The primary aims of this systematic review were to assess objective levels of wear of removable orthodontic appliances and components vs both stipulated and self-reported levels. We also aimed to consider patient experiences and the effectiveness of interventions geared at enhancing compliance.
Electronic databases and reference lists of relevant studies were searched with no language restriction (PROSPERO: CRD42016036059). Randomized and nonrandomized controlled trials, prospective cohort studies, case series, qualitative and mixed-methods studies objectively assessing compliance levels were identified. The quality of the studies was assessed using the Cochrane Collaboration’s risk of bias tool, risk of bias in non-randomized studies of interventions (ROBINS-I), or mixed-methods appraisal tool based on their design.
Of 4269 records, 80 full texts were obtained, with 24 studies meeting the selection criteria. Of these, 11 were included in the quantitative synthesis. A weighted estimate of objectively assessed compliance levels in relation to stipulated wear time was calculated with the discrepancy highest in the headgear group (5.81 hours per day, 95% confidence interval, 4.98, 6.64) based on 6 studies. The mean discrepancy between self-reported and objectively assessed headgear wear was 5.02 hours per day (95% confidence interval, 3.64, 6.40). Compliance level was not directly related to appliance type ( P = 0.211). Thematic synthesis was not undertaken because of the limited number of qualitative studies.
Compliance with removable orthodontic appliances and adjuncts is suboptimal, and patients routinely overestimate duration of wear. Techniques for improving compliance have promise but require further evaluation in high-level research.
Suboptimal compliance is typical with all types of orthodontic appliances.
Actual wear time is approximately 5 hours per day less than stipulated.
Overreporting of appliance wear was found with all types of appliances.
Factors influencing compliance levels are not fully understood.
Compliance with removable orthodontic components can have a telling bearing on the efficiency and success or failure of orthodontics in the short and long term. Removable appliances continue to be popular despite the availability of compliance-free alternatives including fixed functional appliances, implant-supported devices, and fixed retainers. The continued use of removable components can be attributed to the relative simplicity of fabrication and adjustment, low cost, and reduced chair-side time. Moreover, fixed appliances place a higher premium on optimal oral hygiene and, although breakages of both removable and fixed orthodontic appliances are frequent, fractures of fixed appliances are considerably more common at least in the short term.
In terms of clinical effectiveness and associated harms, a lack of high-quality evidence to differentiate fixed and removable adjuncts including functional appliances and retainers has been demonstrated. Notwithstanding this, it is accepted that achievement of optimal outcomes with removable appliances hinges on good compliance. A wealth of research has pointed to suboptimal compliance levels among orthodontic patients; moreover, candid patient reports of wear duration are typically not forthcoming. Consequently, indwelling microelectronic timers have gained traction, primarily as a research tool, to corroborate reported estimates of appliance wear. Depending on the timing and extent of poor compliance and overreporting of wear, these issues risk stagnation of treatment, unnecessary changes in treatment plans, compromised treatment outcomes, and orthodontic relapse after treatment.
The problem of suboptimal compliance with removable appliances and adjuncts has been exposed in numerous contexts; however, relatively few interventions aiming to optimize appliance wear time have been studied. Furthermore, there is a lack of consensus regarding the anticipated wear levels associated with removable adjuncts and how this relates to reported wear durations. The primary aim of this review was therefore to assess levels of compliance with various removable orthodontic appliances and adjuncts. Secondary aims were to assess the effectiveness of interventions used to improve compliance levels, to explore patient experiences and interventions to enhance compliance with removable adjuncts, and to identify factors affecting cooperation.
Material and methods
Protocol and registration
The protocol for this systematic review was prospectively registered on PROSPERO ( www.crd.york.ac.uk/PROSPERO ; CRD42016036059).
The following inclusion and exclusion criteria were applied.
Study design. Quantitative studies including randomized, nonrandomized controlled clinical trials, prospective cohort studies, and case series (minimum sample size, 20 patients) incorporating objective data on compliance levels were eligible. Qualitative studies exploring patients’ views and experiences of removable orthodontic appliances or adjuncts and the interventions used to improve compliance levels (including barriers and facilitators affecting wear of the appliance) were included. Mixed-methods studies in which quantitative or qualitative components met the above criteria were also included.
Participants. Patients of any age treated with headgear, protraction facemask, chincup, removable appliances, removable retainers, or fixed appliances with intraoral elastics as adjuncts were eligible.
Interventions and comparators. Orthodontic interventions including headgear, protraction facemask, chincup, removable appliances, removable retainers, or fixed appliances with intraoral elastics as adjuncts were included. The use of means of improving compliance was also to be assessed.
