Background
Getting effective healthcare innovations into practice is often suboptimal [
1‐
3]. This implementation issue also applies to orthodontic mini implants (OMIs) because surveys worldwide have shown that many clinicians rarely or never use these devices [
4‐
8] notwithstanding their promising success rates, effectiveness, and applicability [
9‐
11]. To understand the causes of this problem, it is important to identify the barriers and facilitators to the implementation of OMIs in clinical practice. This systematic review identified and quantified these implementation constructs.
Most orthodontic treatment plans need some form of anchorage to counteract the reciprocal forces of orthodontic tooth movement [
12]. Numerous anchorage systems have been developed for this purpose. They generally apply forces to groups of teeth or use extra-oral traction to the neck or cranium. These techniques are effective, but they can still cause a loss of anchorage, have a limited area of application, and often depend on the constant collaboration of the patient [
12]. OMIs are not limited by most of these characteristics; they can be implemented in a wide variety of orthodontic treatment plans and can be used in the maxilla and mandible for long periods of time [
10]. The most frequently used OMIs are machine surfaced bone screws with a diameter of 1.3–2 mm and a length of 6–10 mm [
11]. Both single and multiple OMIs with or without connecting plates are used for anchorage purposes. After insertion, OMIs are usually loaded immediately with orthodontic forces, and they are removed after the completion of the orthodontic treatment objectives.
Since the introduction of OMIs in 1997 by Kanomi [
13], the number of publications on these devices has increased exponentially [
14] and systematic reviews on OMIs have recorded promising low implant failure rates [
11,
15] and favorable effectiveness [
9,
10]. Numerous orthodontic companies have been founded and presentations of orthodontic treatments with OMIs have become the norm at orthodontic meetings [
16]. International orthodontic conferences are even organized that focus exclusively on treatment with OMIs [
17].
However, surveys in the USA, India, Germany, and England showed that many orthodontists never or rarely use these devices [
4‐
8,
18‐
20]. This knowledge-to-action (KTA) gap, which is the gap between evidence-based knowledge and the use of this information in practice, was unexpected for a technique that has been introduced almost 20 years ago and has received such high acclaims [
3,
9‐
11,
13]. This KTA gap was also unexpected in the context that most orthodontic treatment plans need some form of anchorage [
12]. In addition, more than 75 % of 108 surveyed doctors included OMIs in their treatment plan for a common orthodontic patient [
21].
KTA gaps are not just limited to interventional procedures with OMIs but are common for a variety of medical conditions and are a global problem [
22,
23]. Only a small fraction of healthcare innovations gets incorporated into practice, and it has been estimated that 45 % of patients in the USA do not receive recommended care [
23,
24]. In recent years, interest in the causes of KTA gaps and strategies for dealing with them has increased dramatically [
25]. Identifying and quantifying implementation constructs of healthcare interventions is an important step to address these issues.
We therefore developed a systematic review that identified and quantified barriers and facilitators to the implementation of OMIs in clinical practice. We applied this primary research objective to both demand-side stakeholders, i.e., orthodontic patients and their family members, and potential supply-side stakeholders, e.g., clinicians, office staff, clinic owners, researchers, guideline developers, policy makers, and implant companies. Barriers to the implementation of OMIs could refer to the invasiveness of the interventional procedure, learning a new technique, the adverse effects of interventions and the fear of complications, financial barriers, the large volume of research evidence, the lack of trust in research evidence, the applicability of the new health technology to a local context, etc. [
5,
8,
26‐
28]. Facilitators to implementation of OMIs could refer to shorter treatment time, better outcomes, improved esthetics during orthodontic treatment, recommendations by patients who had undergone treatment with OMIs, etc. [
5,
29,
30]. Research data on the identified and quantified implementation constructs can be subsequently consulted to develop tailored strategies to deal with them.
For our secondary objectives, we recorded the prevalence of clinicians that do not use OMIs in clinical practice. This statistic was extracted from the studies that were identified as “eligible” for our primary research question. During our scoping searches, we identified no systematic reviews that addressed our primary and secondary research questions on the implementation of OMIs.
