Association between crowding estimation and extraction recommendations in orthodontics

Introduction

Little is known about how precisely orthodontists in the United States (US) assess crowding or at what range of crowding they recommend extraction. This study aimed to assess the relationship between estimated crowding in patients with a Class I relationship and extraction recommendation by orthodontists in the US. The secondary aims were to evaluate the accuracy and precision of clinician estimations and determine if clinician background traits play a role in extraction decision-making.

Methods

An electronic survey was prepared using 4 patients with a Class I relationship with anterior crowding selected from a University Orthodontics Clinic and was sent to approximately 10,400 subjects through Facebook and the American Association of Orthodontists Partners in Research program.

Results

From the 297 responses received, most clinicians recommended extraction once crowding reached 9-10 mm in either the maxilla or the mandible. The data from 2 patients suggest this decision was more strongly correlated with mandibular crowding. Clinician estimations varied widely but, on average, were precise within approximately 2 mm of objective measurements. There was a tendency to overestimate crowding, especially by Northeastern practitioners. Clinicians who reported routinely measuring crowding or who reported that they recommended extractions to >10% of their patients were 1.2-2.0 and 1.4-1.6 times more likely, respectively, to recommend extraction in the patients.

Conclusions

Crowding estimation was highly subjective and varied widely among clinicians. Most clinicians recommended extraction once maxillary or mandibular crowding approximated 9-10 mm. Some clinician demographics were correlated with the precision and accuracy of estimations and the likelihood of extraction in the patients.

Highlights

•

Crowding estimation and treatment preferences were assessed in Class I patients.
•

The survey participants included 297 orthodontic clinicians who evaluated 4 patients.
•

Clinicians reported wide ranges and were precise within 2 mm of measured values.
•

Most preferred extraction with 9-10 mm crowding in the maxilla or the mandible.

Dental crowding has been identified as the main reason people seek orthodontic treatment ^, and may be defined as the rotation, ectopic eruption, or impaction of teeth that occurs when the space required for ideal alignment exceeds the available space in the dental arch. This malocclusion is more prevalent in the anterior dentition and affects the mandible more than the maxilla. Mandibular crowding alone has been found to impact around 63% of the population and tends to worsen with age.

Crowding can be resolved in several ways, depending on the severity or amount present. Nonextraction treatment using methods such as mesiodistal reduction of teeth or expansion of the dental arch is typically recommended for mild-to-moderate crowding, whereas extractions may be considered as crowding becomes severe. Several studies have examined variation in crowding estimations and treatment decisions among clinicians and found considerable differences. Several factors might contribute to these inconsistencies, such as differences in training and types of diagnostic tools used. However, one of the biggest contributors is the lack of clearly defined standards for classifying orthodontic problems and making treatment decisions.

It is widely accepted that the extraction decision is multifactorial and depends on several other factors, including facial profile, incisor inclination, periodontal health, and curve of Spee. ^, ^, Although many of these factors have been well studied in connection to the extraction decision, there are fewer evidence-based guidelines with regard to crowding amounts. The only known parameters are those by Proffit et al, who suggest either extraction or nonextraction treatment when crowding is 5-9 mm and extraction treatment once crowding reaches 10 mm. Identifying a range for crowding at which extractions are generally considered can help orthodontists make this decision, especially when it is a patient with a borderline extraction, and there is still uncertainty after considering all other factors. The extraction of permanent teeth is irreversible and is considered by patients to be one of the most anxiety-provoking dental procedures. Furthermore, the decision between extraction and nonextraction treatment can significantly impact the health of the periodontium. Therefore, the practitioner must consider all diagnostic factors before recommending this.

There is little known about how precisely orthodontists in the United States (US) assess crowding in their practices or at what range of crowding they decide to recommend extraction. This study aimed to assess the relationship between estimated crowding and the recommendation for extraction by orthodontists in the US. The secondary aims are to evaluate the accuracy and precision of clinicians’ estimations of crowding compared with objective manual measurements and determine if clinicians’ sex, region, experience, method of crowding assessment, and extraction prevalence play a role in extraction recommendation.

Material and methods

This study was approved by the University’s Institutional Review Board. Patients were selected from the Department of Orthodontics on the basis of the following inclusion criteria: full permanent dentition, Class I molar classification, normal cephalometric incisor inclination, presence of maxillary and or mandibular crowding, lip competency, and complete records available (photographs, intraoral scan, panoramic and lateral cephalometric radiographs). Exclusion criteria included missing teeth other than third molars, worn or fractured dentition, restorations that make up ≥50% of the crown structure, defective or grossly contoured restorations, root canal treatment, ankylosis, moderate-severe decay or recession, periodontal disease, and other bone pathology. Four of the 10 screened patients met the inclusion criteria and were selected for the survey.