Outcome measures. Primary outcomes included compliance levels with orthodontic regimens (hours per day of wear or percentage of compliance) in relation to both stipulated and patient-reported levels of wear. Secondary outcomes were the impact of the interventions used to improve compliance levels and delineation of patient experiences and factors influencing compliance levels with wear regimens.
Information sources, search strategy, and study selection
The following electronic databases were searched from inception to May 2016 without language restrictions: MEDLINE via OVID using specific search terms ( Appendix 1 ), PubMed, the Cochrane Central Register of Controlled Trials, Web of Science Core Collection, and LILACS and BBO databases. Unpublished clinical trials were accessed electronically using the following online portals: ClinicalTrials.gov ( www.clinicaltrials.gov ), the National Research Register ( www.controlled-trials.com ), and ProQuest Dissertation and Thesis database ( pqdtopen.proquest.com.easyaccess2.lib.cuhk.edu.hk ). Citation tracking and searching of reference lists of the included studies was performed to identify relevant research. The authors of the included studies were contacted via e-mail if additional information was required.
Risk of bias and quality assessment in individual studies
After identification and retrieval of relevant abstracts, 2 authors (D.A., P.S.F.) independently identified studies that met the inclusion and exclusion criteria and assessed their quality. Reconciliation of disagreement followed discussion. The quality of randomized controlled trials was assessed using the Cochrane Collaboration’s risk of bias tool with only studies at low or unclear risk of bias included in the meta-analysis. The following domains were considered: sequence generation, allocation concealment, blinding of outcome assessors, incomplete outcome data, selective reporting, and other biases. The quality of nonrandomized clinical trials was assessed using the risk of bias in non-randomized studies of interventions (ROBINS-I), with studies of low or unclear risk of bias included in the meta-analyses. The following domains were assessed: bias due to confounding, selection bias, bias in classification of intervention, bias due to missing data, bias in measurement of the outcomes, and selective reporting. The quality of mixed-methods studies was assessed using the mixed-methods appraisal tool, with a threshold score of 50% for inclusion assessing qualitative and quantitative aspects, as well as mixed-methods. The quality of the included studies was assessed independently by the same 2 authors, with disagreements resolved by discussion.
Data items and collection
Data from the included studies were abstracted using prepiloted data collection forms. Data extracted included (1) study design and setting, (2) characteristics of participants, (3) appliance type and objective measure used, (4) treatment interventions, (5) stipulated wear time, (6) duration of the study and treatment phase, (7) objective and self-reported wear times, (8) patient awareness of the timer, (9) treatment cost, and (10) factors influencing compliance levels.
Summary measures and approach to synthesis
Clinical heterogeneity in relation to the type of orthodontic appliance or adjunct was considered prior to the comparisons. Statistical heterogeneity was investigated by examining a graphic display of the estimated compliance levels in conjunction with 95% confidence intervals (95% CI). Statistical heterogeneity was quantified using the I-squared statistic. A weighted estimate of objective compliance level for different orthodontic appliances in relation to stipulated wear time and relative to self-reported levels was calculated from the included studies. Data from qualitative studies were to be synthesized thematically, followed by integration of quantitative and qualitative results.
Risk of bias across studies
To identify publication bias, standard funnel plots and contoured enhanced funnel plots were to be drawn if sufficient numbers of studies were identified (>10 studies).
Meta-regression was undertaken to assess the relative effects of appliance type on compliance levels (extraoral vs intraoral appliances). All statistical analyses were undertaken using the Stata statistical software package (version 12.1; StataCorp, College Station, Tex).
Study selection and characteristics
The search of electronic databases yielded 4263 records; an additional 6 records were identified from hand searching. After excluding 190 duplicates, 4079 titles and abstracts were scanned to identify relevant studies by 2 authors (D.A., P.S.F.). Eighty full-text articles (74 from electronic searches) were obtained and evaluated. Fifty-six full texts were excluded ( Appendix 2 ) with 24 meeting the selection criteria; of these, 2 were randomized controlled trials, 21 were prospective cohort studies, and 1 had a mixed-methods design ( Fig 1 ). The designs, patient characteristics, appliance type, and interventions within the included studies are summarized in Table I .
