An index system for the clinical supervision of clear aligner in extraction patients based on the Delphi method and the analytic hierarchy process

Abstract

Background

Clinical supervision of clear aligner treatment in extraction patients is essential during the long process of follow-up period. However, there is insufficient clinical guidance available to define the key points of a standard examination for clinical supervision.

Methods

An initial index system was constructed based on a comprehensive literature review. The final indexes to be used in the system were determined through Delphi process. Weightings for the final index system were determined using the Analytic Hierarchy Process (AHP), which required the development of a judgment matrix to assess the relative significance of each index through pairwise comparisons.

Results

A tiered indexing system was created, consisting of two primary indexes, 10 secondary indexes, and 30 tertiary indexes. The expert authority coefficient (Cr) was 0.89 for the first round of consultation and 0.91 for the second round. The weighting values for the first-level indexes ‘tooth movement’ and ‘dentofacial health’ were 0.38215 and 0.61785, respectively, thus indicating that dentofacial health holds greater importance.

Conclusions

We integrated both qualitative and quantitative analyses to establish a novel index system for clinical supervision of clear aligner treatment in extraction patients to reduce the incidence of complications.

Introduction

In recent years, there is a notable rise in the popularity of clear aligners for orthodontic treatment among patients, primarily because of their improved appearance, comfort, and better oral health benefits [1]. However, clear aligner treatment requires a long follow-up period, thus increasing the incidence of complications due to defects in the design of the treatment plan or problems when the patients wear aligners. Therefore, follow-up monitoring has become an essential aspect of the process of clear aligner treatment. However, there is insufficient clinical guidance to define specific key points for the standard examination used during follow-up monitoring.

It is important to acknowledge that clear aligners are not sufficient to cope with extraction cases. The clear aligner technique is known to exert only limited control over root movement, thus posing a notable drawback in addressing extraction cases in an effective manner [2]. The ability to adequately manage molar anchorage control and incisor torque control is of utmost importance in achieving successful space closure in extraction treatments. A study found that nearly half of the extraction cases showed a decrease in incisor torque, even with the use of attachments or power ridges for torque control [3]. Another research project examined the shifting of teeth near premolar extraction sites as the spaces were closed, uncovering significant tilting [4]. If tilting movement occurs, it may cause fenestration or cortical anchorage, thus resulting in a deterioration of treatment outcomes [5].

Orthodontists design tooth movement patterns based on specific treatment objectives. Nevertheless, the actual control of tooth movement and anchorage often deviates from the intended plan. In addition to tooth movement, a number of dentofacial problems, including periodontal health, temporal-mandibular joint disorders (TMD), and facial aesthetics, significantly impact patient satisfaction with orthodontic treatment due to their influence on performance and functionality [6].

Regular clinical supervision is essential for improving orthodontic effectiveness and ensuring successful results. According to the British Orthodontic Society (BOS), proper supervision involves patients being examined in the presence of the supervising dentist. If this cannot be done, the overseeing dentist must examine the patient every other visit at a minimum [7, 8]. Moreover, patient compliance has been identified as a vital factor in determining the success of clear aligner treatment, with patients expected to wear the appliance for at least 22 h daily. Throughout the treatment process, it is imperative for orthodontists to reevaluate the alignment of the appliance with the teeth and assess the patient's dentofacial health status. Based on this evaluation, a decision should be made regarding the continuation of wearing the appliance or the need for redesigning and producing a new appliance. However, there is a significant lack of research pertaining to clinical supervision indexes for clear aligner treatment. Orthodontists who intend to utilize clear aligners for extraction patients can only rely on their own clinical expertise, expert opinions, and the limited evidence-based results available in the published literature [9].