Objectives
The following objectives of this systematic review were defined:
Primary objective
-
To identify and quantify barriers and facilitators to the implementation of OMIs for all potential stakeholders such as patients and their family members, clinicians, office staff, clinic owners, researchers, guideline developers, policy makers, and implant companies.
Methods
We applied the methods that were described previously in our published protocol [
14]. Various sources were consulted to develop these methods: (1) established conceptual models for assessing barriers and facilitators to the implementation of knowledge use [
22,
25,
26,
31,
32]; (2) guidelines and handbooks for designing and reporting qualitative and quantitative systematic reviews [
33‐
38]; (3) guidelines and checklists for reporting research studies [
39]; (4) systematic reviews that were specifically designed to identify implementation constructs and addressed similar research objectives as our research study [
27,
28,
40‐
44]; and (5) systematic reviews and their protocols on OMIs that we have published previously [
14,
45‐
48].
The characteristics and application of these sources are further explained in the pertinent sections of this systematic review. Differences between our protocol and the final systematic review are reported in the “
Results” section together with the rationale and the consequences of these modifications. We adopted the PRISMA 2009 Statement for reporting systematic reviews and presented this manuscript according to the order of this guideline (Additional file
1)[
36,
49]. This systematic review was not registered in the PROSPERO database because our research questions are not covered by the inclusion criteria of this register [
50].
Eligibility criteria
We applied the eligibility criteria that were outlined in our published protocol [
14]. These criteria are summarized under here.
Study designs
For the primary objectives, we defined the following eligibility criteria:
-
Studies that collected original data on identified barriers and facilitators to the implementation of OMIs in clinical practice were eligible. To avoid exclusion of pertinent studies, we did not set eligibility criteria for specific research designs. We expected to find interviews, focus groups, surveys, and questionnaires as the most common eligible research designs.
-
Studies that addressed our primary research question as their primary or secondary objectives were eligible. For example, studies that identified barriers and facilitators to the implementation of OMIs in clinical practice as a part of a larger mixed methods model were eligible.
For the secondary objectives, we defined the following eligibility criteria:
-
Only quantitative studies, for example, surveys, that addressed the primary objectives of this systematic review were eligible.
Stakeholders (participants)
We applied broad-spectrum selection criteria that included all potential demand- and supply-side stakeholders. The former category refers to patients undergoing the interventional procedure and the pertinent family members. The latter category refers to the following stakeholders: clinicians, clinic owners, researchers, office staff, guideline developers, policy makers, and implant companies.
Interventions
Interventions that used single or multiple implants with diameters smaller than 2.5 mm for orthodontic anchorage objectives were eligible. We applied no restrictions to the length or design of the implant, its connection to plates, its location of insertion, the type of method for implant insertion, the type of orthodontic loading, and the type of implant maintenance after its insertion. Interventions with OMIs on patients of any age, sex, or demographics were eligible.
Outcomes
Any type of barrier or facilitator to the use of OMIs in clinical practice was selected as our primary outcome. A barrier was defined as any variable that impedes or obstructs their use. A facilitator was defined as any variable that eases and promotes the use of OMIs. Barriers and facilitators were eligible when they were described as implementation constructs by the eligible stakeholders [
44]. For example, patient’s perceptions of the interventional procedure or assessments of health experiences such as pain and discomfort during implant insertion were not considered as eligible outcomes when they were not specifically defined as barriers to the use of OMIs by these patients. These eligibility criteria avoid mislabelling of implementation constructs during qualitative analyses as a result of bias or misinterpretation of outcomes by systematic reviewers.
Setting and language
No setting and language restrictions were applied.
Information sources were searched from 1 January 1997, the year of the first publication on OMIs, until 15 January 2016 [
13].
Methods to find pertinent subject headings and key words were adopted from our previous systematic reviews and protocols on OMIs [
14,
45‐
48]. An information specialist (NR) in computerized searches of healthcare publications assisted with the development of the search strategies. In our protocol, we presented the full electronic search strategies for both PubMed and Google Scholar Beta [
14]. For PubMed, we used the following search strategy: (orthodontics OR orthodontic OR orthodontist OR orthodontists) (implant OR implants OR “mini implant” OR “mini implants” OR screw OR screws OR “mini screw” OR “mini screws” OR “miniscrew” OR “miniscrews” OR “microscrew” OR “temporary anchorage device”).
