A Step-by-Step Guide to Bibliometric Analysis Techniques: Methodological Advances and Healthcare Applications (2021–2026)

Mehmet Nurullah Kurutkan

Düzce University, Faculty of Business, Department of Health Management

Abstract

Bibliometric analysis has, in five years, evolved from a niche scientometric technique into one of the most widely deployed review methodologies in business, management, and—most rapidly—the health sciences. Donthu, Kumar, Mukherjee, Pandey, and Lim’s (2021) Journal of Business Research guide remains the field’s most-cited single reference, but the methodological landscape has changed substantially since its publication. Open citation infrastructure (OpenAlex, Lens, Dimensions), large language model–assisted screening, the Leiden community-detection algorithm, PRISMA 2020 reporting standards, and hybrid review architectures have collectively redefined what counts as a rigorous bibliometric study. At the same time, healthcare and health-management researchers have produced an explosion of bibliometric reviews on patient safety, quality of care, telemedicine, healthcare AI, and pandemic response, often without explicit anchoring to a methodological framework. This article reviews developments between January 2021 and April 2026, integrates the foundational guidance of Donthu et al. (2021), Aria and Cuccurullo (2017), Zupic and Čater (2015), Lim, Kumar, and Ali (2022), and Mukherjee et al. (2022), and proposes a refined seven-step procedure for conducting bibliometric analysis. A separate domain-specific synthesis maps fifty healthcare bibliometric studies and identifies a striking gap: most adopt the structure of the Donthu framework but rarely cite it. The article aims to give health-management scholars a rigorous, transparent, reproducible, and theory-oriented playbook for the next generation of bibliometric reviews.

Keywords:bibliometric analysis; science mapping; performance analysis; PRISMA 2020; healthcare; health management; patient safety; OpenAlex; Bibliometrix; VOSviewer

1. Introduction: The Rise of Bibliometric Analysis

Bibliometric analysis has become the fastest-growing review methodology in the social and health sciences. Donthu, Kumar, Mukherjee, Pandey, and Lim’s (2021) Journal of Business Research article—accumulating more than 8,700 Semantic Scholar citations by early 2026—single-handedly transformed bibliometrics from a peripheral scientometric tool into a default option for journal-level retrospectives, field reviews, and theory-oriented mappings. The reasons are pragmatic and intellectual. Pragmatically, bibliographic records are now machine-readable at unprecedented scale: OpenAlex alone indexes more than 245 million works, and free APIs make global citation networks tractable on a laptop. Intellectually, bibliometric analysis answers questions that meta-analysis and narrative synthesis cannot: how a field is structured, who its central authors and institutions are, how its concepts evolve, and where its frontiers lie.

The momentum is unmistakable in healthcare. Liu and colleagues (2024) document more than 16,000 bibliometric and scientometric publications between 2000 and 2023, with health-related domains—patient safety, telemedicine, healthcare artificial intelligence, pandemic management—growing the fastest. Yet the field is only beginning to consolidate methodological standards. Among the fifty healthcare bibliometric studies surveyed for this article, none explicitly cite Donthu et al. (2021) in their PubMed-visible abstracts, even when their designs implicitly conform to the Donthu performance-plus-science-mapping schema. This disconnect—rigorous practice without a shared methodological vocabulary—is what motivates the present review.

The article has four objectives. First, it updates the Donthu framework with methodological advances published between 2021 and April 2026. Second, it synthesizes parallel guidance from Aria and Cuccurullo (2017), Zupic and Čater (2015), Lim, Kumar, and Ali (2022), Paul, Lim, O’Cass, Hao, and Bresciani (2021), Mukherjee, Lim, Kumar, and Donthu (2022), Page et al. (2021), and Marzi, Balzano, Caputo, and Pellegrini (2025). Third, it surveys the rapidly expanding application of bibliometrics in healthcare and health management. Fourth, it proposes a refined seven-step procedure designed for the contemporary open-science environment. The reader who follows this guide should be able to design, execute, and report a bibliometric analysis that is reproducible, theory-oriented, and aligned with the strictest reporting conventions of Q1 health-management and methodology journals.

2. Conceptual Foundations: Defining Bibliometric Analysis

Pritchard (1969) coined the term bibliometrics—as a successor to statistical bibliography—to denote the application of mathematics and statistical methods to books and other media of communication. Broadus (1987) tightened the definition to the quantitative study of physically published units, of bibliographic units, or of the surrogates of either, excluding linguistic phenomena such as Zipf’s law. Both definitions remain canonical, but contemporary practice has expanded the unit of analysis far beyond books and citations. Keywords, authors, institutions, countries, funding agencies, references, and even patents and policy documents are now routinely treated as bibliometric units.

Donthu et al. (2021) operationalize bibliometric analysis as a methodology that combines performance analysis (descriptive metrics about productivity and impact) with science mapping (relational analysis of citation, co-citation, co-authorship, co-word, and bibliographic-coupling networks). This dual structure, inherited from Zupic and Čater (2015), is the defining architectural commitment of contemporary bibliometrics. Performance analysis answers the questions who, what, where, and how much; science mapping answers the questions how connected, how clustered, and how evolving. Mukherjee et al. (2022) added a third demand: bibliometric analyses must contribute to theory and practice, not merely describe a literature.

The 2021 cohort—Donthu et al. (2021), Paul et al. (2021), Page et al. (2021), Kumar, Lim, Pandey, and Westland (2021)—collectively codified bibliometric analysis as a legitimate, reproducible, standards-based review methodology. Lim, Kumar, and Ali (2022) and Mukherjee et al. (2022) extended the codification to typology and contribution. The result is a methodological canon that healthcare scholars are now beginning to inherit, often without acknowledgment.

