|Home | About | Journals | Submit | Contact Us | Français|
The aim of this study was to investigate whether the addition of records can influence intra- and inter-rated agreement on torque choices made to treat a group of patients with various malocclusions.
Forty-eight patients were presented to five orthodontic specialists in three different occasions. During the first session, the participants were shown only the models and intraoral photos of the patients; extraoral photos were added during the second session, and cephalometric X-rays were further supplemented during the third session. Mean weighted kappa coefficients were calculated to measure agreement.
The inter-observer agreement was low with the mean coefficients measured:κ1 = 0.34 (SD ± 0.09), κ2 = 0.57 (SD ± 0.12), and κ3 = 0.54 (SD ± 0.28) for the three attempts, respectively. The mean kappa coefficients for the intra-rater agreement were also low ranging from 0.18 to 0.66 and the mean coefficients were 0.27 (SD ± 0.11) between first and second, and 0.53 (SD ± 0.11) between second and third attempt, respectively.
This study shows that the addition of extraoral photographs, and subsequently cephalograms to plaster models and intraoral photos, does affect intra-, and inter-rater agreement on torque selection. It seems that the addition of extraoral photos plays a more important role in torque selection decisions than lateral cephalograms. Different clinicians do not have a uniform opinion on the size of torque required to treat cases. Further research is required to define rules on torque choices.
Andrews (1) introduced the concept of proper crown inclination in his seminal ‘Six keys to normal occlusion’ article and he presented it as a keystone characteristic for normal overbite and solid posterior occlusion. The development of the straight-wire appliance where first-order offsets, mesiodistal root tip (second-order), and slot inclination (third-order) bends were incorporated into the brackets came as a consequence of the previous work, increasing the efficiency of treatment by decreasing the amount of wire bending needed for the typical patient (2). Since then, a great number of prescriptions have been proposed, used, and promoted by various clinicians (e.g. Roth, Ricketts, Alexander, etc.) based almost entirely on individual personal beliefs as to what constitutes the best prescription for optimum treatment results.
Ideally, any prescription employed should be capable to bring each tooth to the desired faciolingual inclination in terms of both function and aesthetics. Sarver and Ackerman suggest that all four dimensions of smile analysis—anterior, lateral, oblique, and time—must be taken into account when planning the orthodontic rebuilding of a smile and they drew attention to incisor position and angulation (3, 4). With regard to the latter, it has been shown that both the maxillary labiolingual inclination and AP position play an essential role in the aesthetics of the smiling profile (5). A maxillary incisor that is upright or almost perpendicular to the horizontal plane is more preferable, while labial inclination of the maxillary incisors could easily ruin a pleasing smile appearance (5, 6).
Variations in the morphology of the crown (7, 8), bonding deficiencies expressed as inaccuracies (9, 10), the torque play between wire and bracket (11, 12), as well as a number of unknown factors like the muscle activity all may play a role in the final tooth orientation. All of the previous combined with the plenitude of prescriptions available, and the different treatment mechanics used by clinicians on a global scale while pursuing excellence in both aesthetics and occlusion, make finishing with a straight wire extremely demanding.
Advances in technology in the form of digital model acquisition, three-dimensional imaging of the dentition and skeleton, robotics, and computer aided design/computer aided manufacturing allow new individualized finishing archwires or customized bracket-archwires systems to be fabricated. These systems based on virtual set-ups show promising results in terms of efficiency as it is expressed in shorter treatment times or better occlusal finishing when compared with cases treated in a conventional way (13–17). However, since these technologies are rather new, their application requires additional investment and training, and their efficiency has not been proven beyond doubt, their use in not yet widely spread.
Many bracket systems offer more than one prescription values for a single tooth. Thus, as an alternative to using customized systems, and in order to achieve desirable occlusal and aesthetic results—in an effort to remain loyal to the straight-wire principle—prominent clinicians may depart from the employment of a single constellation of bracket prescription in all of their cases, and use variable torque values on different teeth according to their treatment goals (12).
It has been shown that the clinical examination, when supplemented only with plaster models and photographs, offer sufficient information to draw an orthodontic treatment plan for most patients (18). Deveraux et al. (19) concluded that on most treatment-planning decisions in the six patients that they used in their study, the availability of a lateral cephalometric radiograph and its tracing did not make a significant difference to the treatment decisions, and this highlights the uncertainty surrounding the necessity for lateral cephalometric radiographs in treatment planning.
A similar obscurity exists when clinicians are faced with decisions regarding the torque values they may choose to treat efficiently cases covering a great span of different malocclusions, and using a straight-wire approach. There is a serious lack of evidence on the value that various orthodontic records may have on making such choices.
Thus, the aim of this study thus was to 1. investigate whether adding records can influence the agreement among different clinicians, and 2. the effect that the additional information might have on the same individual, regarding the torque values that they may choose to treat a group of patients with various dentofacial anomalies.
