Ethics

This study fulfilled the criteria for a service evaluation, and approval to commence was provided by the Department of Innovation and Improvement, Cardiff and Vale University Health Board (Ref A/01). All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1975, as revised in 2008.

Subjects

The sample consisted of consecutive new patients aged 16 years and older referred to the University Dental Hospital, Cardiff, UK, during a 15-month period from 2012 through 2014. These patients were judged by senior clinicians to have moderate-to-severe malocclusions, as categorized using Angle’s Classification, for the skeletal pattern and required orthognathic surgery to correct the interarch relationships. Patients included as part of this study were UK residents and as such were offered orthognathic surgery within a public health service in which treatment is provided free at the point of use. Patients were excluded if they had been referred for or had reported temporomandibular disorder problems, had a history of sleep apnea, or reported or exhibited psychological disorders likely to confound QoL. Patients with syndromes or previous treatment for trauma were also excluded. As this was a pilot study, a formal hypothesis was not tested, and, therefore, no sample size calculation was performed. Instead, the sample was recruited and variables analyzed periodically until data saturation was reached (ie, further recruitment of participants would have no significant influence on the data).

Variables

Demographic explanatory variables, including patient age, gender, and malocclusion type, were recorded. The Index of Orthognathic Functional Treatment Need (IOFTN) was used to assess objective treatment need, with the single most severe occlusal or facial trait scored categorically from 1 to 5. Complexity of treatment was measured using the Index of Complexity, Outcome and Need (ICON), with a threshold for treatment of 43 points.

Pretreatment lateral cephalometric radiographs were traced and analyzed using Dolphin 10 Imaging & Management Solutions software. The severity of the dentofacial discrepancy was assessed via seven key cephalometric measurements: (1) ANB angle, (2) SNB angle, (3) lower facial height proportion, (4) upper incisor proclination, (5) Holdaway angle, (6) overbite, and (7) overjet. An overall radiographic severity value for each patient ranging from 0 to 7 was calculated by recording the number of traits that lay outside a range of two standard deviations from Eastman normal values.¹⁴

Patient-reported QoL was evaluated using the OQLQ. This condition-specific, self-completion questionnaire has been previously validated and contains 22 questions, rated on a 4-point Likert scale. Questions are spread across four domains that include social impediment, facial esthetics, oral function, and awareness of dentofacial deformity. Scores range from 0 to 88, with higher scores indicating poorer QoL and a greater patient-perceived need for treatment, although no threshold values are indicated.¹⁵ Patients were given the OQLQ to complete prior to their consultation at the multidisciplinary clinic.

Reliability of Measurements

One operator calibrated in the use of IOFTN and ICON scored all occlusal indices and performed all cephalometric measurements. The intraoperator reliability of the IOFTN and ICON scores was assessed by rescoring 20% of the sample, randomly selected using an integer sequence (www.random.org) 6 months later. A second calibrated operator also scored the variables from the 20% random sample to account for interoperator reliability.

Statistical Methods

All statistical analyses were carried out using IBM SPSS Statistics (v. 20) software (SPSS Inc, Chicago, Ill). Descriptive statistics were produced for the demographic variables. Reliability was tested for the categorical IOFTN scores using cross-tabulation and Kappa statistics¹⁶ and the Bland-Altman method for the continuous ICON scores.¹⁷

All occlusal and cephalometric values were designated as independent variables and the OQLQ scores as the dependent variable. Bivariate analyses were then carried out for the independent variables and the overall OQLQ score. Dichotomous independent variables were tested using an independent sample t-test, categorical variables were tested with a one-way analysis of variance (ANOVA), and continuous variables tested using linear regression. Prior to conducting the analyses, all continuous variables were checked for a normal distribution. The cephalometric variables were found to violate the assumption of normality as a result of the extremes of positive and negative readings. These values were all transformed to a positive magnitude away from accepted White norms.¹⁴ The analyses were extended to include the four individual domains of the OQLQ as additional dependent variables. Significant variables were extracted and used within a multivariate linear regression model to investigate their predictive effect on the overall OQLQ score. The significance level was set at P < .05 for all statistical analyses.