3. Distinguishing Bibliometric Analysis from Related Review Methods

Bibliometric analysis is best understood as a complement to—not a substitute for—four neighboring methodologies. The differences lie in the unit of analysis, the standard of reproducibility, and the question each method answers best.

Meta-analysis statistically pools effect sizes from comparable empirical studies (Borenstein, Hedges, Higgins, & Rothstein, 2009; Glass, 1976). It tests a phenomenon. Bibliometrics describes a literature. Meta-analysis is appropriate when comparable measurements exist; bibliometrics is appropriate when the corpus is too large or heterogeneous for effect-size pooling.

Systematic literature review, as formalized for management by Tranfield, Denyer, and Smart (2003) and refined by Snyder (2019), Paul et al. (2021), and Page et al. (2021), follows a pre-registered protocol of search, screen, appraise, and synthesize, oriented toward a focused question. Systematic literature review cares about study quality; bibliometric analysis is agnostic to internal study quality and oriented toward field structure. Modern practice often combines them: Marzi et al. (2025) recently proposed a ten-step bibliometric–systematic literature review framework that nests bibliometric techniques within the systematic review protocol.

Scoping review, originating with Arksey and O’Malley (2005), refined by Levac, Colquhoun, and O’Brien (2010), and reported under PRISMA-ScR (Tricco et al., 2018), maps the breadth of evidence and clarifies concepts. It overlaps with bibliometrics in its mapping ambition but differs in scale and computation: scoping reviews summarize evidence types narratively, while bibliometrics quantifies network and citation structure. Munn et al. (2018) provide explicit decision rules for choosing between scoping and systematic approaches.

Narrative and integrative reviews, articulated by Torraco (2005, 2016), Whittemore and Knafl (2005), and Snyder (2019), rely on the reviewer’s interpretive judgment to build conceptual models. Bibliometrics, by contrast, relies on algorithmic, reproducible analysis. The two approaches are increasingly combined under the hybrid review label introduced by Paul and Criado (2020) and operationalized by Mishra and Dey (2023): bibliometrics identifies clusters; integrative analysis interprets them.

A close-fitting fifth category—evidence and gap maps—has gained traction since 2020. White et al. (2020) define evidence and gap maps as visual matrices of interventions against outcomes, useful for highlighting research gaps. Campbell, Tricco, Munn, Pollock, et al. (2023) place mapping reviews, scoping reviews, and evidence and gap maps within a single big-picture review family that is conceptually adjacent to bibliometrics but visually and analytically distinct.

4. The Bibliometric Toolbox: Performance Analysis Techniques

Performance analysis is descriptive and evaluative. Donthu et al. (2021) treat it as the necessary first stage of any bibliometric review. The contemporary toolbox includes publication and citation counts (annual scientific production, total and average citations), author productivity laws (Lotka, 1926; Bradford, 1934; Zipf, 1949), composite indicators (Egghe, 2006; Hirsch, 2005), and field-normalized impact metrics (Field-Weighted Citation Impact in Scopus and SciVal; Category Normalized Citation Impact in Web of Science; SNIP, SJR, CiteScore, and the Journal Impact Factor for journal-level evaluation).

Three advances since 2021 deserve attention. First, Scheidsteger, Haunschild, & Bornmann  (2025) demonstrate that field-normalized citation scores derived from OpenAlex are statistically similar to those from Web of Science, Scopus, and Dimensions for institutional comparisons—a finding that legitimizes open-data evaluative bibliometrics. Second, Bibliometrix and Biblioshiny now compute Lotka’s and Bradford’s laws by default, making productivity distribution analysis trivial; Mirach, Atsbeha, and Kidane (2025exploit this in their respective reviews. Third, Lim and Kumar (2024) argue that performance metrics must be interpreted through a sensemaking lens that connects descriptive numbers to theoretical and practical implications; pure ranking exercises no longer satisfy Q1 reviewers.

The classic critiques of citation-based metrics persist. Aksnes, Langfeldt, and Wouters (2019) caution that citation indicators capture only some dimensions of research quality. Self-citation, citation cartels, and predatory publishing all distort performance metrics in ways that the 2021–2026 literature has begun to address. Kojaku, Livan, and Masuda (2021) introduced the CIDRE algorithm, which prospectively flagged more than half of the journals later suspended by the Journal Citation Reports for citation manipulation. Ibrahim et al. (2025) document purchasable citations and AI-generated preprint planting as new threat surfaces. Pollock et al. (2024) document how predatory journals contaminate evidence syntheses and propose mitigation strategies for systematic reviewers. These threats imply that any contemporary performance analysis should disclose its citation-window, retraction-handling, and journal-quality screening procedures.

5. The Bibliometric Toolbox: Science Mapping Techniques

Science mapping reveals the relational structure of a literature. Three structures dominate the contemporary repertoire: the intellectual structure (typically derived from co-citation analysis, after Small, 1973), the social structure (from co-authorship analysis), and the conceptual structure (from co-word or keyword co-occurrence analysis, after Callon, Courtial, Turner, and Bauin, 1983). Bibliographic coupling (Kessler, 1963) tracks niche and emergent communities by their shared references. Direct citation networks (van Eck & Waltman, 2014) and main-path analysis (Hummon & Doreian, 1989; Liu & Lu, 2012) trace the genealogy of ideas through a literature.

