Differences in productivity and collaboration patterns on spine-related research between neurosurgeons and orthopedic spine surgeons: Bibliometric analysis

Abstract

Background:

Spinal surgeries are commonly performed by neurosurgeons and orthopedic spine surgeons, with many spine-related articles published by them. However, there has been limited research that directly compares their research achievements. This article conducted a comparative analysis of spine-related research achievements between neurosurgeons and orthopedic spine surgeons. This study examines differences in productivity and impact on spine-related research between them using these measures, particularly with a novel clustering algorithm.

Methods:

We gathered 2148 articles written by neurosurgeons and orthopedic spine surgeons from the Web of Science core collections, covering the period from 2013 to 2022. To analyze author collaborations, we employed the follower-leader clustering algorithm (FLCA) and conducted cluster analysis. A 3-part analysis was carried out: cluster analysis of author collaborations; mean citation analysis; and a category, journal, authorship, L-index (CJAL) score based on article category, journal impact factors, authorships, and L-indices. We then utilized R to create visual displays of our findings, including circle bar charts, heatmaps with dendrograms, 4-quadrant radar plots, and forest plots. The mean citations and CJAL scores were compared between neurosurgeons and orthopedic spine surgeons.

Results:

When considering first and corresponding authors, orthopedics authors wrote a greater proportion of the articles in the article collections, accounting for 75% (1600 out of 2148). The CJAL score based on the top 10 units each also favored orthopedic spine surgeons, with 71% (3626 out of 6139) of the total score attributed to them. Using the FLCA, we observed that orthopedic spine surgeons tended to have more collaborations across countries. Additionally, while citation per article favored orthopedic spine surgeons with standard mean difference (= −0.66) and 95%CI: −0.76, −0.56, the mean CJAL score in difference (= 0.34) favored neurosurgeons with 95%CI: 0.24 0.44.

Conclusion:

Orthopedic spine surgeons have a higher number of publications, citations, and CJAL scores in spine research than those in neurosurgeons. Orthopedic spine surgeons tend to have more collaborations and coauthored papers in the field. The study highlights the differences in research productivity and collaboration patterns between the 2 authors in spine research and sheds light on potential contributing factors. The study recommends the use of FLCA for future bibliographical studies.

Key points:

1. Orthopedic spine surgeons have a higher publication count, number of citations, and CJAL scores in the domain of spine research compared to their counterparts in neurology. This indicates a higher level of research productivity and impact from the orthopedics authors.

2. The use of the follower-leader clustering algorithm (FLCA) revealed that orthopedics authors tend to collaborate more extensively across countries. This increased level of collaboration is another indicator of their influence and reach in the field of spinal research.

3. Despite the above trends, the mean CJAL score, which measures research achievement based on article category, journal impact factors, authorships, and L-indices, actually favored nurosurgeons. This suggests that nurosurgeons may have fewer but more impactful publications in the field of spine research.

1. Introduction

Spinal surgery is a discipline that falls under 2 surgical specialties: neurological surgery and orthopedic surgery.^[1] Previous studies have compared outcomes between neurosurgeons and orthopedic surgeons across various areas, such as early outcomes and complications of pedicle subtraction osteotomies,^[2] representation of women on editorial boards of spine-related journals,^[3] resident training courses in adult spine surgery,^[4] workforce involvement,^[5] competency for common spinal conditions,^[6] differences in classifying cervical dislocation injuries and making assessment and treatment decisions,^[7] and research achievements (RAs).^[1,8] However, there has been little research conducted on author collaboration in this field, and the definition of author collaboration (AC)^[9] based on clustering remains unclear. This needs to be explored in more detail using a reliable clustering algorithm.

1.1. Issues of author collaborations

ACs in bibliometrics can be analyzed using various metrics, such as coauthorship analysis, network analysis, and bibliographic coupling.^[10–12] Coauthorship analysis examines the collaboration between authors, while network analysis visualizes the relationships between authors in a network graph. Bibliographic coupling looks at the shared references between authors to identify their common research interests.^[13,14]

While cluster analysis has become a popular technique in bibliometric studies to identify and visualize research areas and collaborations, there are some potential drawbacks to this approach.^[15] One major limitation is the potential for bias in the selection of clustering algorithms and parameters used. Different algorithms and parameters can yield different results, which may affect the interpretation of the research areas and collaborations. Therefore, it is important to carefully consider the choice of clustering algorithm and parameters and to test the robustness of the results.

