The purpose of this study was to assess the long-term posttreatment changes in all criteria of the American Board of Orthodontics’ (ABO) model grading system.
We used plaster models from patients’ final and posttreatment records. Thirty patients treated by 1 orthodontist using 1 bracket prescription were selected. An initial discrepancy index for each subject was performed to determine the complexity of each case. The final models were then graded using the ABO’s model grading system immediately at posttreatment and postretention. Statistical analysis was performed on the 8 criteria of the model grading system, including paired t tests and Pearson correlations. An alpha of 0.05 was considered statistically significant.
The average length of time between the posttreatment and postretention records was 12.7 ± 4.4 years. It was shown that alignment and rotations worsened by postretention ( P = 0.014), and a weak statistically significant correlation at posttreatment and postretention was found (0.44; P = 0.016). Both marginal ridges and occlusal contacts scored less well at posttreatment. These criteria showed a significant decrease in scores between posttreatment and postretention ( P <0.001), but the correlations were not statistically significant. The average total score showed a significant decrease between posttreatment and postretention ( P <0.001), partly because of the large decrease in the previous 2 criteria.
Higher scores for occlusal contacts and marginal ridges were found at the end of treatment; however, those scores and the overall scores for the 30 subjects improved in the postretention phase.
Orthodontics is concerned with occlusion and, more importantly, treating malocclusion. Several attempts were made to quantify the different types of occlusion. The American Board of Orthodontics (ABO) has tried to quantify the proper arrangement of teeth after orthodontic treatment. It has developed a model grading system for orthodontists to use in assessing the adequacy and quality of their orthodontic treatment more objectively. The ABO’s model grading system was developed in the early 1990s in 4 field tests over 5 years. It relies on 8 criteria and has been tested recently to determine the overall quality of the orthodontic results of ABO graded cases. The ABO has divided its grading criteria into 2 sections: the first section of the discrepancy index is a way of evaluating complexity before treatment, and the second section uses the ABO’s model grading system to grade the quality of the overall orthodontic result.
Relapse is an inevitable part of orthodontic treatment, but it is valuable to determine which portion of the orthodontic treatment might worsen or improve with time. The stability of a patient after orthodontic treatment has been difficult to predict. Over the years, there has been much research in this area, and several goals have been shown to be important during and after orthodontic treatment. The patient’s mandibular arch form should be maintained during orthodontic treatment, and every effort should be made to keep it as close to the original arch form as possible. Moreover, the mandibular intercanine width should be maintained as close to the original intercanine width, or it will tend to relapse after the orthodontic treatment. The original position of the mandibular incisors before orthodontic treatment is considered to be the most stable incisor position. Also, mandibular incisor interproximal reduction provides a long-term improvement in posttreatment stability.
In this study, we set out to assess the long-term posttreatment changes using all criteria of the ABO’s model grading system and to quantify its relationship to an overall excellent orthodontic result. We also addressed the areas that are of concern to orthodontists after treatment and which criteria are more likely to worsen and which will improve.
Material and methods
The sample consisted of 30 patients treated by 1 orthodontist (R. G. “Wick” Alexander, who treated them with the Alexander preadjusted appliance), selected from a larger sample of his records. These subjects were chosen based on completeness of the records required to undertake the study. The inclusion criteria for the study were that both arches had to be in the permanent dentition and treated orthodontically by 1 orthodontist using the Alexander bracket prescription. We included both extraction and nonextraction patients, but we excluded those who had mandibular incisor extractions because those extractions would have skewed all overjet measurements.
One examiner (R.M.A.) did all analyses on the models including the grading assessment. An ABO discrepancy index was performed on each subject to determine the complexity of the case. Plaster models for each subject were graded using the ABO’s model grading system immediately posttreatment and from 4 to 22 years postretention.
The forms used to evaluate the cases were obtained directly from the ABO Web site: Discrepancy Index Worksheet, Cast Radiograph Evaluation Reference, and the Cast Radiograph Evaluation Form. The ABO measuring gauge was used to measure 7 of the 8 criteria of the model grading assessment, whereas the final panoramic radiographs were used to measure root angulation. The model grading system was used to measure all subjects. This tool relies on 8 criteria to determine the adequacy of the treatment results: alignment and rotations, marginal ridges, buccolingual inclination, occlusal relationships, occlusal contacts, overjet, interproximal contacts, and root angulation. Because of the incomplete records for root angulation at postretention (15 panoramic x-rays were not taken), the statistical analysis was disregarded for this criterion. A definition of each criterion along with the point system was defined by the ABO and taken directly from the Cast Radiograph Evaluation Reference. The total score was calculated by adding the scores of each category. To ensure the reliability of the 1 examiner, 5 subjects were randomly selected and regraded. The total scores for these subjects were redone at both posttreatment and postretention.