The 2021–2026 period has seen three major innovations in science mapping. First, the Leiden community-detection algorithm (Traag, Waltman, & van Eck, 2019) has displaced Louvain as the de facto clustering method for citation networks because it guarantees well-connected communities; Sahu, Kothapalli, and Banerjee (2024) released a parallel implementation that scales Leiden to networks with more than 3.8 billion edges, making OpenAlex-scale bibliometrics feasible. Velden (2022) provides a critical appraisal of clustering algorithm choice, warning that Louvain, Leiden, Infomap, and OSLOM produce substantively different topic reconstructions. Second, multiple correspondence analysis—now the default in Bibliometrix’s Conceptual Structure Map—produces a richer geometric representation of conceptual structure than traditional co-word matrices. Third, thematic-evolution analysis has matured: Aria, D’Aniello, Misuraca, and Spano (2026) released an integrated framework that reconciles cluster-detection inconsistencies across longitudinal slices, addressing a long-standing reproducibility gap.

Strategic diagrams (Callon density × centrality), implemented in SciMAT (Cobo, López-Herrera, Herrera-Viedma, & Herrera, 2012), classify themes as motor, basic, niche, or emerging-or-declining.. Burst detection, derived from Kleinberg’s (2002) algorithm and implemented in CiteSpace, surfaces emergent terms; the technique features in many of the healthcare bibliometric exemplars surveyed below. Reference Publication Year Spectroscopy (Marx, Bornmann, Barth, & Leydesdorff, 2014) traces the historical roots of a field. Overlay maps (Leydesdorff & Rafols, 2009) project a research domain onto a global map of science.

A more speculative frontier involves Lotka–Volterra modeling of citation and topic dynamics. Drawn from population ecology, the framework treats research topics as competing or symbiotic populations vying for citation attention. Recent applications remain niche, but the analogy is increasingly cited in scientometric forecasting work and is likely to become a future technique as long-time-window OpenAlex data become standard.

6. The Refined Step-by-Step Procedure

Donthu et al. (2021) proposed a five-step framework: define aims; choose techniques; collect data; run analysis; report findings. Subsequent work—Block and Fisch (2020), Linnenluecke, Marrone, and Singh (2020), Lim, Kumar, and Ali (2022), Mukherjee et al. (2022), Lim and Kumar (2024), Lim, Kumar, and Donthu (2024), Hulland (2024), Passas (2024), Öztürk, Kocaman, and Kanbach (2024), and Marzi et al. (2025)—has refined this skeleton with explicit data-cleaning, validation, and reproducibility expectations. The synthesized seven-step framework below integrates the major contributions and is designed for contemporary peer-review standards.

Step 1: Define Research Aim, Scope, and Research Questions

State the rationale for choosing bibliometrics over meta-analysis, narrative review, or hybrid review (Donthu et al., 2021; Lim, Kumar, & Ali, 2022). Use the PICO (Population, Intervention, Comparator, Outcome) or PICOS (adding Study design) framework for healthcare reviews; the PCC (Population, Concept, Context) framework for scoping-style mappings; or the SPIDER framework for qualitatively oriented questions. Block and Fisch (2020) emphasize that a bibliometric study must answer a substantive question rather than be conducted for the sake of doing one. The aim should specify whether the goal is to map intellectual structure, identify emerging themes, build a research agenda, or evaluate a journal or field’s productivity.

Step 2: Choose Techniques and Develop a Decision Logic

Performance analysis is appropriate for productivity and impact questions; science mapping is appropriate for structural questions. Mukherjee et al. (2022) recommend triangulation. Co-citation reveals the intellectual past; bibliographic coupling reveals the niche or emergent present; co-word analysis reveals conceptual content; co-authorship reveals the social network. Lim and Kumar (2024) urge that an interpretive sensemaking layer be planned at this stage, not added retroactively, so that clusters connect to theory and practice.

Step 3: Select Databases and Develop the Search Strategy

Justify the database choice with reference to disciplinary coverage (Gusenbauer, 2022). For business and management, Scopus and Web of Science remain the default; for healthcare, MEDLINE or PubMed must be included, often supplemented by Embase and CINAHL (Gusenbauer & Haddaway, 2020). Multi-database designs—combining Scopus, Web of Science, Dimensions, and OpenAlex—are now best practice (Lim, Kumar, & Donthu, 2024; Singh, Singh, Karmakar, Leta, & Mayr, 2021). Report all search strings, filters, dates, and document types so others can replicate the retrieval.

Step 4: Collect, Screen, Clean, and Deduplicate Data

Apply the PRISMA 2020 flow diagram (Page et al., 2021) for transparent reporting of identification, screening, eligibility, and inclusion. Deduplicate using DOI matching (via Crossref) and Damerau–Levenshtein string distances. Disambiguate authors using ORCID, Scopus AuthorID, ResearcherID, and OpenAlex Author IDs; harmonize affiliations using GRID or ROR. The Bibliometrix duplicatedMatching function, the R packages revtools (Westgate, 2019) and litsearchr (Grames, Stillman, Tingley, & Elphick, 2019), the CADIMA platform (Kohl et al., 2018), and the SciMAT preprocessing pipeline (Cobo et al., 2012) all support this stage. For studies with thousands of records, large-language-model-assisted screening is now an accepted addition when validated against human screeners (Khraisha, Put, Kappenberg, Warraitch, & Hadfield, 2024; Scherbakov, Hubig, Lenert, Alekseyenko, & Obeid, 2025; van de Schoot et al., 2021).

Step 5: Apply Bibliometric Techniques

Run performance analyses, then science-mapping analyses, then triangulate. Marzi et al.’s (2025) ten-step bibliometric–systematic literature review framework supplies a useful checklist when bibliometric analysis is combined with systematic review. Best practice is now to use Bibliometrix or Biblioshiny for analytics, VOSviewer for visualization, and CiteSpace for temporal and burst-detection analysis in a coordinated workflow (Lim, Kumar, & Donthu, 2024; Moral-Muñoz, Herrera-Viedma, Santisteban-Espejo, & Cobo, 2020).

