30 (48.4) |
Age group (y) | ||||||
10-12 | 63 (36.2) | 35.6 (8.4) | 6 (9.5) | 16 (25.4) | 14 (22.2) | 27 (42.9) |
13 | 50 (28.7) | 36.1 (8.2) | 5 (10.0) | 13 (26.0) | 7 (14.0) | 25 (50.0) |
14+ | 61 (35.1) | 35.8 (8.7) | 6 (9.8) | 15 (24.6) | 13 (21.3) | 27 (44.3) |
Ethnicity | ||||||
European | 142 (81.6) | 36.0 (8.2) | 12 (8.5) | 36 (25.4) | 27 (19.0) | 67 (47.2) |
Non-European | 32 (18.4) | 34.7 (9.5) | 5 (15.6) | 8 (25.0) | 7 (21.9) | 12 (37.5) |
Household SES | ||||||
High | 65 (37.4) | 34.4 (7.4) † | 4 (6.2) | 22 (33.8) | 13 (20.0) | 26 (40.0) |
Medium | 81 (46.6) | 35.5 (8.7) | 11 (13.6) | 18 (22.2) | 18 (22.2) | 34 (42.0) |
Low | 28 (16.1) | 39.9 (8.8) | 2 (7.1) | 4 (14.3) | 3 (10.7) | 19 (67.9) |
∗ <25, Minor or none; 26-30, definite; 31-35; severe; >35, very severe or handicapping (after Estioko et al ).
† P <0.05; 1-way ANOVA: the low-SES group differed from the other 2 groups.
Baseline CPQ 11-14 scores ranged from 0 to 89 and were positively skewed. Mean CPQ 11-14 scores across the ordinal response categories of the global self-rated oral health question were 11.3 (excellent), 17.1 (very good), 23.6 (good), 27.1 (fair), and 43.0 (poor; P < 0.05). Mean overall CPQ 11-14 scores were highest among those of low household SES, and there was a marked SES gradient in scores on the emotional well-being subscale ( Table II ).
CPQ 11-14 overall score (SD) | CPQ 11-14 subscale scores (SD) | ||||
---|---|---|---|---|---|
Oral symptoms | Functional limitations | Emotional well-being | Social well-being | ||
All combined | 20.5 (14.2) | 5.3 (2.8) | 5.5 (4.3) | 5.6 (6.5) | 4.2 (4.7) |
Sex | |||||
Female | 21.0 (14.2) | 5.2 (2.8) | 5.3 (4.1) | 6.3 (6.7) | 4.2 (4.7) |
Male | 19.8 (14.3) | 5.4 (2.8) | 5.8 (4.7) | 4.5 (5.9) | 4.1 (4.7) |
Age group (y) | |||||
10-12 | 19.2 (14.6) | 5.2 (2.8) | 5.5 (4.7) | 4.5 (6.0) | 4.0 (4.7) |
13 | 21.4 (14.8) | 5.5 (2.7) | 5.3 (4.1) | 6.1 (6.8) | 4.5 (5.0) |
14+ | 21.2 (13.3) | 5.2 (2.9) | 5.6 (4.1) | 6.4 (6.6) | 4.1 (4.4) |
Ethnicity | |||||
European | 20.8 (14.8) | 5.4 (2.8) | 5.5 (4.5) | 5.6 (6.7) | 4.3 (4.9) |
Non-European | 19.4 (11.2) | 4.9 (2.9) | 5.1 (3.7) | 5.9 (5.7) | 3.6 (3.2) |
Household SES | |||||
High | 20.5 (14.1) ∗ | 5.1 (3.0) | 5.4 (4.2) | 5.7 (5.6) † | 4.3 (4.8) ‡ |
Medium | 18.6 (12.3) | 5.3 (2.6) | 5.4 (4.3) | 4.5 (6.2) | 3.4 (3.6) |
Low | 26.4 (17.9) | 5.6 (2.7) | 5.8 (4.7) | 8.8 (8.1) | 6.2 (6.4) |
∗ P <0.05, 1-way ANOVA: the medium category differs from the low category.
† P <0.05, 1-way ANOVA: all categories differ from one another.
‡ P <0.05, 1-way ANOVA: the low category differs from the medium and high categories, which do not differ significantly from each other.
