Bibliometric and visualized analysis of randomized controlled trials in orthodontics between 1991 and 2022

Introduction

In many evidence-based approaches to orthodontic research, randomized controlled trials (RCTs) represent authoritative evidence to identify rational therapeutics. This study aimed to perform mappings of bibliometric networks on orthodontic RCTs and summarize visual characteristics between 1991 and 2022.

Methods

The articles were retrieved from the Web of Science Core Collection in October 2022 without an initial time limit. Only orthodontic RCTs were eligible. Some bibliometric tools (HistCite, VOSviewer, SCImago Graphica, and CiteSpace) were applied for visualized analysis. Data such as geography, productive institutions, hot articles, journals, authors, references, and keywords were extracted and summarized for analysis.

Results

A total of 1122 orthodontic RCTs were searched. A total of 3841 authors from 1157 institutions in 65 countries published orthodontic RCTs. The United States (149) was the most prolific country, and the University of Sao Paulo (35) was the most productive institution. The American Journal of Orthodontics and Dentofacial Orthopedics (206) was the most popular journal for scholars. The visualization results of keyword co-occurrence identified 5 clusters: (1) tooth movement and auxiliary measures, (2) appliances and oral health, (3) orthodontic discomfort and symptomatic therapy, (4) periodontal disease in orthodontics and health maintenance, and (5) retention and relapse.

Conclusions

Over the past 31 years, publications and citations on orthodontic RCTs from the Web of Science Core Collection have increased notably across many countries, authors, and institutions. Recently, there has been a significant increase in the attention to orthodontic RCTs that focus on accelerating tooth movement.

Highlights

•

This study performed a bibliometric analysis of orthodontic randomized controlled trials (RCTs).
•

We aggregated a large amount of data related to orthodontic RCTs between 1991 and 2022.
•

The attention of orthodontic RCTs to accelerating tooth movement has increased significantly.
•

Topics related to accelerating tooth movement may become future hotspots.

Evidence-based orthodontics proposes using the best evidence when making clinical decisions. Levels of the evidence pyramid are ranked in descending order of internal validity (rigor or without bias) from top to bottom, with systematic reviews (SRs) and meta-analyses representing the highest level of the available evidence, followed by randomized controlled trials (RCTs). In the process of orthodontic diagnosis and treatment, professional knowledge of different orthodontists is subjective, and patients’ preferences differ. Accordingly, SRs and meta-analyses based on RCTs are considered the highest level of evidence in the assessment of the effectiveness of interventions, which assist orthodontists who lack an accurate grasp of the nuances of the guidelines. As a significant contributor to orthodontic practice in the 21st century, the development of evidence-based orthodontics is inseparable from high-quality RCTs.

RCTs use control and randomization for treatment allocation to evaluate the effects or benefits of at least 1 treatment in humans. The number of citations measures a paper’s influence in the relevant field. Highly cited RCTs assist practitioners in integrating clinical practice with clinical scientific evidence to achieve improved treatment outcomes. The study of evidence quality is always part of the development of RCTs. The modified Consolidated Standards of Reporting Trials statement is frequently used to assess the quality of reports on orthodontic RCTs. ^,

Bibliometrics is part of scientometrics and is extensively used in different academic disciplines because of its high data processing power. Bibliometrics became a medical subject heading for the first time in 1990. Bibliometric visualization can analyze abundant publications at both macro- and microlevels and return information containing the outputs, cooperation among universities or authors, past and future trends, and collaboration models of various science funding. ^, To summarize the historic evolution of a research field, especially trends and frontiers, bibliometrics has been widely applied in dentistry, such as in oral and maxillofacial surgery, endodontics, periodontology, and dental implantology. However, a comprehensive bibliometric analysis of orthodontic RCTs is still lacking.

This research adopted bibliometric visualization to analyze orthodontic RCTs. It can contribute by evaluating the current focus, hot topics, and emerging trends in orthodontic RCTs between 1991 and 2022, indicating landmark achievements and future directions.

Material and methods

To avoid bias from database updates, we collected publications from the Web of Science Core Collection (WoSCC) within 1 day and saved them as plain text. The retrieval strategy had to conform to the following criteria ( Fig 1 ):

1.

The search formula was as follows: TS=(random∗ trial OR random∗ control∗ OR random∗ clinical OR placebo∗ OR “random∗ allocat∗” OR “random∗ assign∗” OR “randomly divided“ OR RCT) AND TS=(orthodont∗).
2.