Study selection
Eligible studies for this systematic review were selected independently by three review authors (RMR, LR, and LL). Disagreements between these operators on the eligibility of an article were resolved through rereading of the pertinent article, discussions, and if necessary through contacting its authors [
49]. In the case of persistent disagreements, a fourth author (NDG) was asked to arbiter.
The selection of studies was summarized in a PRISMA flow diagram [
36,
49]. Excluded articles were listed in a table with the reasons for their exclusion. A detailed description of the procedures for selecting studies, contacting authors, and for assessing multiple publications of the same research data and dealing with these issues were presented in our published protocol [
14]. All study selection procedures were conducted according to this protocol.
Data collection process and data items
Prior to the formal study selection and data extraction process, a list of “potential” barriers and facilitators to the implementation of OMIs was developed by the three reviewers (RMR, LR, and LL). This list was developed through the assessment and discussions of three groups of publications: (1) systematic reviews that focused on the identification of barriers and facilitators to the implementation of health-related issues and technologies [
27,
28,
40‐
44]; (2) conceptual models for assessing barriers and facilitators to knowledge use [
25,
26,
31,
32,
58]; and (3) our previous systematic reviews and protocols on OMIs [
14,
45‐
48]. This list of “potential” barriers and facilitators was created prior to the study selection and data extraction process and was used to calibrate the three reviewers and to increase their background knowledge on implementation constructs. Our list of “potential” implementation constructs was not used as a reference checklist during the study selection and data collection process because this could have resulted in the inappropriate exclusion of “unexpected” barriers and facilitators to the implementation of OMIs in clinical practice.
For the development of the data collection forms, we explored (1) the reporting checklists of pertinent research designs of the Equator network [
39]; (2) the data extraction forms of previous systematic reviews and protocols on OMIs [
14,
45‐
48]; and (3) the three groups of publications that were used to develop the list of potential implementation constructs [
25‐
28,
31,
32,
40‐
48,
58]. Pertinent items for the extraction of data for the secondary research question were also assessed during this research phase. Extracted data items for our primary and secondary objectives include: the source, eligibility, duplicate publication, the study design, selection procedures, stakeholders, the setting, interventions, outcomes, flow and timing, adverse effects, withdrawals and missing outcomes, the funding, and miscellaneous data of the selected studies [
36,
38]. Many of these items were further subdivided and all extracted entries were listed in tables. Descriptions of each item were presented in these tables and entries that could bias the outcomes were also recorded.
All data extraction procedures were conducted independently by the three aforementioned operators, who are experienced systematic reviewers and topic experts. Disagreements on extracted items were resolved through rereading and discussions and if necessary an arbitrator (NDG) was consulted to adjudicate these disagreements. Our pilot-tested data extraction forms and a detailed description of our data collection procedures were given in the published protocol of this systematic review [
14].
Outcomes and prioritization
Primary outcomes
-
The primary outcomes were all barriers and facilitators to the implementation of OMIs in clinical practice identified by all demand-and supply-side stakeholders. When given, we recorded also the prevalence of implementation constructs among subgroups of pertinent stakeholders.
-
Pertinent stakeholders were defined in Table
1 and were divided in “users” and “non-users.” Potential subgroups for which outcomes were recorded are presented in Table
1.
Table 1
Potential subgroups for which outcomes were recorded
Research design | Surveys: outcomes obtained through surveys (questionnaires) of stakeholders Interviews: outcomes obtained from interviews with stakeholders Focus groups: outcomes obtained from focus groups with stakeholders |
Stakeholders | Demand-side stakeholders: orthodontic patients and their family members Supply-side stakeholders: clinicians, office staff, clinic owners, researchers, guideline developers, policy makers, implant companies, etc. Users/non-users: both demand- and supply-side stakeholders can be further subdivided in those that have used OMIs previously (users) and those that have never used these devices (non-users) |
Interventions | Specified interventions: these interventions refer to a specific phase or type of the interventional procedure. Phases of the intervention refer to the anesthetics, implant insertion, orthodontic treatment with OMIs, implant removal, or the healing phase. Types of interventions refer to the implant type and dimensions, number of implants, use of plates, the surgical procedure, implant location, timing and forces of orthodontic loading, etc. [ 48]. “Non specified” interventions: these interventions refer to “any orthodontic treatment with OMIs.” Additional information on the specific phase or type of the interventional procedure is not provided |
Time points | Pre-intervention recordings, i.e., recordings of outcomes prior to the interventional procedure Immediate post-intervention recordings, i.e., recordings of outcomes within 2 weeks after the completion of the interventional procedure Long-term post-intervention recordings, i.e., recordings of outcomes more than 2 weeks after the completion of the interventional procedure |
Setting/country | Private practice: stakeholders treated or working in a private practice University setting: stakeholders treated or working in a university clinic Country: stakeholders treated or working in a specific country |
-
A barrier is defined as any variable that impedes or obstructs the use of OMIs. A facilitator is defined as any variable that eases and promotes their use.