Participants

A total of 102 patients were recruited to the study, 97% of whom were white British. The remaining 3% of patients were Asian British. The patient completion rate for the OQLQ was 100%. The sample had a median age of 21 years, and participants ranged from 16 to 59 years of age. Only seven patients in the cohort were above 40 years of age. The majority of patients were female (74%). A mix of malocclusions was observed, with Class III being most prevalent (51%), followed by Class II (38%) and, finally, Class I (11%). Exploration of the demographic variables confirmed bias in the sample (ie, age, gender, and malocclusion), and therefore subsequent statistical analyses were carried out using robust bootstrapping methods incorporating a bias-corrected and accelerated confidence interval.¹⁸

Reliability of Measurements

The reliability of the IOFTN was very good, with Kappa scores of 0.91 for intraoperator testing and of 0.82 for interoperator testing. The Bland-Altman plots showed a mean difference in intraoperator scores for the ICON of 1.2 points, with 95% limits of agreement between −8.6 and 11.0. Mean difference for interoperator scores was −0.7, with 95% limits of agreement at −11.6 to 10.2. This represents a maximum random error of 11.6 points, which is below the level of clinical significance for the ICON, set at ±18 points.¹⁹

Descriptive Data

OQLQ scores ranged from 8 to 88, with a mean of 53.5 (standard deviation [SD]  =  18.7, 95% confidence interval [CI]  =  49.9–57.2). Individual domain scores showed that esthetics scored the highest, with a mean of 15.6 (SD  =  4.5). The remaining three domains were more closely grouped, although the functional domain had the lowest mean score at 10.3 (SD  =  5.8).

Treatment Need and Complexity Analyses

All patients were eligible for orthognathic treatment under the IOFTN guidelines, scoring 3 and over. All patients scored above the 43-point ICON treatment-need threshold. ICON scores ranged from 58 to 111, with a mean of 83.1 (SD  =  12.1, 95% CI  =  80.7–85.5). A significant proportion of patients (72%) were in the highest treatment complexity group (>77 ICON points).

Cephalometric analysis gave a mean severity score for the entire cohort of 3.6 (SD  =  1.3) traits out of 7. Class III patients scored a higher mean and median severity when compared to Class I and Class II patients, but this was not considered statistically significant with a one-way ANOVA (P  =  .010).

Independent Variable Significance Testing

The P-values obtained from the bivariate significance tests are shown in Table 1. Values that were significant for an association with the overall OQLQ score were gender and starting overjet. Females reported higher scores than males (12.6 points; 95% CI  =  2.0–17.2) and showed a significant association in the social (P  =  .016) and awareness (P  =  .013) domains of the OQLQ. The magnitude of overjet away from white normal values was associated with poorer overall QoL, with significant relationships in the esthetic (P  =  .017) and functional domains (P  =  .002).

Table 1.  Individual Variable Significance Testing, Expressed as P-Values. Bold Values Represent Significant Findings (P < .05)^a

View Fullscreen

When considering the functional domain as a separate dependent variable, significant associations were seen with overjet (P  =  .002) and overbite (P  =  .028) and the IOFTN score (P  =  .006). ICON and radiographic severity score were not associated with any aspect of patient-reported QoL.

When using the significant variables within a multivariate linear regression model (Table 2), the IOFTN category becomes a significant predictor for the overall OQLQ score (P  =  .013), with category 5 patients scoring a mean of 10.0 points (95% CI  =  2.1–17.8) more than category 4 (Table 2). Being female increases the OQLQ by 15.6 points when compared to males (95% CI  =  7.4–23.9). While overjet was shown to be significant predictor of OQLQ score (P  =  .049), the effect was small, with a 1-mm increase in overjet away from normal values, resulting in a 1.5 increase in OQLQ score (95% CI  =  0.1 to 3.0).

Table 2.  Results of Multivariate Regression Model for Factors Predicting Overall Orthognathic Quality of Life Questionnaire (OQLQ) Score

View Fullscreen