Another potential drawback in bibliometrics is the reliance on visual displays yielded by software, which may not fully capture the complexity and nuance of search areas and collaborations.

Readers were not provided with a clear definition of ACs if they did not describe AC relationships based on their similarity using correlation distance, Euclidean distance, K-means clustering, hierarchical clustering, density-based clustering, or other spectral methods (e.g., the eigenvalues and eigenvectors of a similarity matrix to partition the data points into clusters). In tradition, cluster analysis does not provide a definitive answer about the boundaries and relationships between research areas, and further analysis and interpretation may be needed to fully understand the results.^[16]

One of the research questions pertains to the clustering algorithm that was created and developed for implementation in ACs. This algorithm (namely, following leader cluster algorithm [FLCA] denoted by follower-leader clustering^[17–19]) relies on the principal link between followers and leaders, such as the cluster head.

1.2. RAs between neurosurgeons and orthopedic surgeons

Studies^[1,8] have presented a comparison of RAs between neurosurgeons and orthopedic surgeons utilizing the h-index.^[20] However, using the h-index has some limitations, including: coauthors receiving equal credit in an article^[21]; not taking into consideration journal impact factors and document types^[22]; and difficulty in distinguishing RAs within and between groups due to the use of an integer value.^[23]

To compare the RAs of neurosurgeons and orthopedic surgeons, a category, journal, authorship, L-index (CJAL) score^[11,24] that considers article category, journal impact factors, authorships, and L-indices (on article citation)^[25] is necessary. The second research question focuses on this RA comparison, which was based on the total CJAL scores and mean citations.

1.3. Study goals

The inquiry into whether there exist noteworthy distinctions in the research accomplishments between neurosurgeons and orthopedic surgeons in the realm of spinal surgery is a thought-provoking one. Despite both specialties concentrating on the spine, they have distinct methodologies and educational backgrounds that could potentially affect their research productivity and author collaborations.

The focus of this study is to compare the research productivity and influence of spine-related publications authored by neurosurgeons and orthopedic spine surgeons utilizing various metrics, including the FLCA algorithm used for cluster analysis.

2. Methods

2.1. Data source

We gathered a total of 2148 articles related to spines using specific keywords in the Web of Science Core Collection database. The search terms used were “adult spinal deformity, “sagittal alignment, “spinopelvic alignment, “kyphoscoliosis, “sagittal plane deformity, “sagittal imbalance, and “degenerative scoliosis.”^[26] The articles were collected between 2013 and 2022, and we excluded first and corresponding authors who were not affiliated with identical neurosurgery or orthopedic departments. Only first and corresponding authors from identical departments were included for comparison purposes.

We obtained a comprehensive range of metadata, including author names, research institutes, departments, countries of origin, and other relevant information. A complete list of metadata is available in Supplemental Digital Content 1, http://links.lww.com/MD/K305.

All data presented in this study are publicly available on Web of Science Core Collection, and no participant identification information was collected or disclosed. As such, ethical approval was waived for this study.

2.2. Three major parts in this section

2.2.1. Develop the FLA for AC cluster analysis.

In social network analysis,^[10,27] data is often divided into 2 parts: the main data and their dyadic relationships (as shown in Table 1, represented by data A and data B). Firstly, the network association frequencies are arranged from highest to lowest, and then each keyword (or author entity) is assigned a temporary serial number (or label) from 1 to n, where n is the number of entities. The score of the association frequency determines the size of the bubble, and the weights of the dyadic relationships are summed up for the main data.^[28] For example, A-B 2 (indicating that A and B are coauthors of 2 papers), both A and B are assigned a weight of 2. A-A 2 (indicating that A is the sole author of 2 papers), A is assigned a weight of 2.

Table 1.

The following leader cluster algorithm (FLCA) used in this study (Searching leaders).