This study was based on a sample of 48 patients (mean age 20.40 years, SD ± 12.6, range: 11–61; 38 per cent males, 62 per cent females) seeking orthodontic treatment in five different private practices. Each practice contributed with the records of 10 consecutive patients who were referred for orthodontic evaluation representing a great spectrum of malocclusions (cleft and syndromic patients were not included) with the exception of one practice, which contributed only eight patients. All patients had full dentitions and patients with impacted teeth were excluded.
Photographs of the patients (facial frontal, smiling, and profile, intraoral frontal, and upper and lower occlusal) were available. The collected records also included plaster models and cephalometric X-rays.
Five orthodontic specialists (four females and one male—given numbers from 1 to 5) familiar with the principle of differential torques were asked to make torque choices for the right maxillary central incisor, that they would think as more appropriate for treating these patients. The available torque choices were −2 degrees (low torque—L), 15 degrees (standard torque—S), and 22 degrees (high torque—H), as offered by the Damon System, (Ormco, Glendora, California, USA). A value of 1, 2, and 3 was given for the corresponding bracket choice following the above bracket order.
The torque choices during the first attempt were made with the participants shown only the models and the intraoral photos. The second attempt was made after the participants were shown the previous records and in addition photos of the face. Lateral cephalograms X-rays were added during the third attempt.
Descriptive statistics are presented as frequencies (per cent) for gender, and mean (± SD) for age. To measure the inter-observer agreement for the three attempts, weighted Kappa coefficients (using quadratic weights) were calculated. At each attempt, 10 Kappa coefficients were produced. Subsequently, the mean and standard deviation were calculated for all 10 coefficients, providing an estimation for the overall agreement at each attempt. This method was repeated for both the right and left maxillary incisors.
The intra-observer agreement was measured by calculating weighted Kappa coefficients between the first and second attempt and between the second and the third attempt for each of the five participants for the right maxillary incisor. Calculations were performed using MedCalc® v14.8 (MedCalc Software bvba, Ostend, Belgium).
The mean kappa coefficients were κ1 = 0.34 (SD ± 0.09) for the first, κ2 = 0.57 (SD ± 0.12) for the second, and κ3 = 0.53 (SD ± 0.28) for the third attempt, indicating a weak overall inter-observer agreement. The individual kappa coefficients ranged from as low as 0.16 to as high as 0.87 (Table 1).
The intra-observer agreement, which was also tested for the right maxillary incisor, was also low with kappa coefficients ranging from 0.18 to 0.66. Between attempt two and three the coefficients increased, however they remained low from a statistical viewpoint. The mean kappa coefficients were 0.27 (SD ± 0.11) and 0.53 (SD ± 0.11) between first and second, and between second and third attempt, respectively (Table 2). The agreement interpretation of weighted kappa value was as follows: less than 0 less than chance agreement; 0.01–0.20 poor agreement; 0.21–0.40 fair agreement; 0.41–0.60 moderate agreement; 0.61–0.80 good agreement; 0.81–0.99 very good agreement (20, 21).
Forty-eight patients were chosen because it was thought to be a reasonable, sufficiently large number that would not allow the participants to recall the choices that they had made during the initial and subsequent selection sessions. A minimum wash out period of 1 week between the three sessions was maintained to enhance this effect. The age range of the patients was from 11 to 61 years old, and their types comprised of all three types of Angle’s malocclusion classification. The malocclusion types and ages range were considered to represent a broad spectrum of orthodontic patients. The gender distribution, favouring females in an almost 2:1 ratio was not believed to have any influence on the choices to be made. Syndromic, cleft, and impacted teeth patients were excluded to prevent the selection process from becoming even further complicated for the participants, by the extreme occlusal and aesthetic variations that this type of patients present.
Dental casts are traditionally used in orthodontic diagnosis and treatment planning but there is no evidence that they are definitely needed. Digital models are a valid alternative to traditional plaster study models in treatment planning for Class II malocclusion patients (22). Intra-oral photographs could be a good alternative to dental casts (23). Han et al. (24), and Nijkamp et al. (25) have shown that the access to cephalometric radiographs and consequent analysis does not affect treatment decisions in adolescents with a Class II, division 1 malocclusion. However, these records are included among the types of records suggested for routine orthodontic diagnosis and treatment planning (26, 27), and therefore they were integrated into the records tested in the present study.
The inability of the participants to observe the patients in postured positions, and the deprivation to discuss treatment options (e.g. extraction versus non-extraction) with the patients might be considered among the weaknesses of this study.
Weighted Kappa coefficients were calculated instead of standard kappa coefficients, due to the fact that the data produced were ordinal, and differences between a low and a high torque bracket represent a much higher disagreement than between a low and standard, or a standard and a high torque bracket (28). Calculating the mean Kappa gives an indication of the overall agreement at each attempt. The standard deviation indicates the variation between all ten Kappa coefficients at each attempt.