Resin models for each patient were printed with a SprintRay (SprintRay Inc, Los Angeles, Calif) 3-dimensional printer. Blinded to all records and patient identifiers, 1 orthodontic graduate student (K.T.) measured crowding (mm) on each model by subtracting the required arch perimeter, or the sum of the tooth widths mesial to the first molars, from the available arch perimeter. A Boley gauge was used for all measurements, and the main method for measuring the available arch perimeter was with linear segments. The sum of 5 linear segments comprised the available arch perimeter: mesial of the first molars to the mesial of the first premolars, mesial of the first premolars to the mesial of the lateral incisors, and from the mesial of the lateral incisor to the mesial of the contralateral lateral incisor. ^, For comparison, the available arch perimeter was also assessed by adapting a strand of brass wire around the arch mesial to the first molars over the buccal cusp tips of the premolars and along the incisal edges of the anterior teeth in a symmetrical arch form. ^, The same student assessor repeated the crowding analyses in a random order 4 weeks later. The average crowding measurements taken from the 2-time points (T1 and T2) were used as the measured crowding value for each patient. All landmarks on the lateral cephalometric radiographs were traced by the same orthodontic resident and reviewed by the same faculty member.

To be eligible to participate in this study, an orthodontist or orthodontic resident needed to be practicing in the US or training at an accredited US orthodontic residency program.

An online survey was sent to all American Association of Orthodontists (AAO) Partners in Research Program members and international Orthodontic Pearls Facebook (Facebook Inc, Menlo Park, Calif) group members. The consent and recruitment form included with the survey specified that the survey was for clinicians practicing in the US. Survey questions were multiple-choice and open-ended ( Supplementary Material ). The first part of the survey included questions about the clinicians’ background: sex, region, years of experience, extraction prevalence, and preferred method for crowding assessment. The second part of the survey included images of the records for each patient ( Supplementary Figs I-IV ) along with follow-up questions about crowding assessment and preferred extraction or nonextraction treatment. Patients were randomized in the order that they were presented.

This study defined accuracy as the average difference between the clinician’s crowding estimation and the objective measurement. A negative value indicates underestimation by the clinician, whereas a positive value indicates overestimation. Precision, or measurement error, is the average absolute difference between the clinician’s crowding estimation and the objective measurement. A lower absolute difference indicates that the clinician’s estimation is more precise and closer to the objective measurement.

Statistical analysis

Based on a 2016 survey of AAO members regarding orthodontist decisions for extraction, the expected sample size was 200-300 practitioners, which was determined to have 80% power with a 2-sided 0.05 significance level to detect a 40%-50% increase in the odds (odds ratio, 1.4-1.5) for recommending extraction for each 1 mm increase in crowding if the rate of recommending extraction is 40%-60% for crowding around 5 mm. Intraexaminer error in crowding measurements was calculated with Dahlberg’s formula. Descriptive summaries were run to determine general trends in the data. Paired t -tests and linear regression using the generalized estimating equations method to account for the repeated measures were used to compare the accuracy and precision between the patients and arches. Two-sample t -test, 1-way analysis of variance, nonparametric Fisher exact test, and Pearson’s chi-square test determined the association between practitioners’ characteristics and precision and extraction recommendation. Holm’s method was used to adjust the statistical significance ( P value) for post-hoc multiple testing. The relationship between crowding estimation and extraction recommendation was examined using scatter plots and locally weighted smoothing.

Results

The survey was distributed to 1904 subjects through the AAO Partners in Research Program and approximately 8500 subjects through the Facebook Orthodontic Pearls group for 10,404 subjects. Of the 373 clinicians who responded to initial background questions, 266 completed all questions for all patients, 31 completed questions for 1-3 patients, and 76 did not complete any patient questions. A breakdown of 1-3 patients completed by the 31 subjects is shown in Table I . One hundred seventy-nine participants who completed questions were from Facebook and 118 from the AAO. Patients A and B received 280 responses each, whereas C and D received 281. As the Orthodontic Pearls Facebook group is international, a true response rate for this group could not be determined. It is unknown how many total members were in the US survey response rate for the AAO group, which was 7.2%. Overall, the experience of clinicians varied, with 23.3% having practiced for <5 years, 28.2% for 5-15 years, and 38.9% for ≥15 years ( Table II ). Most responders from the AAO group had more than 15 years of experience (60.1%), whereas those who responded from Facebook were more evenly distributed by experience ( Supplementary Tables I and II ). Most respondents were male (63%), practiced in the Western region (30%), visually estimated crowding (81.8%), and reported they had recommended extraction to 10%-25% of their patients (57%) ( Table II ). There was no association between years practiced or clinician sex and reported frequency of extractions ( Table III ).