Input k, Sorting dataset A (name, connections, cluster#)
And B(couple names and connections) by
descending order and cluster#(1至n) assigned each
Output for temporary clusters
1	Followers search leaders from lower to higher in connections by observing the maximum as the leader
2	For jk = n To 1 Step -1 for DatasetA(size = n)
3	Entity = dataset A. Cells(jk, 1)
4	For j = 1 To m datasetB(size = m)
5	Name 1 = dataset B. Cells(j, 4)
6	Name 2 = dataset B. Cells(j, 5)
7	connections = dataset B. Cells(j, 6)
8	If Entity = Name1 Or Entity = Name 2
	And Name 1 <>Name 2 Then
9	If Entity = Name 1Then
10	Leader = Name 2
11	Else
12	Leader = Name 1
13	End If
14	For jk2 = 1 To jk—1
	(Counts for leader > counts for follower)
15	□□If datasetA. Cells(jk2, 1) = leader Then
16	□Dataset A. Cells(jk, 3) = jk2
17	□(leader found)
18	□□□□□□Goto 22 end the loopjk2及j
19	□□□End If
20	□□□Next jk2□end dataset A
21	□Next J□□end dataset B(size = m)
22	Next jk□□end dataset A(size = n)

Table 1 illustrates the algorithm for finding leaders. Firstly, for the main data (such as countries in data A), their leaders are traced from smallest to largest (as in steps 3–23). Except for the entity with the maximum count (e.g., country), all others are followers. When each follower finds a unique leader (the leader with the highest number of links, and in case of ties, the leader with the higher total number of links), the follower is assigned the cluster label of that leader (as in step 17). If a follower is not associated with any other entity, it becomes an isolated entity, forming its own cluster (as the initial label in the Input step).

Next, we proceed to the matching process in Table 2, starting from the largest entity and looking for its members. At this stage, there is a parameter k in the algorithm. When the rank of the entity is within the limit of k (≤k, as in step 3) and the entity has a sufficiently large total number of links (i.e., higher network centrality), and there are followers pointing to it, the entity can form a new cluster (similar to children growing up, finding partners, and leaving their parents to form their own families). Otherwise, the algorithm corrects the previously linked followers and assigns them the label of the new leader (i.e., changing their original cluster label due to following another leader, as in steps 11–17). This completes the entire cluster analysis operation.

Table 2.

The following leader cluster algorithm (FLCA) used in this study (matching up).

Input k, sorting dataset (name, connections, cluster#)
Output the final clusters
1	Ensure k (in k, all entities having at least one follower become a leader)
2	For jk = 1 To n dataset A(size = n)
3	If jk <=k Then
4	Cluster#= dataset A.Cells(jk, 3)
5	Search for followers(with identical cluster#)
6	If found() Then
7	Dataset ACells(jk, 3) = giving the initial cluster#
8	End If
9	Else
10	Cluster#= dataset A.Cells(jk, 3)
11	For j = jk + 1 To n
12	If dateset A.Cells(j, 3) = jkThen
13	(=leader cluster#)
14	Dataset A Cells(j, 3) = cluster #
15	(=set identical cluster# with the leader cluster#)
16	End If
17	Next j end cluster# update
18	End If
19	Next jk end dataset A(size = n)
20	Renew cluster# with counts from 1 to the number of clusters

The FLCA algorithm^[29,30] can be represented by mathematical expressions (1) and (2): followers (based on the maximum association attribute, having at most one unique leader); sorting the weights of the leaders (i.e., the link counts in the network)^[17–19] in descending order and finding their followers; determining the number of cluster leaders based on the value of k. However, the number of clusters may vary depending on the value of k in the algorithm, typically more clusters are formed with a larger k value. To set an ideal value for k, we can use the minimum value of the absolute advantage coefficient (AAC)^[29,30] for weighting.

$${\displaystyle \begin{array}{l}{\displaystyle AAC=\left(R_{12}/R_{23}\right)/(1+\left(R_{12}/R_{23}\right)),}\end{array}}$$ (1)

$${\displaystyle \begin{array}{l}{\displaystyle R_{12}=A1/A2,}\end{array}}$$ (2)

$${\displaystyle \begin{array}{l}{\displaystyle R_{23}=A2/A3,}\end{array}}$$ (3)

where the AAC ratio is determined by the 3 consecutive numbers of values (e.g., top 3 CJAL scores in descending order denoted by A1, A2, and A3 in Eqs. 2 and 3). The ACC ranged from 0 to 1.0, representing the strength of dominance for the top member when compared to the next 2 members.

$$\{\begin{array}{l}{L}_i=\sum _{i=1}\sum _{j=}\left(F_j\in L_i\right), \left(4\right)\\ {\text{F}}_i=\underset{i=j+1}{\sum ^n}\underset{j=1}{\sum ^n}\left(F_j\in L_i\right), \left(5\right)\\ \text{if} \text{j}\le \text{k} \text{and} \text{with} \text{at} \text{least} \text{one} \text{follower}, \text{Li}=\text{j}\\ \text{if} \text{having} \text{maxmimum} \text{connections} \text{as} \text{the} \text{cluster}\\ \#\text{of} \text{the} \text{leader} j\\ \text{if} \text{j}>\text{k},the follower cluster=\text{the} \text{leader} \text{cluster}\\ {\displaystyle \sum \text{denotes}} \text{the} \text{search} \text{loop.}\end{array}$$

For simplification in computation of AAC for the determination of parameter k in Equation 5, A2 and a3 are assigned to 1.0 if the cluster number is one. A3 is assigned to 1.0 if the cluster number is 2.