The Delphi technique is a research methodology employed to gather expert opinion on a specific research inquiry. This technique operates under the assumption that collective intelligence enhances individual judgment and effectively captures the consensus of a group of experts [10]. The analytical hierarchy process (AHP) aims to objectify subjective judgments by introducing numerical rating system and amplifying the weight disparity among the indexes, thereby improving the detection of crucial indexes [11]. The Delphi and AHP methodologies have been extensively employed in various biomedical domains for constructing evaluation index systems [12,13,14]. However, in the field of orthodontics and dentofacial orthopedics, these mathematical approaches have not yet been utilized.

The goal of this research was to build a framework for overseeing clinical procedures in patients undergoing clear aligner treatment and tooth extraction, utilizing both the Delphi and AHP techniques. The results of our study have the potential to enhance the clinical efficiency of clear aligner therapy and lower the occurrence of negative clinical outcomes.

Methods

The research involved both qualitative and quantitative analysis, utilizing a two-round Delphi consultation and AHP to create an index system for clinical supervision in extraction cases during clear aligner treatment, along with determining appropriate weightings. Figure 1 illustrates the procedure of creating an index system and determining weights.

Fig. 1

The process of index construction and weight determination

Assembling experts and constructing the initial indices

To develop the index system, a research group consisting of 20 experts specializing in orthodontics was assembled. To be eligible for this study, participants had to meet the following criteria: (1) have over 5 years of experience working with clear aligners, (2) hold a lecturer position or higher in orthodontics, and (3) provide written informed consent and willingly participate in the study. The criteria were set up to guarantee the dependability of the consultation results. Moreover, we delved into their publication records in prominent orthodontic journals, with a particular focus on those pertaining to clear aligner therapy and clinical supervision. Additionally, we requested them to submit case reports and share their firsthand experiences in handling diverse challenges and complications encountered in such cases. This enabled us to comprehensively evaluate their practical skills and problem—solving capabilities. The initial indexes used in this study were developed according to literature documents as well as the perspective of the research group.

The Delphi method

In accordance with the guidelines for conducting and reporting Delphi Studies, a two-round Delphi survey was carried out [15]. A consensus was reached by the group of 20 experts through the utilization of the Delphi process, which involved two rounds of questionnaires delivered and received via email. Participants were allotted a period of two weeks to fill out the surveys, during which they had to evaluate the importance of the criteria in every cycle. Each item's significance was evaluated using a 5-point Likert scale ranging from 1 (unimportant) to 5 (highly important) [16]. The indexes were calculated based on the coefficient of variability (CV) and the mean importance score. Following this, experts evaluated the judgment criteria (Ca) and their level of familiarity (Cs), considering theoretical analysis, practical experience, and reference literature from both domestic and international sources. To be considered for the ultimate indexes, the minimum average importance score of 2 and a maximum CV of 0.4 in the initial round were required, while in the subsequent round, the minimum average importance score was raised to 2.5 and the maximum CV was lowered to 0.3. Indexes failing to meet the inclusion criteria in terms of mean value or CV were either adjusted or eliminated based on expert judgment. Following this, experts proceeded to rate the judgment criteria (Ca) and their level of familiarity (Cs). Four aspects were considered in the judgment criteria: practical experience, theoretical analysis, intuitive feeling and reference literature from both domestic and international sources (Supplementary File 3). The specialists' knowledge of the indexes was subsequently classified as extremely familiar equals 1.00 point, very familiar equals 0.80 points, generally familiar equals 0.60 points, less familiar equals 0.40 points, and unfamiliar equals 0.20 points. Indexes that failed to reach consensus in the first round were carried over to the following survey until agreement was reached, ultimately resulting in the creation of the final index system.