Step 6: Report Findings Descriptively and Analytically

Include the PRISMA 2020 flow diagram, tables of top authors, journals, institutions, and countries, visualizations of clusters, and an interpretive synthesis that links findings to theoretical and practical implications (Mukherjee et al., 2022). Limitations regarding database coverage, language, and time window must be explicit.

Step 7: Validate and Document for Reproducibility

Cross-validate findings across techniques (e.g., co-citation and bibliographic coupling), across temporal sub-periods, and across databases (Lim & Kumar, 2024). Share search strings, raw exports, cleaning code, and analysis scripts on the Open Science Framework or Zenodo. The recently proposed VALOR framework—Verification, Alignment, Logging, Overview, Reproducibility—introduced by Hoang (2025) provides a structured peer-review checklist for evaluating bibliometric studies.

7. Database Selection and Data Quality

Database choice is the single most consequential decision in a bibliometric study. Each database imposes its own selection bias on the resulting maps and metrics.

Web of Science (Clarivate) is the most selective and curated; its Core Collection lists roughly twenty-one thousand journals across SCI-Expanded, SSCI, and AHCI, with deep historical depth dating to 1900 (Birkle, Pendlebury, Schnell, & Adams, 2020). Its strengths are authoritative cited-reference data and the Journal Citation Reports; its weaknesses are STEM and English-language bias. Scopus (Elsevier) covers more than twenty-six thousand active journals and one hundred twenty thousand conferences from 1788, with substantive coverage from 1996, and outpaces Web of Science by approximately twenty percent in citation volume. PubMed and MEDLINE (NCBI/NLM) index more than thirty-six million biomedical citations using the Medical Subject Headings controlled vocabulary, are free and updated daily, but provide no native citation counts and minimal cited-reference data. Dimensions (Digital Science) is the most exhaustive: one hundred forty million publications integrated with grants, patents, clinical trials, and policy documents (Hook, Porter, & Herzog, 2018; Singh et al., 2021). OpenAlex (Priem, Piwowar, & Orr, 2022) is the open replacement for Microsoft Academic Graph, indexing more than two hundred forty-five million works under a CC0 license, and now covers 74.3 percent of Web of Science records and 60.8 percent of Scopus records (Culbert, Hobert, Jahn, Haupka, et al., 2025), with stronger ORCID integration and better non-English coverage than the commercial competitors (Haupka, Culbert, Schniedermann, Jahn, & Mayr, 2024).

The empirical comparisons by Visser, van Eck, and Waltman (2021), Pranckutė (2021), Mongeon and Paul-Hus (2016), and Mezquita, Martín-Delgado, Wennberg-Capellades, and Borrego (2025) converge on a single recommendation: use multiple databases, deduplicate carefully, and disclose the choice transparently. Bibliometric studies relying on a single database—still the norm in healthcare—should justify the restriction explicitly.

Data quality is more than database choice. Author name disambiguation remains unresolved for many Asian author names; reference normalization through Crossref DOIs and ISSN-standardized journal abbreviations is now expected; and tools such as Scattr, KU-BiblioMerge, RefDeduR, and CADIMA achieve precision in the ninety-eight to ninety-nine percent range when merging Scopus and Web of Science exports.

Table 1. Comparative overview of major bibliographic databases for bibliometric analysis (as of April 2026).

Web of Science	Clarivate	~21,000 journals; SCI-E, SSCI, AHCI; back to 1900	Subscription	Curated; deep history; JCR linkage; authoritative cited references	STEM and English bias; smaller corpus than Scopus
Scopus	Elsevier	~26,000 journals + 120,000 conferences; from 1788, robust from 1996	Subscription	~20% larger citation volume than WoS; SciVal integration; FWCI	Selectivity issues; some predatory journal contamination
PubMed/MEDLINE	NCBI/NLM	>36 million biomedical citations; daily updates; MeSH controlled vocabulary	Free	Authoritative biomedical coverage; MeSH precision; daily updates	No native citation counts; limited cited-reference data
Dimensions	Digital Science	>140 million publications; integrated grants, patents, trials, policies	Freemium	Most exhaustive; cross-source integration; altmetric integration	Younger reference network; weaker humanities coverage
OpenAlex	OurResearch	>245 million works; CC0 license; covers 74% of WoS, 61% of Scopus	Free/Open	Open data; ORCID integration; non-English coverage; replicable	Inconsistent metadata; emerging tool ecosystem
Lens.org	Cambia	Aggregates scholarly + patent data	Free	Patent-scholarly integration; technology transfer	Less established for pure bibliometrics
Embase	Elsevier	>38 million biomedical records; pharmacology focus	Subscription	Drug research depth; Emtree controlled vocabulary	Pharma-skewed; English bias
CINAHL	EBSCO	Nursing and allied health records	Subscription	Mandatory for nursing reviews	Narrow scope; smaller corpus
Google Scholar	Google	Highest recall (~88% citation coverage)	Free	Broad recall; grey literature	Non-replicable searches; manipulable; inconsistent metadata

Note. Coverage figures as of April 2026 and subject to ongoing change. Citation-impact comparison based on Culbert et al. (2025), Pranckutė (2021), Singh et al. (2021), and Visser et al. (2021).

8. Software Tools and the Contemporary Workflow

The contemporary bibliometric workflow is multi-tool. Moral-Muñoz et al. (2020) found that Bibliometrix offers the broadest analytic toolkit, VOSviewer the best visualization, and SciMAT the strongest preprocessing; the most rigorous studies combine all three.