2.2.2. Comparison of RAs in 2 facets.

2.2.2.1. Based on the CJAL score.

The calculation of the CJAL score^[11,24] involves analyzing the publications of both the first author and corresponding author,^[31] using a 4-quadrant plot^[31] to display the leading entities in spine research. Generally, authors with more publications tend to have higher JCAL scores. To compare RAs between neurosurgeons and orthopedic authors, only those in the same department were included, and the mean citations and CJAL scores were further analyzed.

2.2.2.2. Based on the mean citations and CJAL scores.

The mean citations and CJAL scores were compared using a forest plot^[32] based on the standardized mean difference.

2.2.3. Trend analysis of publications in journals and authors.

Three journals with the most publications were compared on their trends over years, along with the 3 categories of publications by authors in neurology, orthopedics, and both (i.e., 1st or corresponding authors are mixed).^[31]

2.2.4. Statistics and tools for visual representations.

Visualizations for this study were generated using R software^[33] and code written in R.^[34] These visualizations included circle bar charts, heatmaps with dendrograms, and forest plots, with a significance level set at Type I error (0.05). The AAC^[29,30] was applied to measure the dominance strength over the next 2 based on the CJAL. Through the computation of AAC, the dominance strength in a variable (i.e., country, institute, department, or author) can be measured and judged by the effect size, with criteria of <0.5, between 0.5 and 0.7, and not <0.7 as the small, medium, and large effect sizes, respectively.^[29]

To create the graphs, we utilized author-made modules on a website,^[34] specifically designed for bibliographical research. Additionally, CJAL scores for each member in each entity were plotted on a 4-quadrant radar plot, with HTML pages that were displayed on Google Maps. Supplemental Digital Content 2, http://links.lww.com/MD/K306 contains a detailed description of the visualization method used in this study.

3. Results

3.1. Comparison of country-based author collaborations

Orthopedic spine surgeons generally have more collaborations and coauthored papers in the field. This is evident when comparing a single cluster of authors from the US, South Korea, and Germany in neurosurgeons (Fig. 1) to 9 members in orthopedic spine surgeons (Fig. 2). Orthopedic spine surgeons from China and Japan contribute more publications than neurosurgeons. The FLCA process, depicted in Figures 3 and 4, shows that the clustering employed in this study is based on primary connections, distinguishing it from other algorithms that rely on similarity measures, such as correlation distance, Euclidean distance, K-means clustering, hierarchical clustering, density-based clustering, or others with eigenvalues and eigenvectors of a similarity matrix to partition the data points into clusters.

Figure 1.

Country-based author collaborations in neurosurgeons.

Figure 2.

Country-based author collaborations in orthopedic surgeons.

Figure 3.

Cluster analysis using a heatmap with a dendrogram for neurosurgeons.

Figure 4.

Cluster analysis using heatmap with dendrogram for orthopedic surgeons.

3.2. Comparison of total CJAL scores

Compared to authors in neurology, those in orthopedics have a greater number of publications and citations and higher CJAL scores in the field of spine research. The top entities for neurosurgeons include the US, Univ Calif San Francisco (U.S.), Neurosurgery, and Justin S Smith (U.S.), with both the dominant institute and author showing significant effects (AAC = 0.79 and 0.78, respectively). The computation of CJAL is presented in the bottom panel of Figure 5.

Figure 5.

Comparison of research achievements among elements in countries, institutes, departments, and authors of neurosurgeons.

The leading entities in orthopedics are the US, Capital Med Univ (China), Orthopedic Surgery, and Peter G Passias (U.S.), with only a dominant author showing significant effects over the next 2 authors (AAC = 0.81) in Figure 6.

Figure 6.

Comparison of research achievements among elements in entities of countries, institutes, departments, and authors in orthopedic surgeons.