Weight determination and consistency test

Following the determination of the final index system, the specialists evaluated the importance of each ultimate indexes using the AHP method. Hierarchical models, consisting of three levels, were constructed using AHP to address intricate problems [13]. Subsequently, a judgment matrix is constructed using Satty’s fundamental scales (1–9) for pairwise comparisons and to determine the relative significance of each index within the same hierarchical level (Supplementary File 4) [12]. Each indicator's initial weight and combination weight were calculated, and then a consistency check was performed using yaahp12.4 (yaahp software, Taiyuan, Shanxi, China). The average of each row in the judgment matrix was computed by finding the arithmetic mean; where 'a_ij' represents the element in the i-th row and j-th column of the original judgment matrix, and 'n' represents the number of indexes. The consistency test of the judgement matrix was conducted to ensure its reliability. The highest eigenvalue of the decision matrix was determined as $\lambda_{max}=\frac{1}{\mathrm{n}}\sum^n_{i=1}\frac{\left(\sum^n_{j=1}\alpha_{ij}w_j\right)}{w_i}$; the CI was calculated using the formula ${\text{CI}} = \frac{{\mathop \lambda \nolimits_{\max } - n}}{n - 1}$; the CR was calculated as $CR = \frac{CI}{{RI}}$ where RI represents the mean random consistency index; the RI value is based on the average consistency index of randomly created reciprocal matrices [13].

Statistical analysis

The statistical analysis was conducted using IBM SPSS Statistics for Windows, Version 26.0 (IBM SPSS Inc., Chicago, USA). Mean and standard deviation are used to present descriptive statistics. The authority coefficient (Cr) was equal to half of the summary of Ca and Cs. Ca takes into account four aspects: practical experience, theoretical analysis, intuitive feeling, and reference to literature(Supplementary File 2). Cs reflects the experts' familiarity of the indexes, which is evaluated by the experts. If an expert is "extremely familiar" with an indicator, it is scored 1.00 point; "very familiar" corresponds to 0.80 points; "generally familiar" is 0.60 points; "less familiar" is 0.40 points; and "unfamiliar" is scored 0.20 points.

Kendall's coefficients of concordance (Kendall's W) and the Chi-squared test were employed to measure the level of agreement among expert opinions. Kendall's W is a non—parametric statistical measure used to assess the degree of agreement among multiple raters. In our study, we use it to measure the level of agreement among expert opinions on the importance of various indexes. Kendall's W is applied to measure the degree of consensus among expert opinions. The calculation process is as follows: First, assume there are n experts evaluating m indices. For each index, we need to rank the scores given by all experts. Then, calculate the sum of squared differences of the ranks for each index. R_ij represent the rank of the score given by the i—th expert for the j—th index. The sum of squared differences for the j—th index is calculated as $S_{j} = \sum\limits_{i = 1}^{n} {\sum\limits_{k = i + 1}^{n} {(R_{ij} } } - R_{kj} )^{2}$. Then, we calculate the total sum of squared differences $S = \sum\limits_{j = 1}^{m} {S_{j} }$. Finally, Kendall's W is calculated using the formula $W = \frac{12S}{{m^{2} (n^{3} - n)}}$. A higher Kendall's W value indicates a greater degree of consensus among the experts, lending more credibility to the index system we developed. The average random consistency index was utilized to assess the consistency of various judgment matrices with Yaahp 12.4 software. A CR value of 0.10 or less suggests that the judgment matrix shows acceptable consistency [13].

Results

The authority coefficient and degree of collaboration among the authorities

This study involved two round expert consultations. Supplementary File 1 contains comprehensive fundamental details about the specialists. 20 specialists were asked to join the Delphi procedure via email. Experts participated at a rate of 100% in both rounds, demonstrating a significant level of engagement. Experts' authority coefficient was determined using their self-assessment scores from Table 1. During the first round, the mean credibility stood at 0.89, rising to 0.91 in the subsequent round, surpassing the threshold of 0.7, indicating a significant level of trustworthiness [17]. Thus, the results are reliable and persuasive. We then assessed the level of collaboration among specialists using Kendall's W statistic. Typically, Kendall's W ranges from 0 to 1, with a higher value indicating a greater level of coordination among the experts. In the first round of the Delphi process, Kendal’s W was 0.73. Kendall's W in the second round was 0.53. There was good coordination among specialists in both sessions.

