INTRODUCTION

Evaluating the likely duration of a patient's orthodontic treatment remains an elusive component of clinical management. Although a finished orthodontic case could have an esthetic result and meet American Board of Orthodontics (ABO) standards, the patient might consider it unsatisfactory if treatment duration exceeded the initial prediction. Patients who receive accurate information about predicted treatment duration are found to be more satisfied with their treatment and have more realistic expectations of outcomes.1 Satisfied patients also are more likely to refer other patients.2 An accurate estimate of treatment length is not only valuable to the patient to determine time absent from work or school, but also allows the orthodontist to predict overhead costs and determine appropriate fees.3–5 It is for these reasons that it is critical for practicing clinicians to predict accurate treatment duration.

One possible method of prediction utilizes the discrepancy index (DI), a pretreatment scoring system developed by the ABO for Phase III of the orthodontic board certification exam. The DI has become an accepted and reliable index for quantifying the complexity of cases based on pretreatment orthodontic record analysis and measurements from dental casts and radiographs.6,7 However, the DI scoring method has been modified during the past few years in the “other” category.

Multiple studies have sought to determine which factors best predict treatment duration. Poor patient compliance, missed appointments, broken appliances, and poor oral hygiene can contribute to longer duration.1,2,4,5,8–12 Furthermore, excessive duration may lead to poor oral hygiene, patient compliance, and case outcomes.13

To date, no studies have quantified the relationship of DI score to treatment duration. The purpose of this study was to determine if there was a significant association between total DI score or DI components with treatment duration in a graduate orthodontic clinic.

MATERIALS AND METHODS

Institutional review board approval was obtained for this retrospective study of patient records from 1720 consecutively completed cases started and finished by orthodontic residents in a 2-year university clinic from 1998–2004. Cases were excluded if: (1) they were in the mixed dentition, (2) casts were broken, or (3) other records were incomplete. A total of 732 cases were analyzed.

Data collection was completed in two parts. In part one, DI scores for 716 cases were collected by the senior author and five undergraduate dental students. Records were randomized for each year and divided among the six examiners. Pretreatment cephalometric tracings (ANB, SN-GO-GN, and IMP angles), panoramic radiographs (axial inclinations) and casts (overjet, overbite, anterior open bite, lateral open bite, crowding, occlusion, lingual posterior crossbite, buccal posterior crossbite, and other) were used to collect the DI score. Treatment duration, age, sex, ethnicity, and year debonded (removal of fixed appliances) were also recorded. The DI scores calculated followed the formula outlined by Cangiolosi et al6 and established in the ABO November 2006 guidelines.

Calibration of the six examiners was achieved by scoring 20 cases at 1–2 week intervals until a ±5% margin of error was achieved. Cohen's kappa coefficients ranged from 0.68–0.94 depending on the variable scored. Data collection began once the kappa coefficient was greater than 0.61, indicating substantial agreement.14 In part two, the senior author reviewed the first data set and found an additional 16 cases that were not scored. These cases were scored the same as part one and increased the sample size to 732. Of the 732 cases, 189 had a score in the “other” category of the DI. Since the DI was revised in 2009 to specify components of the “other” category with definitive point values after the original scoring, these components and point values were recorded.14,15 The DI scores were updated to reflect the revised scoring. Pretreatment radiographs were used to evaluate supernumerary teeth, impacted teeth, missing teeth, skeletal asymmetries, and additional treatment complexities (ie, ectopic eruption). Patient photos and casts were used to evaluate midline discrepancy, anomalous morphology, spacing, ankylosis of permanent teeth, and tooth transposition.

The second part of data collection was completed by the senior author. Twenty randomly selected cases included in the 732 cases were rescored to assess reliability. During the 2 months of data collection, the same 20 cases were scored at the beginning, after 2 weeks, and at the end of data collection. Test-retest reliability was excellent (ICC > 0.90).

Pearson and Spearman correlation coefficients were calculated along with plots to evaluate the associations of the DI score and the individual DI components with treatment duration. A multiple variable regression analysis was performed for significant DI score (or individual components) to assess significance of additional possible predictors of treatment duration (age, sex, ethnicity, and the year treatment finished) and significance of individual DI components after adjusting for other DI components. With a sample size of 732, the study had >99% power to detect a difference in the correlation between total DI and treatment duration if the true correlation was 0.30, assuming a two-sided 5% significance level for the test.

RESULTS

The population had 59% women and 41% men; mean age was 19 ± 9.4 years. The majority were White (84%), followed by African American (8%), other (4%), and Asian and Hispanic (2% each). Between 80 and 123 cases (11%–17%) were used each year from 1998–2004 (Table 1). None of the demographic variables were significant in any of the models.