The orthodontists who were recruited in the study were exposed to the concept of variable torque orthodontics, and they had been applying it to their practices for a minimum period of 6 months before enrollment. However, it was not possible to define and compare their competence level in clinical application before recruiting them, and this might be considered as an additional weakness of this study. Thus, the differences detected in the results could be attributed to calibration differences or in other words to the difference in absorbing and implementing the variable torque concept on an individual basis (performance bias).
It has been suggested that an orthodontist using a straight-wire appliance may reduce the need for routine torque adjustments by selecting a torque prescription and wire size combination that will cancel out the slot play at the crown’s desired final inclination (12). In case this is the sole criterion when selecting torques, the results of this study would be more uniform. To the contrary, a rather impressive intra- and inter-individual disagreement was observed in our study. This type of low inter- and intra-examiner agreement is not unusual when orthodontic treatment planning decisions are considered. Treatment planning even by the same individual on the same set of records may lead to significant variation (29, 30). The demand on torque selection presupposes that each time the clinicians has to make—on a preliminary subliminal level—, choices in treatment planning and the differences in selections might reflect in reality differences in treatment planning.
Smile aesthetics and inclination of teeth are closely related. In the profile view, a preferred smile is the one with the incisor inclination slightly augmented in the labial direction (6). Torque loss and lingual inclination of maxillary incisors leads to an unpleasant smile and an ‘aged’ appearance (31). When the maxillary incisors are left protrusive in the face they receive higher aesthetic scores (32). Even the buccolingual inclination of posterior teeth has been documented as an important feature of smile attractiveness (33, 34) with a broad smile being preferred as a more pleasant aesthetic feature. Clinicians are faced with the challenge to combine these aesthetic demands with cephalometric standards—which have served for years as guidelines to assess attractiveness—and form a treatment plan. It has been shown though that even a case treated to the American Board of Orthodontic standards might produce a smile which is not aesthetically pleasing (35) and facial harmony may exist within a range of cephalometric values (36). The inter-rater agreement, which was fair (κ1 = 0.34 SD ± 0.09) at the beginning, improved to moderate with the addition of records in both successive attempts (κ2 = 0.57 SD ± 0.12, and κ3 = 0.53 SD ± 0.28, respectively). This low initial agreement could as well represent true differences among clinicians in the types of torques required to treat this group of malocclusions, and achieve an aesthetically pleasing result originating in differences in their final aesthetic goals. The extra photographs providing them with a more comprehensive picture as to the patients’ aesthetic treatment needs might have caused the observed improvement. A meticulous study of the inter-rater data revealed a small drop in agreement at the third attempt (when lateral cephalograms became also available), which was mainly due to a strong disagreement between one of the observers and the other four. It seems that the addition of lateral cephalograms had a tremendous impact on this individual’s perception of torque needs.
It might have been useful to determine rater consistency by repeating torque choices decisions at each stage rather than just between stages. This could be considered another weakness of this study. The low intra-rater agreement (κ = 0.27 SD ± 0.11) between the first and second attempts indicated that either the addition of facial photographs altered the perception the participants about the required torque or the consistency of such decisions is low due to the complicated nature of the problem. The intra-rater agreement improved and become moderate (κ = 0.53 SD ± 0.11) when X-rays were added, indicating that the lateral cephalograms were not of such great importance as to changing the previous choice but they were more confirmatory of it. The addition of extraoral photographs seemed to have played a more catalytic role in altering the initial choices.
Extraoral photographs as well as cephalometric X-rays could be altering the initial assessment of the participating clinicians as to the aesthetic and biomechanical requirements of the reviewed cases leading them to different torque choices at each separate session.
Torque in orthodontics can be a function of a plethora of variables in addition to the torque value. The bracket slot size—wire size difference (37), the bracket slot modulus (38, 39), the wire modulus (40), the mode of ligation (39, 41), the wire corner radius (42), the morphology of teeth (7), the manufacturing precision (43), the tipping of neighbouring teeth (44), and definitely the mechanics that will be employed during treatment can all play a role in establishing the final tooth inclination. These factors may interact in many different combinations not fully understood. Past failures or successes as a result of the these combinations could have been mistakenly attributed to right or wrong torque prescriptions and this could have led to the random pattern of choices observed by the five clinicians who participated in the study. Future research could focus on a comparison between the expectations of the participants and the final tooth inclination achieved as a result of the choices they made. This assessment could add some more insight on the effectiveness of the torque selection process.
Treatment and chair time can probably be reduced by careful selection of torque prescription on an individual basis by reducing the wire adjustments required to achieve the correct inclination. Almost without doubt final detailing wire bending will be necessary but it will probably be minimal and less demanding. However, the indications on how to select torques are not as clear as the orthodontic community might think and more specific guidelines should be developed.
These days a computer algorithm can select the prescription (torque, and tip) that may be milled in the slot and fabricate a fully customized bracket for every single tooth or calculate the bends needed on a wire on a tooth-to-tooth basis, and prepare such a wire. However, even with the most sophisticated of algorithms, occasionally bends may have to be incorporated in the final wire by the clinician when he/she works having excellence in mind.