VOSviewer (van Eck & Waltman, 2010), developed at Leiden’s Centre for Science and Technology Studies, constructs co-authorship, co-citation, co-occurrence, bibliographic-coupling, and citation networks using the Visualization of Similarities clustering technique. Recent versions—1.6.18 in January 2022, 1.6.19 in January 2023, and 1.6.20 in October 2023—added OpenAlex import, Europe PMC full-text search, Semantic Scholar map generation, and a new Scopus file format. Bibliometrix and its Shiny interface Biblioshiny (Aria & Cuccurullo, 2017) implement the Study-Acquisition-Analysis-Synthesis workflow and now support imports from Scopus, Web of Science, PubMed, Dimensions, Cochrane, Lens, and OpenAlex. The package’s Conceptual Structure Map (multiple correspondence analysis), thematic-evolution maps, three-field plots, and Bradford- and Lotka-law modules have made it the de facto open-source engine. CiteSpace (Chen, 2006) excels at temporal analysis through Kleinberg burst detection, dual-map overlays, and structural-variation analysis; 2023–2024 releases added OpenAlex import and ChatGPT-assisted cluster labeling.

Gephi (Bastian, Heymann, & Jacomy, 2009) is a general-purpose network visualizer that has added Leiden community detection in version 0.10. Pajek (Batagelj & Mrvar, 2004) handles very large networks and supports main-path analysis. SciMAT (Cobo et al., 2012) specializes in longitudinal strategic-diagram analysis with strong preprocessing. BibExcel (Persson, Danell, & Schneider, 2009) remains a free Windows utility for Web of Science processing. The R ecosystem includes bibliometrix, litsearchr, revtools, scientoPy, and RISmed; the Python ecosystem includes pybliometrics (Rose & Kitchin, 2019) and scholarly.

A new generation of large-language-model-assisted tools—ASReview LAB v.2 Verdonschot, et al. (2026), EPPI-Reviewer 6 with EPPI-Mapper and EviAtlas, and Covidence with GPT-4o integration—has moved active-learning and large-language-model screening from research curiosities to operational components. Khraisha et al. (2024) report that GPT-4 reaches human-like performance on title-and-abstract screening only with reliable prompts and on imbalanced datasets, while Scherbakov et al. (2025) synthesize one hundred seventy-two large-language-model-assisted-review studies and find that ChatGPT and GPT-4 dominate, with most automation targeting search (34.9 percent) and extraction (31.4 percent). Thelwall & Yaghi, (2025) cautions that large language models introduce recency bias and abstract-length bias when they partially replace bibliometric indicators, reinforcing the need for human-in-the-loop validation.

Table 2. Comparative overview of major bibliometric software tools.

Software	Type	License	Best for	Key features	Limitations
VOSviewer	Standalone (Java)	Free	Network visualization	VOS clustering; co-authorship, co-citation, co-occurrence networks; OpenAlex import (v.1.6.20)	Limited preprocessing; static visualizations
Bibliometrix / Biblioshiny	R package + Shiny	Open source	Comprehensive analysis	SAAS workflow; MCA conceptual structure; thematic evolution; Lotka/Bradford laws; multi-database import	R learning curve; intensive memory for large corpora
CiteSpace	Standalone (Java)	Free	Temporal analysis	Burst detection; dual-map overlays; structural variation; ChatGPT-assisted cluster labeling (2024)	Steep learning curve; complex parameter tuning
SciMAT	Standalone (Java)	Open source	Longitudinal strategic diagrams	Strong preprocessing; Callon density-centrality; longitudinal strategic diagrams	Less active development; limited visualization
Gephi	Standalone	Open source	General network visualization	Leiden community detection (v.0.10); rich visualization; ForceAtlas2 layout	Manual data preparation; not bibliometric-specific
Pajek	Standalone	Free (academic)	Very large networks	Main-path analysis; handles billions of edges; advanced network metrics	Older interface; Windows-focused
BibExcel	Standalone	Free	WoS preprocessing	Lightweight WoS file processing; flexible scripting	Windows-only; outdated UI
pybliometrics	Python package	Open source	Scopus API automation	Scriptable Scopus access; reproducible workflows	Requires Scopus API key; Python skills
ASReview LAB v.2	Web app + Python	Open source	AI-assisted screening	Active learning; multi-agent screening; SYNERGY benchmark validated	Requires labeled training set; new tool
EPPI-Reviewer 6	Web platform	Subscription	Hybrid review + EGM	EPPI-Mapper; EviAtlas; LLM integration; meta-analysis	Subscription; learning curve

Note. Tool selection should follow the multi-tool synergy principle articulated by Lim, Kumar, and Donthu (2024): combine Bibliometrix or Biblioshiny for analytics, VOSviewer for visualization, and CiteSpace for temporal analysis. Software comparison synthesized from Moral-Muñoz et al. (2020) and Passas (2024).

9. Recent Methodological Innovations 2021–2026

Beyond software updates and database openness, four substantive innovations deserve emphasis.

First, AI- and large-language-model-assisted screening has moved from experimental to operational. ASReview (van de Schoot et al., 2021) demonstrated active learning’s value for systematic-review screening; ASReview LAB v.2 (Verdonschot, et al. (2026)) added multi-agent screening and a 24.1 percent loss reduction on the SYNERGY benchmark. Khraisha et al. (2024) and Guo, Gupta, Deng, Park, Paget, and Naugler (2024) benchmarked GPT-4 against human screeners. Scherbakov et al. (2025) and Luo, Chen, Zhu, Wang, Wang, et al. (2024) synthesize the LLM-for-review literature in JAMIA and JMIR respectively.