|Study||Design||Participants and setting||Appliance/adjunct||Objective measure||Intervention to improve compliance||Stipulated wear time (h/d)||Study duration/treatment phase||Subjective recording of wear||Awareness of being monitored||Treatment cost|
|Clemmer and Hayes, 1979||Prospective cohort study||n = 20 (11 M, 9 F)
Mean age, 13.8 y
Age range, 11-17 y
St Louis University Department of Orthodontics, St Louis, Mo
|Cervical pull HG||Aledyne Timer, Aledyne Co., Los Altos, Calif||NA||12-14||Up to 6-9 wk||No||No||Unclear|
|Cureton et al, 1993
Cureton et al, 1993
|Prospective cohort study||n = 28 (10 M, 18 F)
Age, >10 y
US Army dental facility, Fort Knox, Ky
|Cervical pull HG||Fabricated from commercial wrist watch||
With calendar (n = 14; 7 M, 7 F)
|12||Up to 3 mo, with at least 3 mo of HG wear||Yes||No||Unclear|
|Bartsch et al, 1993||Prospective cohort study||n = 77 (40 M, 37 F)
Mean age, 10.2 ± 1.51 y
Würzburg Orthodontic Department, Würzburg, Germany
|Bionator||Fabricated from commercial wrist watch||NA||15||Mean: 3.9 mo/initial and advanced stages of treatment||No||Unclear||Unclear|
|Cole, 2002||Prospective cohort study||n = 16 (8 M, 8 F)
Mean age, 12.7 y
|Cervical pull HG||Compliance Science System, Ortho Kinetics, Vista, Calif||NA||10-12||Up to 6 mo, with at least 3 mo of HG wear||Yes||Unclear||Unclear|
|Brandão et al, 2006||Prospective cohort study||n = 21 (10 M, 11 F) ∗
Mean age, 14.8 y ∗
Age range, 11-19.5 y ∗
Graduate clinic at the Federal University, Paraná, Brazil
|HG||Compliance Science System and Affirm smart headgear modules, Ortho Kinetics, Vista, Calif||Patients were informed about the monitor of HG wear at T1||14||First 3 mo of HG wear||Yes||Patients were informed at T1||Unclear|
|Bos et al, 2007||Prospective cohort study||n = 56 (19 M, 37 F)
Mean age, 12.89 ± 2.16 y
Age range, 10-22 y
Department of Orthodontics, Academic Centre for Dentistry, Amsterdam, Netherlands
|Cervical pull HG||Thermonchron i-Button, Maxim Integrated Products, Sunnyvale, Calif||NA||12||Up to 29 d/initial and advanced stages of treatment||Yes||No||Unclear|
|Trakyali et al, 2008||Prospective cohort study||n = 30 (16 M, 14 F)||Cervical pull HG||Compliance Science System, Ortho Kinetics, Vista, Calif||Conscious hypnosis motivation group and HG timetables,
Mean age (n = 15; mean age, 10.78 ± 1.06 y)
Verbal motivation group and HG timetables, (n = 15; mean age, 10.07 ± 1.09 y)
|16||Up to 6 mo||Yes||No||State funded|
|Schott et al, 2014||Prospective cohort study||n = 28 (16 M, 12 F) ∗
Mean age, 10.6 ± 2.2 y ∗
Age range, 7.7-17.4 y ∗
|Upper removable appliance||TheraMon Sensor, Handelsagentur Gschladt, Hargelsberg, Austria or Forestadent, Pforzheim, Germany||NA||15||First 6 mo of appliance wear||No||Unclear||Unclear|
|Veeroo et al, 2014||Mixed-methods study||Semistructured interviews:
n = 14 (5 M, 9 F); age range, 15-16 y
RCT: n = 12 (5 M, 7 F), age range, 13-16 y
Orthodontic departments in Aylesbury and Amersham, United Kingdom
|Intraoral elastics||Number of used elastics||Intervention group (“if-then” planning) (n = 6)
Control group (n = 6)
|Full time (changed twice daily)||First visit after start of elastic wear||Yes (duration in h)||Unclear||Unclear|
|Hyun et al, 2015||RCT (crossover study)||n = 18 (7 M, 11 F)
Mean age, 15.44 ± 1.38 y
University of Buffalo, NY, Orthodontic Clinic
|Maxillary Hawley retainer||SMART Microsensor, Great Lakes Orthodontics, Buffalo, New York||Group 1: aware of monitoring from start of study (n = 8)
Group 2: aware of monitoring at T1 (n = 10)
|Schäfer et al, 2015||Prospective cohort study||n = 141 (88 M, 53 F)
Mean age, 10.95 ± 1.87 y
Age range, 7-15 y
Private practice and university hospital
|Maxillary removable appliance (n = 70)
Functional appliance (standard activator or Class IIII activator) (n = 71)
|TheraMon Sensor, Handelsagentur Gschladt, Hargelsberg, Austria or Forestadent, Pforzheim, Germany||NA||15||First 3 mo of appliance wear||No||Unclear||Statutory health insurance (n = 98)
Private health insurance (n = 43)
|Schott et al, 2016||Prospective cohort study||n = 109 (54 M, 55 F)