Table 1 The Delphi expert authority coefficient, degree of opinion concentration and opinion coordination

Creation of the index framework

The original index structure, consisting of two primary, 13 secondary, and 39 tertiary indices, was created using information from literature and the research team's perspective (Supplementary File 2). During the initial consultation phase, the average significance of each index element ranged from 1.20 to 5.00. During the second round, the average significance of each index item ranged from 2.55 to 5.00. The final index system only included indexes that satisfied the inclusion criteria. After the two-round Delphi process, three second-level indexes and nine third-level indexes were deleted. The ultimate indexing system comprised of two primary indexes, 10 secondary indexes, and 30 tertiary indexes (Table 2 and Fig. 2).

Table 2 The final index system for clinical supervision during clear aligner treatment in extraction patients

Fig. 2

The final index system for clinical supervision during clear aligner treatment in extraction patients. A indicates the first level indexes, (B) indicates the second level indexes, and (C) indicates the third level indexes

Weight distribution of the indexes

The AHP method was used to establish the weight coefficient for the comprehensive evaluation index at every level. AHP was used to calculate the normalized weight and combination weight for each final indexes, as shown in Table 3 and Fig. 3A greater weight suggested that the indicator played a more significant role in clinical supervision for extraction patients undergoing clear aligner treatment. Each altered judgment matrix had a consistency ratio below 0.10, demonstrating good consistency across all matrices. For first-level indexes, the weight values of ‘tooth movements’ (A1) and ‘dentofacial health’ (A2) were 0.38215 and 0.61785, respectively. For second-level indexes, ‘periodontal health’ (B7) had the highest comprehensive weight value of 0.13637, while ‘mandibular molars’ (B4) had the lowest comprehensive weight value of 0.01365. For third-level indexes, the highest comprehensive weight value of 0.01417 was seen in ‘temporal-mandibular joint pain’; furthermore, horizontal movement of ‘mandibular molars’ had the lowest comprehensive weight value of 0.00011.

Table 3 The weight values of the three-level index system

Fig. 3

A multi-nested sector statistical chart showing the weights of different level indexes. Each ring represents the indexes of one level. The area of each index within the ring of its level represents the percentage of its weight. The weight values of each index can be found in Table 3. A includes all the First-level and Second-level indexes and their weights. The inner ring represents the First-level indexes and their weights, while the outer ring represents the corresponding Second-level indexes and their weights. The blue part of the sectors represents the A1 Tooth movements indexes and its subordinate indexes, and the orange part of the sectors represents the A2 Dentofacial health and its subordinate indexes. B shows A1 Tooth movements, its Second-level indexes (the second ring), Third-level indexes (the third ring) and their weights. C shows A2 Dentofacial health, its Second-level indexes (the second ring), Third-level indexes (the third ring) and their weights

Discussion

Clear aligners are not sufficient to cope with extraction cases as the desired tooth movement may not always be achieved as planned [18]; furthermore, it is important to carefully consider dentofacial health [3, 4, 6]. Consequently, clinical supervision is crucial to enhance the efficiency of clear aligner treatment in extraction cases. In this study, we constructed an index system for clinical supervision during clear aligner treatment in extraction patients, and further determined the weightings for this system by employing the Delphi method in conjunction with AHP.