Table I

Number of clinicians completing questions for 1-3 patients

Patient questions completed	n (%) ^†
Patient A only	4 (12.9)
Patient B only	6 (19.3)
Patients A and B	5 (16.1)
Patients C and D	8 (25.8)
Patients B, C, and D	3 (9.6)
Patients A, C, and D	5 (16.1)

† n = 31.

Table II

Clinician demographics

Characteristic	n (%) ^†
Years practiced
0	36 (9.7)
<5	87 (23.3)
5-15	105 (28.2)
>15	145 (38.9)
Sex
Female	134 (35.9)
Male	235 (63.0)
Prefer not to say	4 (1.1)
AAO regional group
Midwest	74 (19.8)
Northeast	66 (17.7)
West	112 (30.0)
South	104 (27.9)
Other	17 (4.6)
Recommend extraction
<10%	90 (24.2)
10-25%	212 (57.0)
25-75%	70 (18.8)
Crowding assessment
Visual estimation only	305 (81.8)
Measurement on study models	43 (11.5)
Visual estimation and other methods	25 (6.7)

† n = 373.

Table III

Reported extraction frequency by years practiced and clinician sex

Variables	Recommend extraction, n (%)			P value
Variables	<10%	10-25%	25-75%	P value
Years practiced				0.307
<5	24 (19.7)	72 (59.0)	26 (21.3)
5-15	27 (25.7)	64 (61.0)	14 (13.3)
>15	39 (26.9)	76 (52.4)	30 (20.7)
Sex				0.972
Female	34 (25.4)	75 (56.0)	25 (18.7)
Male	55 (23.5)	134 (57.3)	45 (19.2)

Dahlberg’s error between the student assessor’s T1 and T2 crowding measurements was small, revealing an error of 0.27 mm and 0.34 mm for the linear segments and brass wire methods, respectively. On average, the measurements were higher using linear segments than the brass wire method (mean difference, 0.39-1.37 mm; P <0.005).

Crowding estimation, severity, and treatment preferences for each case are listed in Tables IV-VI . There was a wide variation in estimation for all patients, ranging from 0 mm to 18 mm for the least crowded case ( Figs I and II ). Of the 4 patients, patient A was perceived to have the least crowding, with most clinicians preferring nonextraction treatment (95%). Patients B and D were found to have the most crowding, with the majority recommending extractions (77.2%-84.3%). Crowding in patient C was estimated to be more than in patient A but less than in patients B and D, with 72.8% wanting to treat nonextraction.

Table IV

Clinician crowding estimation for each patient

Characteristics	Patients
Characteristics	A	B	C	D
Maxillary crowding (mm)
Mean ± standard deviation	5.3 ± 2.6	8.4 ± 2.9	5.3 ± 2.3	9.4 ± 3.1
Median (IQR)	5.0 (4.0-6.0)	8.0 (6.0-10.0)	5.0 (4.0-6.0)	8.0 (8.0-10.0)
Range	0.0-18.0	2.0-20.0	0.0-15.0	5.0-20.0
Mandibular crowding (mm)
Mean ± standard deviation	5.1 ± 2.3	9.1 ± 3.4	7.3 ± 2.8	8.0 ± 2.9
Median (IQR)	5.0 (4.0-6.0)	8.0 (7.0-10.0)	7.0 (5.0-8.0)	8.0 (6.0-10.0)
Range	0.0-15.0	4.0-20.0	1.0-20.0	3.0-20.0

IQR , interquartile range.

Table V

Clinician crowding severity classification for each patient

Characteristics	Patients
Characteristics	A (n = 280)	B (n = 280)	C (n = 282)	D (n = 280)
Maxillary severity
None	2 (0.7)	0 (0.0)	0 (0.0)	0 (0.0)
Mild	101 (36.1)	6 (2.1)	119 (42.2)	0 (0.0)
Moderate	164 (58.6)	149 (53.2)	160 (56.7)	65 (23.2)
Severe	13 (4.6)	125 (44.6)	3 (1.1)	215 (76.8)
Mandibular severity
None	3 (1.1)	0 (0.0)	0 (0.0)	0 (0.0)
Mild	111 (39.6)	0 (0.0)	24 (8.5)	15 (5.4)
Moderate	155 (55.4)	114 (40.7)	186 (65.7)	137 (48.9)
Severe	11 (3.9)	166 (59.3)	73 (25.8)	128 (45.7)

Note. Values are presented as n (%).

Table VI

Clinician treatment recommendation for each patient

Treatment	Patients
Treatment	A (n = 280)	B (n = 281)	C (n = 283)	D (n = 280)
Extractions	14 (5.0)	217 (77.2)	77 (27.2)	236 (84.3)
Nonextraction	266 (95.0)	64 (22.8)	206 (72.8)	44 (15.7)