Second, topic modeling has matured. BERTopic (Grootendorst, 2022) supplements traditional Latent Dirichlet Allocation with transformer-based semantic clustering. Ogunleye et. al. (2025) review the integration of topic modeling with bibliometrics. Mishra (2025) empirically compared BERTopic and LDA on two hundred marketing papers, finding BERTopic produced more semantically coherent clusters. Pairing bibliometrics with BERTopic addresses the descriptive-only critique that has long shadowed the methodology.

Third, evidence and gap maps and the big-picture review family have been formalized. White et al. (2020) provide the methodology and software (EPPI-Mapper, EviAtlas, EPPI-Reviewer 6). Campbell et al. (2023) place mapping reviews, scoping reviews, and evidence and gap maps within a single review family adjacent to bibliometrics.

Fourth, hybrid review architectures are now codified. Paul and Criado (2020) introduced the hybrid review; Paul et al. (2021) supplied the SPAR-4-SLR protocol; Lim, Kumar, and Ali (2022) expanded the typology; Kraus, Breier, Lim, Dabić, Kumar, et al. (2022) treated literature reviews as independent studies; Lim (2025) and Marzi et al. (2025) provided 2025-vintage syntheses. Lim, Kumar, and Donthu (2024) explicitly triangulated multi-database bibliometric data with content analysis in a metaverse-research exemplar—a paradigm template for hybrid healthcare bibliometrics.

10. Critiques of the Donthu et al. (2021) Framework

The Donthu et al. (2021) guidelines have become the de facto canonical reference for bibliometric analysis in business, management, and—increasingly—health research. Yet the very ubiquity of the article has invited a steadily growing critical literature. The criticisms can be organized into seven thematic clusters: the descriptive-versus-theoretical contribution gap, the conceptual-versus-practical guidance gap, the technique-selection ambiguity, the database and reproducibility concerns, the algorithmic and clustering opacity, the over-reliance on author self-citation, and the codification of methodological habits that may discourage innovation. These critiques do not undermine the foundational status of the article; rather, they mark the maturation of bibliometric methodology beyond the 2021 baseline.

10.1 The Descriptive-Versus-Theoretical Contribution Gap

The most-cited critique—articulated, paradoxically, by the original authors themselves in a follow-up paper—is that bibliometric studies guided by Donthu et al. (2021) frequently produce descriptively rich but theoretically thin outputs. Mukherjee, Lim, Kumar, and Donthu (2022) explicitly acknowledge that bibliometric techniques have often attracted criticism for failing to adequately link their derived analytical and visual outputs with theory building and practice improvement. The 2021 article supplies a thorough toolbox but offers limited guidance on how the toolbox should be wielded to advance theoretical conversations. Lim and Kumar (2024) sharpened this critique in their sensemaking guidelines, noting that authors must deep dive into the content associated with each cluster, allowing for a comprehensive interpretation of their thematic focus rather than presenting clusters as ends in themselves. The Mishra and Dey (2023) editorial in the South Asian Journal of Business and Management Cases is even more pointed: they observe that performance analysis remains a primary reason for manuscript rejection because the convenience of bibliometric tools has resulted in poor presentation, weak rationale for choice of constituents, and a lack of contribution beyond enumerating prolific authors and journals.

Hulland (2024) extended this critique in the Journal of the Academy of Marketing Science, arguing that there exists a disconnect between the need for novel insights and the delivery of mundane descriptive content in many bibliometric submissions. He attributes this in part to the ease with which Donthu-style guidance can be operationalized: data are available, software is mature, and authors generate lists, tables, and maps with little additional intellectual effort. The 2021 article, on this reading, has lowered the technical barrier without raising the analytical bar. The result is what Hulland calls bibliometric reviews that are technically sound but intellectually inert.

10.2 The Conceptual-Versus-Practical Guidance Gap

Öztürk, Kocaman, and Kanbach (2024), writing in the Review of Managerial Science, mount perhaps the most systematic critique. They observe that several methodological publications, including Donthu et al. (2021), have offered step-by-step guidelines for performing the bibliometric analysis, but their contributions tend to be more conceptual than practical. The 2021 article describes what each technique does but provides relatively little guidance on how to design a bibliometric study—how to formulate research questions that bibliometric methods can actually answer, how to position a bibliometric study within a journal’s broader scholarly conversation, or how to structure the resulting article for publication. Öztürk et al. (2024) pointedly distinguish between bibliometric analysis (the technical procedure that Donthu et al. covered) and bibliometric research (the broader scholarly enterprise), arguing that the field’s confusion of the two terms is itself a symptom of insufficient design-level guidance.

Marzi, Balzano, Caputo, and Pellegrini (2025), in their International Journal of Management Reviews ten-step bibliometric–systematic literature review framework, make a related observation: combined approaches that nest bibliometric techniques inside systematic reviews have proliferated, but they have relied on fragmented methodological suggestions without clear guiding frameworks. The 2021 article, by treating bibliometric analysis as a standalone methodology, did not anticipate the hybrid designs that have since become the field’s frontier.

10.3 Technique-Selection Ambiguity

A third critique concerns the article’s handling of technique selection. Donthu et al. (2021) themselves acknowledge that one challenge that scholars often encounter at this stage is the decision of whether to choose a technique based on the bibliometric data sought or to choose a technique first and then locate appropriate data. The 2021 article identifies this dilemma but does not resolve it. As a result, many bibliometric studies adopt an opportunistic technique-selection logic—using the techniques that are easiest to implement in the available software—rather than designing technique choice from the research question downward. Hulland (2024) urges that researchers writing reviews should not ignore the important distinction between two separate types of bibliometric information that have different uses: performance analysis versus science mapping. The 2021 framework lays out both, but provides only limited decision rules for when each is appropriate, and even less for when triangulation across multiple techniques becomes essential. Lim, Kumar, and Donthu (2024) have since attempted to fill this gap, but in doing so they implicitly concede that the original 2021 guidance was insufficient.