By consulting the existing literature, we constructed an original index framework with two first-level, 13 s-level, and 39 third-level indices. After two rounds of the Delphi process, a total of three second-level indexes and nine third-level indexes were removed. The ultimate index system consisted of two primary indices, along with a total of 10 secondary indices and 30 tertiary indexes. The calculated weight of the indexes in the AHP process indicates its significance. The first-level indexes include ‘tooth movement’ (A1) and ‘dentofacial health’(A2), of which the latter holds higher weights. Dentofacial health exerts a major impact on the appearance and functionality of patients, which are often neglected by inexperienced practitioners [6]. Orthodontic treatment is frequently perceived by patients as a means to enhance dentofacial appearance, with the anticipation of improved self-esteem, oral function, and social interactions [19]. Several studies have demonstrated that clear aligner treatment is linked to heightened satisfaction, improved oral health-related quality-of-life, and reduced negative impacts on oral health than conventional fixed orthodontics [20, 21]. The most influential second-level index was ‘periodontal health’ (B7), followed by ‘temporal-mandibular joint’ (B9), thus indicating that periodontal health and the TMJ are the most important factors for clinical supervision as dentofacial health indices. The incidence of periodontal pathologies is not increased by orthodontic treatment alone; however, the periodontal status of orthodontic patients is significantly influenced by oral hygiene procedures [22, 23]. Recent studies have concluded that patients undergoing treatment with clear aligners exhibit superior periodontal health indices in comparison to those treated with fixed appliances [22, 24]. However, the clear aligner appliance exerts a significantly higher instantaneous stress when compared to conventional fixed orthodontics [25, 26], which may result in tooth mobility for individuals with poor periodontal conditions. According to a recent finite element study, the step distance of intrusion in patients with mild, moderate and severe periodontal disease should be less than 0.18mm, 0.15mm and 0.10mm, respectively [27]. For numerous years, clinicians and researchers have extensively regarded occlusion as a prominent etiological factor, either directly or indirectly, that contributes to TMD [28]. According to current evidence, orthodontic treatment does not have the ability to cause or improve TMD [28, 29]. Recent studies on clear aligners have found an increase in the electromyographic activity of the jaw muscles after six months of treatment. Additionally, there have been reports of an uptick in individuals experiencing muscle soreness upon waking after a month of using clear aligners. There was also a notable increase in the severity and number of tender areas when examining the TMJ and facial muscles [30, 31]. However, none of the participants included in this previous study exhibited clear clinical manifestations of TMD. Therefore, it is imperative to be cautious when prescribing clear aligners to patients who may be susceptible to experiencing jaw muscle pain. From these results, it was evident that occlusion is another important indicator [30, 31]. Occlusion is also a decisive factor in orthodontic treatment; ideal occlusion can involve healthy masticatory muscles and other oral functions. Additionally, American scientists argued that all types of occlusal variations and 'abnormalities' lead to TMD symptoms, so they must be addressed proactively due to their impact on TMD. Therefore, occlusal adjustment is an essential procedure in the process of orthodontic treatment; the ideal occlusal target and stability are related to orthodontic treatment [32].

With regards to the second-level indices, our findings indicate that the weightings of maxillary tooth movement were higher than those of mandibular tooth movement. Furthermore, the magnitude of anterior tooth movement outweighed that of posterior tooth movement. These observations suggest that heightened clinical supervision should be directed towards maxillary incisors during clear aligner treatment for patients undergoing extractions. Of the third-level indices, sagittal movement emerges as the most crucial index, followed by vertical movement, while horizontal movement had the lowest weighting. In cases where extraction is necessary, the process of bodily retraction of incisors involves intricate tooth movements that require precise control of both intrusion and torque in the incisors. Alternatively, there may be lingual tipping, extrusion, and a clockwise moment of incisors, often known as the roller-coaster effect [33]. The consequences may result in a significant overjet of the front teeth, a open bite in the premolars, and forward tilting of the molars [34, 35]. A study conducted recently discovered that clear aligners were capable of achieving only 42% of the intended root torque and incisor intrusion. As a result, researchers have been actively exploring methods to mitigate this side effect. Liu et al. Showed that over correction can result in the intrusion of incisors and torquing of palatal roots, which can be enhanced by attaching additional devices to the canines, leading to increased anchorage needed from the premolars and molars [36]. Cheng et al. revealed that the upper central incisor could approach translation with a proper power ridge height [37]. Other studies showed that the first molars displayed greater mesial tilting, and intrusive movement than anticipated [38]. In clinical practice, discrepancies between the actual tooth movement and the intended tooth movement can result in a poor fit of the clear aligner appliance. If the aligner's light transmission gap measures less than 1mm, orthodontists may recommend patients bite a glue stick more frequently to encourage tooth movement towards the desired position and prolong the duration of aligner wear. However, when the light transmission gaps are consistently large across multiple teeth sites and the tooth movement becomes significantly unmanageable, it becomes imperative to consider restarting the treatment to prevent uncontrollable tooth movement and potential exacerbation of the dentofacial teeth [39]. In addition, patient compliance directly affects the treatment effect. To improve patient compliance, doctors can enhance communication and education, build a good doctor—patient relationship, increase the frequency of clinical supervision, and provide personalized wearing suggestions based on the patient's age, occupation, living habits and other factors.