The extracted information included complete records and cited references.
3.

Orthodontic articles: Related words could be judged as orthodontics from the content. Articles were excluded if they were related to temporomandibular disorder and sleep apnea syndrome without malocclusion.
4.

RCTs performed on animals and nonrandomized trials were excluded; that is, retrospective studies, observational studies, case series, and case reports.

Relevance evaluation was performed manually by screening the title, abstract, and full text. Terminologies such as “random allocation,” “random assignment,” “randomly divided,” and similar wordings were chosen as indicators of a randomized design.

Bibliometrics can contribute to illustrating the interrelationship of different types of information by creating knowledge mappings. In this research, HistCite (Pro 2.1; Thomson Reuters, Manhattan, NY, USA), VOSviewer (version 1.6.18; created by Nees Jan van Eck and Ludo Waltman, Leiden University, Leiden, Netherlands), SCImago Graphica (version 1.0.21; SCImago Lab, Spain), and CiteSpace (version 6.1. R3; proposed by Professor Chen, Drexel University, Philadelphia, USA) were applied to visualize enormous volumes of scientific information. In addition, the output distribution and growth rate of annual publications were calculated, and regression analysis was performed.

HistCite is a scientific visualization software system that can organize information from source journal searches on the Web of Science. It allows for the use of retrospective information, especially bibliometric indicators such as authors, institutions, and journals, and generates a chronological historiography on the basis of a co-citation analysis of the literature. The 2 critical parameters of TLCS and TGCS in the results represent the total local citation score and total global citation score. TLCS indicates the entire citation frequencies of certain papers in the local data set (the dataset composed of the retrieved papers). TGCS denotes the different sum of citations of publications within Web of Science. Obviously, the number of TLCS is often far smaller than that of TGCS. TLCS reveals domain recognition and focus, signifying the total concentration of experts in this domain. TGCS can indicate public influence in the WoSCC for the all-scientific community without referring to any specific fields. ^,

VOSviewer was combined with SCImago Graphica for geography visualization to describe the distribution of cooperation among countries. Moreover, VOSviewer was used to construct clusters of complex keyword co-occurrence. The results were illustrated by label diagrams with links and nodes. CiteSpace, a Java application, was created by Professor Chaomei Chen, College of Information Science and Technology, Drexel University, Philadelphia, Pa, to visualize emerging trends and progressive knowledge domains. Three core concepts related to CiteSpace (Kleinberg’s burst detection, Freeman’s betweenness centrality metric, and heterogeneous networks) were applied to solve 3 central problems: (1) identifying research fronts, (2) labeling specialties, and (3) detecting emerging trends and abrupt points. This study investigated co-cited literature, burst analysis, cluster mappings, and timeline views. Regarding credibility and accuracy, the software provides an excellent research paradigm and is often used in internationally recognized high-impact journals. ^, ^, Reference co-citation clusters were generated using the log-likelihood ratio (LLR) strategy.

This bibliometric study only analyzed data obtained from the WoSCC. The process of using data did not involve ethical problems. Therefore, this study did not require the approval of the ethics committee.

Results

A total of 1122 publications related to orthodontic RCTs were retrieved from the WoSCC. In Figure 2 , A , curve fitting analysis shows that the number of publications on orthodontic RCTs became an overall upward trend since 1991 ( r ² = 0.9266).

According to annual outputs and growth rates, we roughly divided orthodontic RCTs between 1991 and 2022 into 2 periods ( Fig 2 ). The initial period (1991-2001) was characterized by small outputs (<10) and a significant fluctuation in the growth rates at an average rate of 4 studies per year. The next period showed an apparent growing trend, although the fluctuation of growth rates was more moderate. The outputs before 2019 (97 publications) were relatively low, with the highest output <75. The number of publications in 2021 (137 publications) was the highest.

To date, all studies have been cited 18,625 times, an average of 16.6 times per study. The TLCS and TGCS of the literature published in the initial stage were low. The TLCS showed an overall growth trend between 1991 and 2016, peaking at 237, and has since declined. Between 2019 and 2021, the TLCS dropped significantly from 112 to 46. The growth of the TGCS was more pronounced than that of the TLCS. The TGCS peaked at 1778 in 2005 and has since fluctuated.

In the past 31 years, 65 countries have published orthodontic RCTs. Regarding the number of co-authored articles among countries, 26 countries had at least 15 co-authored articles. Europe and Asia were the most prolific continents ( Fig 3 , A ). The top 10 countries generated approximately 80.4% of the overall publications ( Table I ).