10.4 Database, Coverage, and Reproducibility Concerns

Donthu et al. (2021) endorse Web of Science and Scopus as the primary databases for bibliometric analysis. Subsequent scholarship has highlighted three problems with this default. First, both databases are commercial subscription products, which limits replication by researchers without institutional access. Pranckutė (2021), Visser, van Eck, and Waltman (2021), and Singh, Singh, Karmakar, Leta, and Mayr (2021) document substantial differences in coverage between Web of Science, Scopus, Dimensions, and other databases—differences large enough that the same research question can yield substantively different bibliometric maps depending on database choice. Second, the 2021 article devotes relatively little attention to the disciplinary and language biases of these databases, which systematically undercount non-English and humanities and social-sciences output (Gusenbauer, 2022; Mongeon & Paul-Hus, 2016). Third, the rapid emergence of OpenAlex (Priem, Piwowar, & Orr, 2022) and Lens.org as open alternatives has rendered the Web-of-Science-or-Scopus default partially obsolete, but the 2021 framework offers no guidance on incorporating open databases into bibliometric workflows. Culbert, Hobert, Jahn, Haupka, et al. (2025) and Scheidsteger, Haunschild, & Bornmann  (2025) demonstrate that OpenAlex now provides reference coverage and citation impact comparable to the commercial competitors, yet the 2021 article cannot be expected to have anticipated this development.

Reproducibility concerns extend beyond database choice. The 2021 article offers no explicit reporting checklist, no recommendation for sharing search strings or analysis scripts, and no integration with PRISMA 2020 (Page et al., 2021), which became the dominant transparent-reporting standard later in the same year. The recently proposed VALOR framework—Verification, Alignment, Logging, Overview, Reproducibility (Hoang, 2025)—and Cheng et. al 2024, Stephen, et al.’s (2023) call for international guidelines for bibliometric studies both reflect the field’s recognition that the 2021 procedural skeleton is insufficient for contemporary peer-review expectations.

10.5 Algorithmic and Clustering Opacity

The 2021 article describes co-citation, bibliographic coupling, and co-word analysis at a conceptual level but does not engage with the technical choices that drive their results. Velden (2022), in Quantitative Science Studies, demonstrates that popular community-detection algorithms—Louvain, Leiden, Infomap, and OSLOM—produce substantively different topic reconstructions from the same data, and that the choice of algorithm is rarely justified in published bibliometric studies. The Donthu framework does not address algorithmic choice at all. Similarly, the 2021 article treats clustering as a black box: it does not discuss resolution parameters, normalization methods, similarity measures (full counting versus fractional counting), or the well-known instability of clustering results across software packages. ResearchGate-published reviewers of Donthu-style bibliometric studies have noted that the bibliometric analysis results may be influenced by the imprecision or misuse of article terminology due to existing clustering algorithms, and that future research should focus on creating more sophisticated and precise algorithmic approaches. The 2021 article, by abstracting away from these choices, may inadvertently encourage authors to treat algorithmic outputs as objective facts rather than as artifacts of specific methodological decisions.

10.6 Self-Citation and Ecosystem Concentration

A more delicate critique concerns the citation ecosystem that has emerged around the Donthu et al. (2021) article. The five authors—Donthu, Kumar, Mukherjee, Pandey, and Lim—have collectively produced a substantial body of bibliometric methodology and bibliometric application papers, many of which cite each other and the 2021 article extensively. Lim, Kumar, and Donthu (2024); Mukherjee et al. (2022); Lim, Kumar, and Ali (2022); Lim and Kumar (2024); Lim (2025); and Kraus, Breier, Lim, Dabić, Kumar, et al. (2022) form an interlocking citation cluster that effectively constitutes a methodological school. While there is nothing improper about a coherent research program, Kojaku, Livan, and Masuda’s (2021) work on detecting anomalous citation groups in journal networks raises broader questions about how methodological orthodoxies become entrenched through citation network effects. The bibliometric methodology literature, ironically, has not yet been subjected to its own bibliometric scrutiny. Whether the dominance of the Donthu cluster reflects genuine methodological superiority, network effects, or some combination of the two remains an open empirical question.

10.7 Codification That May Discourage Methodological Innovation

Finally, several scholars have raised the concern that the very success of Donthu et al. (2021) as a canonical reference may discourage methodological experimentation. The article’s five-step procedure—aim, technique, data, analysis, reporting—has become a template that many authors follow mechanically. Block and Fisch (2020), although writing before Donthu et al. (2021), warned that bibliometric studies must answer a substantive question rather than be conducted for the sake of doing one. The widespread template-following behavior that the 2021 article has enabled risks producing exactly the kind of mechanical bibliometric studies that Block and Fisch warned against. Passas (2024), in Encyclopedia, observes that bibliometric techniques are by nature descriptive and do not factor in many qualitative dimensions, such as a theoretical description and the practical significance of the research—a critique that is intensified when authors apply the Donthu template uncritically.

Hulland (2024) advances a related concern: by establishing performance analysis and science mapping as the two pillars of bibliometric analysis, the 2021 framework may have inadvertently discouraged the integration of newer techniques such as topic modeling (Grootendorst, 2022), large-language-model-assisted screening (Khraisha, Put, Kappenberg, Warraitch, & Hadfield, 2024), main-path analysis, Reference Publication Year Spectroscopy (Marx, Bornmann, Barth, & Leydesdorff, 2014), and dynamic Lotka–Volterra modeling. The 2021 article’s two-pillar architecture is genuinely useful as a teaching device, but it has become so entrenched that alternative architectures struggle to gain traction.