Hypothetical scenarios for practical application. Consider we have an orthodontic patient who requires tooth extraction for clear aligner treatment. Prior to the treatment, doctors need to gather detailed patient information, such as oral examination results, X—ray films, and model analysis data. They then compare the collected data with each index in the index system, assign quantitative scores to each index, and perform weighted calculations based on the weight of each index to obtain a comprehensive evaluation. With this evaluation, doctors can clearly identify the patient's oral problems and potential risks, and thus formulate a personalized treatment plan. For instance, if a patient has a low score on the "temporomandibular joint" index and is at risk of temporomandibular joint disorder, the doctor will be extra cautious about the magnitude and direction of the corrective force when designing the clear aligner treatment plan to prevent overburdening the temporomandibular joint. If the treatment plan involves substantial anterior tooth retraction, orthodontists will focus more on the sagittal tooth movement index to avoid adverse side effects like the "roller—coaster effect". During the orthodontic treatment, patients are re—evaluated regularly according to the index system. The oral condition changes are continuously monitored to promptly detect problems and adjust the treatment plan. For example, a comprehensive oral examination is carried out every 3—6 months. Based on the examination results, each index is re—scored. If the patient's periodontal health index deteriorates, the doctor can immediately strengthen oral hygiene guidance or adjust the treatment plan to reduce periodontal tissue irritation. If the weights of temporomandibular joint—related factors suggest a high—risk situation, orthodontists may slow down the treatment progress, adjust the aligner wear time, or even consider additional diagnostic tests to safeguard the patient's temporomandibular joint health. After the treatment is finished, patients are evaluated once more using the index system. By comparing the scores before and after treatment, the treatment effect can be assessed. If the patient's comprehensive score improves significantly after treatment, it shows that the treatment plan has been effective; otherwise, it's necessary to reflect on and summarize the treatment process to gain experience for future treatments.

There are some limitations to this study. First, the reliance on expert opinions via email questionnaires in the Delphi process has potential biases. The experts' individual backgrounds, such as their educational experiences, areas of expertise, and years of clinical practice, could have influenced their judgments and the assigned weights of the indexes. Different experts may have had diverse perspectives on the significance of various factors in clear aligner treatment for extraction cases. Also, the email—based approach might have led to misinterpretations of questions. These biases could cause the index weights determined in the study to deviate from the true values, potentially affecting the accuracy of the index system. Second, the AHP method has limitations in fully capturing the complexity of orthodontic treatments. AHP assumes the independence of criteria, but in reality, factors like tooth movement and dentofacial health are often interrelated. For instance, improper tooth movement can impact periodontal health, and changes in occlusion can affect the temporomandibular joint. The inability of AHP to account for these complex interactions accurately may result in an oversimplified view of the relationships between indices. Future research can aim to address these issues, such as using more diverse data collection methods and refining the AHP process to better account for the complexity of orthodontic treatments. In our future projects, we plan to use this index system for clinical supervision in diverse extraction patients undergoing clear aligner treatment, aiming to improve the evaluation system continuously.

Conclusions

The study's results are expected to improve the effectiveness of orthodontic treatment for patients who require extractions using clear aligners, providing a useful and objective mathematical tool. We established an index system for clinical supervision of clear aligner treatment in extraction patients for the first time by combining a two-round Delphi method and AHP. Nevertheless, this assessment index system needs to be continuously enhanced in practical application.