Mean age, 12.3 ± 2.9 y
Age range, 6-20 y
Private practice and university hospital
|Maxillary removable appliance (n = 33)
Functional appliance (bite-jumping appliance or standard activator) (n = 34)
Modified Hawley retainer (n = 42)
|TheraMon Sensor, Handelsagentur Gschladt, Hargelsberg, Austria or Forestadent, Pforzheim, Germany||NA||15.1 ± 0.9
15.2 ± 0.5
13.4 ± 2.7
|First follow- up appointment||Yes||Yes||Unclear|
Risk of bias within studies
The methodologic quality of the randomized controlled trials is shown in Figure 2 . The study of Hyun et al had an overall score of unclear risk of bias, due to lack of descriptions of allocation concealment and outcome assessment. The other clinical trial was considered to be at high risk of bias, because of attrition bias, selective reporting, and lack of clarity of the method used to deal with missing data. Using ROBINS-I in nonrandomized studies of interventions tool for assessing risk of bias in prospective cohort studies, 8 studies were deemed to have an overall serious risk of bias ( Appendix 3 ). Bias due to confounding was found in 5 studies, caused by failure to consider the variability of the appliances assessed. Three studies suffered from serious bias in the selection of the reported outcomes, another 2 studies had bias due to high drop-out rates, and 1 study had selection bias. One mixed-methods study scored more than 50% using the mixed-methods appraisal tool ( Fig 3 ).
Results of individual studies, meta-analysis, and additional analysis
Levels of objectively measured compliance with various removable orthodontic adjuncts
The mean duration of objectively measured wear was considerably lower than stipulated wear time among all appliance types in the included studies ( Table II ) with a mean discrepancy of 5.81 hours per day (95% CI, 4.98, 6.64) in relation to headgear based on 6 studies involving 171 participants ( Fig 4 ). In view of overlap of the samples in 3 studies, the most representative of these was selected. In relation to functional appliances, the corresponding discrepancy between verified and stipulated wear time was 5.71 hours per day (95% CI, 4.98, 6.45) ( Fig 4 ). Smaller discrepancies were found for maxillary removable appliances (3.53 hours per day; 95% CI, 2.06, 5.00) and Hawley retainers (4.58 hours per day; 95% CI, 3.21, 5.95) ( Fig 4 ).
|Appliance||Study group||Stipulated wear time (hours/day)||Objective wear time (mean in hours/day)||Actual wear time relative to stipulated wear time (%)||Self-reported wear time (mean in hours/day)||Subjective wear time relative to actual wear time (%)||Additional comments|
|Clemmer and Hayes, 1979||Cervical pull HG||–||12-14||7.43||57.2||–||–||–|
|Cureton et al, 1993||Cervical pull HG||With calendar (n = 14)
Without calendar (n = 14)
|Orthodontists overestimated HG wear by 60%
Orthodontic residents: overestimated HG wear by 71%
Dental assistants: overestimated HG wear by 73%
|Bartsch et al, 1993||Bionator||–||15||8.7||58.0||–||–||–|
|Cole, 2002||Cervical pull HG||–||10-12||6.78||52.1||8.89||131.1||–|
|Brandão et al, 2006||HG||T1: Unaware of timer (n = 21)
T2: Aware of timer (n = 21)
|14||T1: 5.6 ± 4.4
T2: 7.0 ± 5.4
|T1: 56.7 ± 45.0 ∗
T2: 62.8 ± 52.5 ∗
|13.6 ± 2.6||242.9 ± 19.1||–
|Bos et al, 2007||Cervical pull HG||–||12||5.58 ± 4.39||46.5 ± 78.7||11.02 ± 3.77||197.5 ± 34.2||Wear time estimated by orthodontists:
mean 9.52 ± 3.59 h/d
Wear time estimated by parents:
mean 11.12 ± 3.97 h/d
|Trakyali et al, 2008||Cervical pull HG||Conscious hypnosis motivation group and HG timetables (n = 15)
Verbal motivation group and HG timetables (n = 15)
|16||3 mo: 13.75 ± 5.29
6 mo: 12.13 ± 4.49
3 mo: 8.92 ± 3.41
6 mo: 9.68 ± 4.43
|3 mo: 85.9 ± 38.5
6 mo: 75.8 ± 37.0
3 mo: 55.75 ± 38.2
6 mo: 60.5 ± 45.8
|3 mo: 18.51 ± 2.96
6 mo: 17.97 ± 3.80
3 mo: 14.55 ± 3.69
6 mo: 16.29 ± 4.09
|3 mo: 134.6 ± 16.0
6 mo: 148.1 ± 21.1
3 mo: 163.1 ± 25.4
6 mo: 168.3 ± 25.1
|Schott et al, 2014||Maxillary removable appliance||–||15||T1 (55 ± 11 d): n = 28, 12.8 ± 3.8
T2 (57 ± 13 d): n = 26, 13.3 ± 3.9
T3 (58 ± 7 d): n = 13, 12.7 ± 4.8