3.3. Comparison of mean citations and CJAL scores

In terms of mean citations and CJAL scores, we found that orthopedic spine surgeons have an advantage in mean citations, while neurosurgeons have an advantage in mean SJAL scores (as shown in Figure 7). This suggests that the advantage for neurosurgeons is more related to the document type and journal impact factor, as citations in the L-index favor orthopedic spine surgeons and there is no effect on authorship (i.e., the 1st and corresponding authors are identical in both disciplines).

Figure 7.

Comparison of standardized mean difference (SMD) for neurosurgeons and orthopedic surgeons on the forest plot.

3.4. Comparison of mean citations and CJAL scores

When considering first and corresponding authors, orthopedics authors wrote a greater proportion of the articles in the article collections, accounting for 75% (1600 out of 2148). The JCAL score based on the top 10 units each also favored orthopedic spine surgeons, with 71% (3626 out of 6139) of the total score attributed to them. Using the FLCA, we observed that orthopedic spine surgeons tended to have more author collaborations across countries. Additionally, while citation per article favored orthopedic spine surgeons, the mean CJAL score favored neurosurgeons.

3.5. Trend analysis of publications in the top 3 journals, neurosurgeons and orthopedic spine surgeons

Figure 8 demonstrates that publications on spine research have been increasing more rapidly in orthopedics than in neurosurgeons. The 3 most prominent journals in this field are SPINE, J Neurosurg. Spine, and Eur. Spine J., with a consistent trend in publications over time.

Figure 8.

Trend analysis of publications in the top 3 journals and the types of authors in this study.

4. Discussion

4.1. Principal findings

Based on the first and corresponding authors, we found that orthopedic spine surgeons wrote a larger proportion of the articles in the article collections, accounting for 75% (1600 out of 2148). The CJAL score, based on the top 10 units, also favored orthopedic spine surgeons, with 71% (3626 out of 6139) of the total score attributed to them. Using the FLCA, we observed that orthopedic spine surgeons tended to have more author collaborations across countries. Furthermore, while citation per article favored orthopedic spine surgeons, the mean CJAL score favored neurosurgeons.

4.2. Additional information

Spinal surgery is taught and practiced within 2 different surgical disciplines, neurological surgery and orthopedic surgery.^[1] Previous studies have compared outcomes yielded by neurosurgeons and orthopedic surgeons in many areas, such as early outcomes and complications of pedicle subtraction osteotomies,^[2] representation of women on editorial boards of spine-related journals,^[3] resident training courses in adult spine surgery,^[4] workforce involvement,^[5] competency for common spinal conditions,^[6] differences in classifying cervical dislocation injuries and making assessment and treatment decisions,^[7] and RAs.^[1,8] However, Author collaborations in bibliometrics can be analyzed using various metrics such as co-authorship analysis, network analysis, and bibliographic coupling. Co-authorship analysis examines the collaboration between authors, while network analysis visualizes the relationships between authors in a network graph. Bibliographic coupling In 2 studies,^[1,8] the h-index was compared between neurosurgeons and orthopedic surgeons, but not based on spine research as we did in this study.

The evaluation^[1] examined 278 Accreditation Council for Graduate Medical Education training programs and found 923 full-time faculty members designated in spinal surgery. The findings revealed that neurological spine surgeons had a marginally higher average h-index than their orthopedic counterparts, and both groups demonstrated comparable trends of increasing h-index scores across all academic ranks.

The study^[8] utilized the American Association of Neurological Surgeons Neurosurgical Residency Training Program Directory to identify U.S. and Canadian academic neurological surgeons listed in the national institutes of health (NIH) Research Portfolio Online Reporting Tools database. A total of 215 neurological spine surgeons and 513 orthopedic spine surgeons were included, and their h-indices were compared. The mean h-index was 21.16 for neurological surgeons and 14.08 for orthopedic surgeons. Among neurosurgeons, those with NIH funding had significantly higher h-indices (34.15) than those without funding (19.29). Similarly, orthopedic surgeons with NIH funding had higher h-indices (42.83) than those without funding (13.39). Analysis of variance revealed that department chairs and professors in both neurological (P < .01) and orthopedic (P < .001) surgery had higher h-indices than associate or assistant professors.