-
The pre-intervention recording of barriers or facilitators to a “specified” intervention with OMIs for “non-user” clinicians or for patients that have not undergone this intervention previously in any type of setting or research design was our “preferred” primary outcome.
-
The prevalence of identified barriers and facilitators among the surveyed or interviewed pertinent stakeholders was calculated as follows:
$$ \begin{array}{l}\mathrm{Prevalence}\ \mathrm{o}\mathrm{f}\ \mathrm{a}\mathrm{n}\ \mathrm{identified}\ \mathrm{barrier}\ \mathrm{o}\mathrm{r}\ \mathrm{f}\mathrm{a}\mathrm{cilitator} = \mathrm{t}\mathrm{he}\ \mathrm{n}\mathrm{umber}\ \mathrm{o}\mathrm{f}\ \mathrm{s}\mathrm{t}\mathrm{a}\mathrm{keholders}\ \mathrm{t}\mathrm{hat}\ \mathrm{s}\mathrm{cored}\\ {}\mathrm{a}\ \mathrm{particular}\ \mathrm{construct}\ \mathrm{a}\mathrm{s}\ \mathrm{a}\ \mathrm{barrier}\ \mathrm{o}\mathrm{r}\ \mathrm{f}\mathrm{a}\mathrm{cilitator}\ \mathrm{t}\mathrm{o}\ \mathrm{t}\mathrm{he}\ \mathrm{implementation}\ \mathrm{o}\mathrm{f}\ \mathrm{OMIs}\ \mathrm{in}\ \mathrm{clinical}\\ {}\mathrm{practice}/\mathrm{the}\ \mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}\ \mathrm{n}\mathrm{umber}\ \mathrm{o}\mathrm{f}\ \mathrm{s}\mathrm{t}\mathrm{a}\mathrm{keholders}\ \mathrm{t}\mathrm{hat}\ \mathrm{s}\mathrm{cored}\ \mathrm{o}\mathrm{n}\ \mathrm{t}\mathrm{he}\ \mathrm{r}\mathrm{o}\mathrm{l}\mathrm{e}\ \mathrm{o}\mathrm{f}\ \mathrm{t}\mathrm{his}\ \mathrm{particular}\ \mathrm{construct}\\ {}\mathrm{a}\mathrm{s}\ \mathrm{a}\ \mathrm{barrier}\ \mathrm{o}\mathrm{r}\ \mathrm{f}\mathrm{a}\mathrm{cilitator}\ \mathrm{t}\mathrm{o}\ \mathrm{t}\mathrm{he}\ \mathrm{implementation}\ \mathrm{o}\mathrm{f}\ \mathrm{OMIs}\ \mathrm{in}\ \mathrm{clinical}\ \mathrm{practice}.\end{array} $$
This prevalence was presented for example as 30/50.
-
In our published protocol, we presented additional information on (1) defining primary outcomes; (2) the procedures to extract and categorize primary outcomes; and (3) anticipated exemplary tables of categorized implementation constructs [
14,
38].