|T1: 85.3 ± 29.7
T2: 88.7 ± 29.3
T3: 84.7 ± 37.8
|Veeroo et al, 2014||Intraoral elastics||Intervention group (“if-then” planning) (n = 6)
Control group (n = 6)
|Full time (changed twice daily)||–||77.0 †
|Median: 20.0 h/d
Median: 19.0 h/d
|Hyun et al, 2015||Maxillary Hawley retainer||Group 1: aware of monitoring from start of study (n = 8)
Group 2: aware of monitoring at T1 (n = 10)
|19||T1: 16.3 ± 4.39
T2: 15.6 ± 4.77
T1: 10.6 ± 5.36
T2: 11.1 ± 6.08
|T1: 85.8 ± 26.9
T2: 82.1 ± 30.6
T1: 55.8 ± 50.6
T2: 58.4 ± 54.8
|Schäfer et al, 2015||Functional appliance||–||15||Median: 9.5 h/d
(25th percentile: 7.4, 75th percentile: 12.0)
|Maxillary removable appliance||–||15||Median: 10.1 h/d
(25th percentile: 8.5, 75th percentile: 12.8)
|Schott et al, 2016||Maxillary removable appliance||–||15.1 ± 0.9||11.9 ± 3.0||78.8 ± 25.2||12.0 ± 3.0||100.8 ± 25.0||Wear time estimated by orthodontists:
mean 11.6 ± 2.7 h/d
|Functional appliance||15.2 ± 0.5||10.2 ± 3.3||67.1 ± 32.3||12.2 ± 3.6||119.6 ± 29.5||Wear time estimated by orthodontists:
mean 11.0 ± 3.5 h/d
|Modified Hawley retainer||13.4 ± 2.7||9.0 ± 4.9||67.2 ± 54.4||10.1 ± 3.8||112.2 ± 37.6||Wear time estimated by orthodontists:
mean 9.4 ± 3.4 h/d
A Monte Carlo permutation test was undertaken to explore the possible effect of the appliance type on compliance levels. Although compliance levels with headgear were lower than other intraoral appliances, meta-regression suggested that compliance was not directly related to the type of the appliance ( P = 0.211).
Objectively measured compliance compared with self-reported compliance with removable orthodontic adjuncts
Self-reported wear time was consistently higher than objectively measured wear time in all types of appliances ( Table II ). The mean difference between objective and self-reported compliance levels with headgear wear was calculated from 5 studies, with self-reported wear found to be 5.02 hours per day (95% CI, 3.64, 6.40) higher than the objectively measured wear ( Fig 5 ; Table II ). Overreporting was found to be somewhat lower in a study focusing on intraoral appliances ( Fig 5 ) ( Table II ). Self-reported wear time was found to approximate stipulated wear time more closely, with the difference not exceeding 3 hours ( Table II ).
Impact of interventions used to improve compliance levels
The effectiveness of the different interventions in improving compliance levels was identified in 5 studies ( Table II ). A slight increase in compliance with headgear and Hawley retainers when patients were aware of monitoring was found in 2 studies. Headgear calendars and conscious hypnosis were both shown to be effective in increasing the duration of headgear wear in single studies. Furthermore, the use of a behavioral intervention (“if-then” planning) to address the potential mismatch between intentions and actions in intraoral elastic wear did not lead to a significant improvement in compliance levels. However, although these interventions appeared to have some effect, the objectively assessed compliance level was less than that stipulated in all of these studies.
Factors affecting compliance levels and patient experiences
Several studies considered factors that may affect compliance levels including age, sex, phase of treatment, and patient motivation and attitudes ( Table III ). Younger age groups were found to be more compliant than older groups in 5 studies. However, no relationship between compliance and age was noted in 2 studies. Three studies found girls to be more compliant, but this was not corroborated in the remaining studies. Compliance level was also found to be better in the early stages of treatment in several studies. In isolated studies, patient motivation and attitudes toward orthodontic treatment were reflected in compliance levels, although limited data were found in this respect. Only 1 mixed-methods study involving semistructured interviews explored barriers to compliance with intraoral elastics ; thematic synthesis was therefore not undertaken, precluding integration of quantitative and qualitative data.