Table I

The top 10 productive countries

Rank	Country	Documents, n (%)	Total link strength	TLCS	TGCS
1	United States	149 (13.3)	55	676	3904
2	United Kingdom	145 (12.9)	30	657	5200
3	Brazil	94 (8.4)	23	177	1057
4	India	93 (8.3)	4	89	719
5	Turkey	87 (7.8)	6	167	1173
6	Sweden	82 (7.3)	14	213	1563
7	Italy	67 (6.0)	21	89	1161
8	Germany	63 (5.6)	19	87	1082
9	China	63 (5.6)	11	75	680
10	Iran	59 (5.3)	17	116	613

In terms of national research strength, the United States showed the highest outputs, publishing a total of 149 (13.3%) orthodontic RCTs, followed by the United Kingdom (145 publications; 12.9%) and Brazil (94 publications; 8.4%). At the level of influence intensity, the most cited country for publications was the United Kingdom (TLCS = 657 and TGCS = 5200), followed by the United States (TLCS = 676 and TGCS = 3904) and Sweden (TLCS = 213 and TGCS = 1563). According to the beginning of countries’ cooperation ( Fig 3 , B ), the United Kingdom and the United States had the minimum average appearing year (AAY) value, whereas the AAY values of India, Saudi Arabia, Syria, Iran, Egypt, Jordan, Malaysia, and Pakistan were large. The visualized collaboration network showed that the cooperation among countries was relatively close. Regarding the cooperation intensity of the top 10 countries ( Fig 3 , C ), the United States had the most cooperation among participating countries, followed by the United Kingdom.

A total of 3841 authors from 1157 institutions with publications on orthodontic RCTs were retrieved. As shown in Table II , Hajeer was the most productive author. The top ten authors published 112 studies, accounting for 10.0%. Compared with the other top 10 authors, the Hirsch index of Hajeer was lower, not more than 20. Figure 4 illustrates a certain degree of collaboration. The betweenness centrality measures the proportion of the shortest paths in the network to which a prescribed node belongs. Moreover, no authors’ betweenness centrality was >0.1.

Table II

The top 10 productive authors (author statistics only included the first author and the corresponding author)

Author	Institution	Country	Documents	Hirsch index
Hajeer MY	University of Damascus	Syria	27	16
Darendeliler MA	University of Sydney	Australia	12	40
Ireland AJ	Royal United Hospital	United Kingdom	12	28
Buschang PH	Texas A&M University System	United States	10	50
Alam MK	Jouf University	Saudi Arabia	9	22
Cobourne MT	King’s College London	United Kingdom	9	36
Scribante A	University of Pavia	Italy	9	33
Benson PE	University of Sheffield	United Kingdom	8	24
Fleming PS	Trinity College Dublin	Ireland	8	34
O’Brien KD	University of Manchester	United Kingdom	8	33

The top 10 most productive institutions are shown in Table III ; 3 were located in Sweden, 2 in the United Kingdom, 2 in Switzerland, and 1 each in Brazil, Syria, and Australia. The University of Sao Paulo (35 publications) was the most productive, followed by the University of Damascus (27 publications), the University of Malmo (22 publications), and the University of Manchester (22 publications). The TGCS of the University of Manchester in the United Kingdom (cited 1997 times) was the highest and was significantly higher than that of any other institution. Figure 5 shows the global distribution of cooperation among institutions, which can provide a basis for seeking cooperative institutions. With respect to centrality, the University of Zurich ranked first with 0.16, followed by the University of Manchester (0.13) and the University of Sao Paulo (0.10).

Table III

The top 10 productive institutions

Rank	Institution	Country	Documents	Centrality	TLCS	TGCS
1	University of Sao Paulo	Brazil	35	0.10	74	481
2	University of Damascus	Syria	27	0.05	34	312
3	University of Malmo	Sweden	22	0.02	62	345
4	University of Manchester	United Kingdom	22	0.13	69	1997
5	University of Sydney	Australia	21	0.09	44	451
6	King’s College London	United Kingdom	20	0.05	81	498
7	Karolinska Institutet	Sweden	18	0.08	41	358
8	University of Zurich	Switzerland	17	0.16	48	297
9	University of Bern	Switzerland	16	0.06	12	164
10	University of Gothenburg	Sweden	16	0.09	45	336