The h-index takes into account both the number of publications and article citations.^[1]. In 2009, Lee et al^[35] used the h-index for the first time in the neurosurgical literature to evaluate 30 neurosurgical programs, which confirmed a positive correlation between the h-index and academic rank. Other researchers followed.^[36,37] Vitzthum et al^[38] used the h-index for the first time in the orthopedic literature in 2009 to perform a scientometric analysis of scoliosis research, which was followed by several other researchers.^[39,40] However, the h-index has limitations, such as giving equal credit to coauthors in an article,^[21] not considering journal impact factors and document types,^[22] and difficulty in distinguishing RAs within and between groups.^[23] Therefore, the CJAL score was used in this study.

Our study yielded contrasting results compared to prior research.^[1,8] According to the CJAL score, orthopedic spine surgeons had higher RAs than neurosurgeons. Nonetheless, it is noteworthy that neurosurgeons had a higher mean JCAL than orthopedic spine surgeons, which may be attributed to their contributions of more original articles published in journals with higher journal impact factors.

4.3. Implications and possible changes

The utilization of FLCA for author collaboration cluster analyses is a significant breakthrough in bibliometrics, providing valuable insights into cooccurrence patterns. To visually represent the FLCA results, we used unique and modern approaches such as circle bar charts and heatmaps with dendrograms.

The R code used to generate these visualizations is provided in Supplemental Digital Content 2, http://links.lww.com/MD/K306. This study encourages the use of FLCA for clustering author collaborations, which can also be applied to coword analysis if the cooccurrence phenomenon is based on the principal link of followers to potential leaders with a higher weighted centrality degree^[28] in the network.

FLCA provides a more comprehensive understanding of cooccurrence patterns than traditional approaches that rely on similarity and Euclidean distance.^[41] It is important to note that different approaches may produce different results, and the focus should be on the connection in cooccurrence rather than the similarity of occurrence patterns.

This study is distinguished by 4 key features: the development of FLCA, as illustrated in Tables 1 and 2; the comparison of FLCA with a heatmap combined with dendrograms in Figures 3 and 4, which can be applied beyond bibliometrics in future studies; the utilization of circle bar charts and the 2 axes with stacked and line charts as a unique visualization tool for presenting data in bibliometrics, accompanied by R code for future reference; and the provision of a method for generating a plot with documents in Supplemental Digital Content 2, http://links.lww.com/MD/K306 and the reference,^[42] which can be imitated and reproduced in the future.

4.4. Limitations and suggestions

Several issues should be considered in future studies. First, while the FLCA algorithm has simple and straightforward principles, further development is necessary, such as prioritizing higher individuals with more weighted centrality degree^[28] to lead followers based on the maximum connection and using a stop criterion to relate to others in the AC network (e.g., k in Eqs. 4 and 5).

Second, we focused only on dominant entities in spine research based on publications from 2013 to 2022. However, to gain a more comprehensive understanding, it would be beneficial to investigate RAs that may differ with different time frames, authors, and using the same indicator, such as CJAL, in future research, as we only considered first author and corresponding author in this study.

Third, while this study utilized R to create visual displays, other software options are available for network analysis. However, the focus should be on creating easily understandable displays that simplify complex information for readers with the help of user-friendly tools on websites.^[34]

Fourth, the CJAL score and analysis of RAs only considers the first and corresponding authors,^[31] as they contribute the most to the articles. This approach may lead to different results in cluster analysis compared to other methods, such as CiteSpace^[43,44] and VOSviewer,^[45] which take all coauthors into account when performing author collaborations.

Finally, accurate placement of variables and samples in columns or rows is crucial for the effectiveness of the heatmap combined with the dendrogram used to cluster them. As highlighted in Figure 1 with 2 green-font key points, inaccurate positioning of definitions may lead to difficulties in interpreting the cluster analysis results, and caution should be taken in the future.

5. Conclusion

Our study achieved 3 primary objectives: the development of FLCA for clustering ACs; demonstrating the application of cluster analysis with FLCA by analyzing a sample of articles related to spines authored by neurosurgeons and orthopedic surgeons; and comparing RAs between neurosurgeons and orthopedic spine surgeons: citation per article favored orthopedic spine surgeons and the mean CJAL score favored neurosurgeons.

The FLCA introduced in this study can make a significant contribution to our understanding of cooccurrence patterns in bibliometrics. Therefore, it is recommended for future bibliographical studies that are not limited to any specific research topic, as demonstrated in this study.

Acknowledgments

We thank Enago (www.enago.tw) for the English language review of this manuscript.

Author contributions

Conceptualization: Chun Hsiung.

Data curation: Po-Hsin Chou.

Formal analysis: Willy Chou.

Investigation: Tsair-Wei Chien.