Secondary outcomes
-
The secondary outcome was the prevalence of clinicians that do not use OMIs and represents the knowledge-to-action gap. This statistic was calculated as follows:
$$ \begin{array}{l}\mathrm{The}\ \mathrm{prevalence}\ \mathrm{of}\ \mathrm{clinicians}\ \mathrm{that}\ \mathrm{do}\ \mathrm{not}\ \mathrm{use}\ \mathrm{OMIs} = \mathrm{the}\ \mathrm{number}\ \mathrm{of}\ \mathrm{clinicians}\ \mathrm{that}\ \mathrm{do}\ \mathrm{not}\ \mathrm{use}\\ {}\mathrm{OMIs}/\mathrm{The}\ \mathrm{total}\ \mathrm{number}\ \mathrm{of}\ \mathrm{surveyed}\ \mathrm{clinicians}\ \mathrm{that}\ \mathrm{reported}\ \mathrm{on}\ \mathrm{the}\ \mathrm{use}\ \mathrm{of}\ \mathrm{OMIs}\ \mathrm{in}\ \mathrm{clinical}\ \mathrm{practice.}\end{array} $$
Information that could give further insights in the understanding of the knowledge-to-action gap, e.g., the number of implants placed per clinician per year, was also recorded.
Risk of bias in individual studies
According to our protocol [
14], we applied critical appraisal instruments that were specific for the type of research design used in the eligible studies. In this review, we adopted The Joanna Briggs Institute critical appraisal tool for quantitative studies that report prevalence and incidence data [
35,
42,
59]. This instrument has been specifically developed for questionnaires and surveys. Differences between reviewers in scoring these tools were resolved through discussions. A fourth reviewer (NDG) was called upon in the case of disagreement between reviewers. Authors of eligible studies were contacted in the case of persistent disagreements on appraisal scores.
The critical appraisal scores for each selected study were listed in tables and for each appraisal tool separately [
34,
35]. We calculated the prevalence of “yes” scores (number of “yes”/number of articles) for each individual appraisal question [
42]. No attempts were made to calculate overall appraisal scores. The potential influence of each of the scored answers on the outcomes of each selected study was weighted during the data synthesis and was used to assess the overall strength of evidence of the review (see “Confidence in cumulative evidence”) [
38]. Additional details on the procedures for the assessment of risk of bias and the instruments for assessing the methodological quality of studies were presented in our published protocol [
14].
Synthesis of results
Criteria for a quantitative synthesis
We only conducted meta-analyses for our primary and secondary outcomes when (1) the risk of bias in the eligible studies was low; (2) outcomes between studies were consistent; (3) publication bias was low; (4) a high number of studies was included; and (5) heterogeneity was low [
60‐
62]. Forest plots were used to display the dispersion of the even rates of both primary and secondary outcomes. Comprehensive meta-analysis (CMA) software was used to conduct all statistical analyses in this systematic review [
63,
64].
Unit-of-analysis issues and missing data
To deal with unit-of-analysis issues, we assessed whether all participants underwent the same intervention, multiple interventions, and whether outcomes were recorded at different or multiple time points [
61]. Our strategies for dealing with missing data were presented in our published protocol [
14].
Qualitative synthesis
According to the PRISMA-P 2015 statement, we undertook a systematic narrative (qualitative) synthesis even when criteria for conducting quantitative syntheses were fulfilled [
38]. Our narrative synthesis was conducted systematically and transparently to reduce the potential for bias [
65]. We refrained from vote counting, i.e., counting those studies that yielded a significant result and those that did not [
66,
67]. As suggested by the PRISMA-P 2015 statement [
38], we adopted the “established methods” for conducting systematic narrative syntheses according to the guidance of the Centre for Reviews and Dissemination (CRD) [
65]. The CRD framework for conducting such a synthesis consists of four phases: (1) developing a theory why and how each barrier or facilitator could affect the implementation of OMIs for each linked stakeholder; (2) developing an initial synthesis of the findings of the eligible studies; (3) exploring relationships within and between studies; and (4) assessing the robustness of the synthesized evidence [
65]. These steps do not have to be conducted exactly according to the order of this framework and were conducted iteratively by the three topic experts (RMR, LR, LL) [
65]. Disagreements were resolved through discussions, and persistent disagreements were resolved through the arbitrage of a fourth author (NDG) or through contacting pertinent authors.
Risk of bias across studies
Meta-bias refers to the biased selection of research data and covers both reporting bias (selective outcome reporting) and publication bias [
38]. Methods to assess the presence and impact of both biases and strategies for dealing with them were described in detail in our published protocol [
14,
54,
68‐
70].