Table 1 Demographics (N  =  732)^a

View Fullscreen

The mean total DI score was 15.7; mean treatment duration was 2.6 years (Table 2). Since the magnitude and inference were similar between Pearson and Spearman coefficients for each DI variable, only Pearson correlation coefficients were reported in Table 2. Total DI score explained 9% of the variability in treatment duration, occlusion 5%, overjet 5%, crowding 3%, and none of the remaining individual components explained more than 2% of the variability in treatment duration. Spearman correlation coefficients are reported in Table 3 for the DI “other” subcategories because of the large proportion of zero values (ie, at least 94% for each subcategory). None of the individual “other” subcategories explained more than 1% of the variability in treatment duration.

Table 2 Descriptive Statistics and Pearson Correlation Coefficient vs Treatment Time for Treatment Time and DI Components (N  =  732)^a

View Fullscreen

Table 3 Frequency and Percentages for DI Other Subcategories and Spearman Correlation vs Treatment Time (N  =  732)^a

View Fullscreen

A multiple variable regression analysis including DI score failed to identify demographic variables as significant, so they were excluded from the model. There was a significant, but weak, correlation between the DI and treatment duration (P < .0001; r²  =  0.09). The average increase in treatment duration in years was 0.031 (approximately 11 days) for each point increase in total DI score. A scatter plot of treatment duration vs total DI showed the regression equation: treatment duration  =  0.031 × DI + 2.098 (Figure 1).

View Figure

• Download as Powerpoint
• Download Figure

A multiple variable regression analysis including DI components, DI “other” components, and demographic variables identified the components of occlusion, crowding, overjet, cephalometrics, overbite, lateral open bite, and tooth transposition as significant. The results of the multiple variable regression analyses and the individual regressions with each significant DI component are reported in Table 4. There was a significant association between the components of the DI and treatment duration (P < .0001; r²  =  0.14). After adjusting for the other significant DI components, the average increase in treatment duration per point increase was as follows: tooth transposition approximately 199 days, crowding 30 days, overbite 27 days, overjet 25 days, occlusion 21 days, lateral open bite 14 days, and cephalometrics 5 days.

Table 4 Regression Results for Analysis of Treatment Time with Significant DI Components Combined and DI Total Individually^a

View Fullscreen

DISCUSSION

The large population used in this study allowed for a precise estimation of the correlation between the DI score and treatment duration (95% CI: 0.23, 0.36). Although various factors contribute to the length of orthodontic treatment, the DI is one method that can be used for prediction. The DI is a reliable index for measuring complexity of malocclusion.16 In agreement with Vu et al.,17 the DI score is correlated with increased treatment duration. The current study showed that the average increase in treatment duration is about 11 days for each point increase in total DI score, so a 10-point increase in DI score will increase treatment duration by 110 days on average.

While the current study was the first to quantify the connection between DI and treatment duration, other studies found similar results. Buccal occlusion,9 specifically Class II division 1, increased treatment duration by approximately 150 days.18 As documented by the current data, a patient who is a full step Class II would score eight points for occlusion, and the treatment would increase by approximately 180 days, which agrees with other studies that found crowding,2 increased overbite, and excessive overjet extend treatment.19 The cephalometric variables (ANB, SN-MP, and IMPA) prolonged treatment by 5 days for each point in this category. Previous studies reported increases in the ANB angle lengthens treatment duration.8,20

While numerous studies reported that cases requiring extractions, particularly premolars, have a significant increase on treatment duration,2,8,10,11,18,19,21,22 the current study could not evaluate this influence because potential extractions are not specifically scored with the DI. Transposed teeth, although composing only a small portion of the population (<1%), posed the greatest increase in treatment duration (approximately 199 days). This study also found a 14-day increase for each point increase in lateral open bite, which had not been previously reported. This small amount of increased treatment duration may not adequately represent the difficulty clinically because lateral open bite is heavily weighted in the DI scoring.

Contrary to the findings of Stewart et al.23 that impacted maxillary canines lengthened treatment by approximately 72 days for the unilateral-impacted canine group, and approximately 297 days for the bilateral-impacted canine group, the current study was unable to identify a significant influence on treatment duration after adjusting for the other DI components. There was a significant bivariate correlation between unilateral canine impaction and treatment time (Table 3), but it was not significant in the multiple variable model (Table 4).

Even though the high variability in treatment duration is often related to compliance, practitioners can use the results of this retrospective study as an aid to predict treatment duration. Though the DI was constructed by the ABO as a measure of “case complexity,” the index also has value beyond its original intent by predicting time in treatment.

CONCLUSIONS

• This retrospective study of university clinical records comprising diverse clinicians showed that treatment duration increased 11 days for each point in the DI score. The components with the strongest influence were about:

• 119 days for each point increase in tooth transposition;

• 28–30 days for each point increase in crowding, overjet, and overbite;

• 21 days for each point increase in occlusion;

• 14 days for each point increase in lateral open bite; and

• 5 days for each point increase in cephalometrics.