The follow-up rate by the end of treatment was 87.4% (n = 152), with 59.8% (n = 104) assessed subsequently at the end of the study. The mean follow-up periods for the treatment phase were 25.8 months (SD, 7.6 months) and 21.3 months (SD, 9.7 months) for the posttreatment phase. The research protocol had called for a posttreatment follow-up at 12 months; however, a large earthquake destroyed a considerable section of the city where the principal investigator was based, and this delayed the collection of final records by several months, thus leading to the 21-month period. We considered that further extending it to 24 months would adversely affect the follow-up rate. An attrition analysis ( Table III ) showed that a significantly greater proportion of non-European participants had been lost by the end of the study, and that there were higher impacts in the emotional well-being subscale at baseline among those lost to follow-up. All subsequent analyses were limited to the 104 who remained in the study until it finished.
Characteristic | Baseline sample | Immediately posttreatment ∗ | Later posttreatment period † | ||
---|---|---|---|---|---|
Followed up | Lost to follow-up | Followed up | Lost to follow-up | ||
Numbers (row %) | 174 (100.0) | 152 (87.4) | 22 (12.6) | 104 (59.8) | 70 (40.2) |
Sex | |||||
Female | 112 (64.4) | 96 (63.2) | 16 (72.7) | 72 (69.2) | 40 (57.1) |
Male | 62 (35.6) | 56 (36.8) | 6 (27.3) | 32 (30.8) | 30 (42.9) |
Mean age (SD) | 13.5 (1.3) | 13.4 (1.3) | 13.5 (1.4) | 13.4 (1.3) | 13.5 (1.3) |
Ethnicity | |||||
European | 142 (81.6) | 125 (82.2) | 17 (77.3) | 90 (86.5) | 52 (74.3) ‡ |
Non-European | 32 (18.4) | 27 (17.8) | 5 (22.7) | 14 (13.5) | 18 (25.7) |
Household SES | |||||
High | 65 (37.4) | 54 (35.5) | 11 (50.0) | 41 (39.4) | 24 (34.3) |
Medium | 81 (46.6) | 73 (48.0) | 8 (36.4) | 45 (43.3) | 36 (51.4) |
Low | 28 (16.1) | 25 (16.4) | 3 (13.6) | 18 (17.3) | 10 (14.3) |
Mean DAI (SD) | 35.8 (8.4) | 36.0 (8.3) | 34.6 (9.1) | 35.9 (7.9) | 35.6 (9.1) |
Mean CPQ 11-14 (SD) | 20.5 (14.1) | 20.3 (13.7) | 22.1 (17.3) | 20.1 (14.2) | 21.1 (13.6) |
Oral symptoms | 5.3 (2.8) | 5.3 (2.8) | 4.9 (2.9) | 5.3 (2.6) | 5.3 (3.1) |
Functional limitations | 5.5 (4.3) | 5.7 (4.5) | 4.0 (2.6) | 5.4 (4.2) | 5.6 (4.5) |
Emotional well-being | 5.6 (6.5) | 5.2 (5.5) | 6.7 (10.8) ‡ | 5.3 (6.3) | 6.1 (6.8) |
Social well-being | 4.2 (4.7) | 4.1 (4.6) | 4.6 (5.3) | 4.2 (5.1) | 4.1 (5.1) |
∗ Mean follow-up time after the baseline assessment was 25.8 months (SD, 7.6).
† Mean follow-up time after the previous assessment was 21.3 months (SD, 9.7).
By the end of treatment, the mean DAI score had reduced from 35.9 (SD, 7.9) to 21.3 (SD, 3.6; P <0.001), representing an effect size of 1.8 (large). Some 84.8% of patients had a DAI score of 25 or less, and there were no scores higher than 30. This represented a significant reduction in DAI scores with treatment.
There were differences by baseline malocclusion category in the amount of change in DAI scores ( Table IV ). Those with the most severe malocclusions before treatment showed the greatest reductions in DAI scores. At debond, the 4 groups’ mean DAI scores did not differ significantly.
DAI category at baseline ∗ | Mean DAI at baseline (SD) | Mean DAI at debond (SD) | Change in DAI (SD) | Effect size † |
---|---|---|---|---|
Minor/none | 25.0 (0.1) ‡ | 22.2 (3.9) | 2.8 (3.9) ‡ | 0.4 (moderate) |
Definite | 29.0 (1.4) | 21.2 (3.7) | 7.8 (3.7) | 1.0 (large) |
Severe | 33.4 (1.1) | 20.9 (3.5) | 12.4 (3.6) | 1.6 (large) |
Handicapping | 43.2 (5.8) | 21.5 (3.7) | 21.6 (6.0) | 2.7 (large) |
∗ Defined (respectively) as DAI score ranges of 13-25, 26-31, 32-35, and 36+.
† Calculated using the overall baseline standard deviation (7.9) as the denominator, because those for the DAI categories at baseline were artificially lowered by the categorization process.