For the assessment of the strength of the body of evidence, we consulted the guidelines described by the GRADE approach [
71]. The robustness of the synthesized evidence depends on (1) the number and size of the eligible studies; (2) within and between study diversity; (3) risk of bias assessments (magnitude and direction); (4) the consistency of the outcomes between studies; (5) the magnitude of the outcomes; and (6) the presence of publication bias. To assess the robustness of identified evidence we (1) weighed the role of these variables; (2) revisited the data collection forms and the critical appraisal tools to assess whether items have been overlooked; and (3) contacted authors to obtain additional information. We did not score the “levels of evidence” according to the GRADE approach. Our research questions do not qualify for this approach because they do not address questions about interventions, management strategies, or policies [
72].
Additional analyses
Investigation of heterogeneity
We considered three sources of heterogeneity: methodological, clinical, and other sources of heterogeneity [
61,
73]. These sources were selected a priori based on information from previous systematic reviews on this research topic and through discussions between the reviewers [
45‐
48,
74]. Our “a priori” defined potential sources of heterogeneity were listed in our published protocol [
14]. The type of stakeholders, i.e., patients, clinicians, and office staff, was excluded as a source of diversity because outcomes were analyzed separately for each type of stakeholder. We reported when “post hoc” defined sources of heterogeneity were investigated.
The presence of statistical heterogeneity was investigated by calculating Cochran’s Q, the degrees of freedom based on the number of eligible studies, and the pertinent
p value [
64,
75‐
77]. We also calculated the following statistics: Kendall’s
τ2,
τ, and
I2 [
64,
75,
77‐
80]. These calculations, their use, and strategies for dealing with heterogeneity were explained in our published protocol [
14,
61].
Our protocol described our planned methods for conducting subgroup analyses, meta-regressions, and sensitivity analyses [
14]. Criteria, rationales, and caveats for undertaking such research procedures were also outlined in this protocol [
61,
81,
82].
Authors of pertinent primary research studies were contacted to obtain additional information on (1) the eligibility of specific research studies and (2) unclear or missing data in primary research studies. The methods for this research procedure were described in detail in our published protocol [
14].
Differences between the protocol and the review
We reported all changes in the methods during the conduct of this research study compared with those planned in the protocol. We described the type, timing, and the rationale of each of these modifications. We also reported the consequences of these changes on the direction, the magnitude, and the validity of the outcomes [
83].
Conclusions
This is the first systematic review that addressed implementation issues of OMI in clinical practice. The three eligible surveys [
5,
8,
85] identified and quantified 17 implementation constructs. Three facilitators and one barrier (costs) were identified by more than 60 % of the orthodontic patients [
85]. Lack of training, fear of risk factors, (almost) no suitable indications, and skepticism about additional benefit of OMIs were identified as barriers to the implementation of OMIs by more than 50 % of orthodontic clinicians [
5,
8].
The main limitations of these studies included (1) the small number of studies; (2) not defining the research questions, i.e., the primary outcomes; (3) the research design (surveys) of the studies and the exclusive use of closed-ended questions; (4) not consulting standards for identifying implementation constructs; (5) the lack of pilot testing; (6) high heterogeneity; (7) the risk of reporting bias; and (8) additional shortcomings. This leaves much space for the exploration of additional constructs through open-ended questions in other research designs, such as in-depth interviews or focus groups. The severe underuse of OMIs that was identified in the selected studies demonstrated the need to identify and quantify such constructs and to develop strategies to deal with them.
Most primary studies on OMIs have addressed variables associated with their stability and effectiveness and few studies have assessed factors associated with their implementation. This study showed the need to change course. Before undertaking new research studies on OMIs, it will be necessary to consult the findings of this systematic review and other reviews and convene a variety of stakeholders, such as patients, clinicians, researchers, government bodies, guideline developers, and implant companies, to develop priority questions in this field of research [
109]. Prioritizing such questions could redirect research on OMIs towards studies on their implementation. This could reduce additional research waste and benefit patients, clinicians, researchers, policymakers, insurance companies, implant companies, and research sponsors [
108].