All statistical analyses were performed using the Statistical Package for Social Sciences (IBM, Armonk, NY) for Windows. The overall scores at posttreatment and postretention were compared to determine whether they improved or worsened with respect to the 8 criteria of the ABO’s model grading system. Paired t tests were performed on each criterion to determine whether there was a statistically significant difference between posttreatment and postretention. A Pearson correlation was performed on each criterion to determine whether there was a correlation between the 2 scores at posttreatment and posttreatment. A time-adjusted score (delta [T3–T2]) was compared with the original average decrease in score, change (T3–T2) (postretention minus posttreatment), to determine whether there was a correlation. An assessment of reliability of the scoring was performed by selecting 5 models from posttreatment and 5 models from postretention and rescoring the ABO criteria. Reproducibility was assessed using the standard deviation of Dahlberg. An alpha of 0.05 was considered statistically significant.
Reproducibility was assessed using the standard deviation of Dahlberg, and the results are presented in Table I . Based on this calculation, it was expected that random differences between the scores at posttreatment and postretention, when no change had actually taken place, would be within 3 SD 90% of the time in general. Therefore, it can be estimated that any difference in total ABO scores greater than 7 units (3 × 2.202 = 6.606) would be judged more likely to be a real change than only a random error. Similar assessments were made for 6 of the 8 criteria. For interproximal contacts and root angulations, the data were not appropriate for this assessment.
Thirty subjects were graded at posttreatment and postretention. For the change over time in total ABO scores, an adjustment was made by dividing the change by the number of years to obtain a yearly change for each subject. The average score, results of the t tests, and correlations for the 8 criteria at posttreatment and postretention are summarized in Table II .
|Criterion||Evaluation||Mean||SD||Change||t test P value||Correlation||Correlation P value|
For the changes in alignment and rotation, the scores increased for 22 of the 30 patients (73.3%) and decreased for 6 subjects (20.0%). A statistically significant increase of 1.23 in alignment and rotations between posttreatment and postretention, with a weak statistically significant correlation (0.44; P = 0.016), was found. For the changes in marginal ridges, the results showed that the scores decreased for 27 patients (90%), with a highly significant decrease of 3.53 in the score for marginal ridges between posttreatment and postretention. However, the correlation was not statistically significant. The results showed that for 5 of the 30 patients (16.7%), the buccolingual scores increased; for 18 patients (60%), the scores decreased; for 7 patients (23.3%), the values remained the same with a statistically significant decrease of buccolingual inclination (average decrease, 1.13 points) ( P = 0.009). However, the correlation was not statistically significant. For the changes between posttreatment and postretention for overjet, the results showed a statistically significant decrease (average decrease, 1.10 points) ( P = 0.006). However, the correlation was not statistically significant ( P = 0.18).
For the changes in occlusal contacts, the scores decreased for 26 patients (86.7%), with a highly significant decrease in the score of occlusal contacts (average decrease, 3.90 points) ( P <0.001). However, the correlation was not statistically significant ( P = 0.16). For the changes in occlusal relationships, the paired t test did not show a statistically significant change between posttreatment and postretention ( P = 0.44) or a statistically significant correlation ( P = 0.96). For interproximal contacts, the scores remained the same for 27 patients (90%). The t test and correlation results were unreliable because 27 patients received a score of zero at both times; this calls into question the appropriateness of these tests. For root angulation, the scores remained the same for 25 patients (83.3%). The results for root angulation were not considered for statistical purposes because of the lack of panoramic x-rays at postretention in the records.
The average scores for a subject were 21.97 at posttreatment and 13.10 at postretention, whereas the average change between the total scores was 8.87 ( Table II ). The results showed that for 4 of the 30 patients (13.3%), the average total score increased; for 25 patients (83.3%), the scores decreased; and for 1 subject (3.3%), the values remained the same. There was a highly significant decrease between posttreatment and postretention ( P <0.001). However, the correlation between posttreatment and postretention was not statistically significant ( P = 0.85).
The raw change in average score of postretention minus posttreatment was called change (T3–T2), and the time-adjusted change of postretention minus posttreatment was called delta (T3–T2). The postretention records for all 30 patients were taken from 4 to 22 years after posttreatment, and the average time between these records was 12.7 ± 4.4 years. The delta (T3–T2) average change in score from postretention to posttreatment after it was adjusted for time was –0.71 ( Table III ), with a highly significant positive correlation between delta (T3–T2) and change (T3–T2) (0.89; P <0.001). The comparison of the time-adjusted change (delta [T3–T2]) with the discrepancy index showed a slight positive association between the 2 variables (0.404, P = 0.027).