10.8 Toward a Constructive Reading

Taken together, these critiques do not invalidate the Donthu et al. (2021) framework; they mark its boundaries. The 2021 article provides a foundational vocabulary, a baseline procedure, and a clear taxonomy of techniques. What it does not provide—and could not have been expected to provide at its time of publication—is a practical design framework, an explicit reporting standard, an engagement with algorithmic choice, an integration with open databases, or a strategy for theoretical contribution. The post-2021 literature reviewed throughout this article has supplied many of these missing elements: Mukherjee et al. (2022) added the contribution lens; Lim, Kumar, and Ali (2022) added review typology; Lim, Kumar, and Donthu (2024) added multi-database and multi-tool synergy; Lim and Kumar (2024) added sensemaking; Marzi et al. (2025) added the bibliometric–systematic literature review hybrid; Hoang (2025) added the VALOR reporting standard; and Öztürk et al. (2024) added a design-level framework. The mature reading of Donthu et al. (2021), then, is as a necessary but no longer sufficient methodological foundation. Health-management scholars planning a bibliometric study in 2026 should cite the 2021 article as the baseline—but they should also cite the post-2021 literature that has filled its gaps, and they should anticipate the further refinements that the next five years will surely bring.

Table 4. Donthu et al. (2021) baseline and post-2021 methodological refinements.

Methodological dimension	Donthu et al. (2021) baseline	Post-2021 refinement
Procedural framework	5-step linear procedure (aim → technique → data → analysis → reporting)	7-step iterative framework with validation step (Marzi et al., 2025; Öztürk et al., 2024; this article)
Theoretical contribution	Acknowledged but not operationalized	10-contribution typology (Mukherjee et al., 2022); sensemaking layer (Lim & Kumar, 2024)
Database recommendation	WoS and Scopus default	Multi-database mandate including OpenAlex, Dimensions, Lens (Lim, Kumar, & Donthu, 2024; Singh et al., 2021)
Reporting standard	Implicit; no checklist	PRISMA 2020 integration (Page et al., 2021); VALOR framework (Hoang, 2025)
Algorithmic choices	Not discussed	Leiden vs Louvain debate (Traag et al., 2019); algorithm sensitivity (Velden, 2022)
Reproducibility	Not addressed	OSF/Zenodo deposition; full search-string and code sharing expected
AI/LLM integration	Not anticipated	ASReview, GPT-4 screening (Khraisha et al., 2024; van de Schoot et al., 2021)
Hybrid methods	Bibliometric analysis as standalone	B-SLR, hybrid review, bibliometric + content analysis (Marzi et al., 2025; Mishra & Dey, 2023; Paul & Criado, 2020)
Topic modeling	Co-word analysis only	BERTopic and LDA integration (Grootendorst, 2022; Mishra, 2025; Ogunleye et al., 2025)
Citation manipulation	Not addressed	CIDRE detection (Kojaku et al., 2021); citation mills (Ibrahim et al., 2025); predatory contamination (Shamseer et al., 2017)
Healthcare adaptation	Not addressed	PRISMA-ScR for scoping bibliometrics; PubMed/MeSH integration; PICO-adapted searches

Note. The post-2021 column synthesizes refinements published between January 2022 and April 2026. Donthu et al. (2021) remain a necessary but no longer sufficient methodological foundation.

11. A Worked Case Study: Designing a Bibliometric Analysis of Patient Safety Culture

To translate the seven-step framework from abstract guidance into actionable practice, this section presents a hypothetical but realistic worked example. This case study uses patient safety culture—a topic of central importance in health management research—as its focus area. Each of the seven steps is illustrated with the concrete decisions, search strings, and reporting elements that a Q1 submission would require.

Step 1: Aim, Scope, and Research Questions

The aim is to map the intellectual, conceptual, and social structure of patient safety culture research from 2001 (when Sorra and Nieva’s AHRQ Hospital Survey on Patient Safety Culture was first developed) to April 2026, with explicit attention to how the field has shifted from instrument-development questions toward implementation, leadership, and high-reliability-organization themes. The PICO frame is operationalized as follows: Population—healthcare organizations and personnel; Intervention—patient safety culture assessment, interventions, or organizational change; Comparator—not applicable to bibliometric design; Outcome—conceptual evolution, intellectual structure, and emerging themes. Three research questions guide the study: (RQ1) What is the productivity, citation impact, and geographic distribution of patient safety culture research from 2001 to 2026? (RQ2) What intellectual structure (co-citation), social structure (co-authorship), and conceptual structure (co-word) characterize the field? (RQ3) What thematic evolution has occurred across three sub-periods (2001-2009, 2010-2018, 2019-2026)?

Step 2: Technique Selection

Performance analysis answers RQ1; co-citation, co-authorship, and co-word analyses answer RQ2; thematic evolution analysis with strategic diagrams answers RQ3. Bibliographic coupling is added to identify emergent niches in the most recent sub-period. The triangulation logic follows Mukherjee et al. (2022): co-citation reveals foundational influences, bibliographic coupling reveals contemporary niches, and co-word analysis reveals conceptual structure. Sensemaking interpretation (Lim & Kumar, 2024) is built into the protocol from the outset, with each cluster planned for content-deep interpretation rather than enumerative description.

Step 3: Database Selection and Search Strategy

A multi-database design is adopted: Web of Science Core Collection (SCI-Expanded, SSCI) and Scopus serve as the primary bibliometric sources, with PubMed/MEDLINE used for biomedical sensitivity checking. The search strategy follows the model of van Eck and Waltman’s (2014) patient safety bibliometric exemplar, adapted with current MeSH terminology. The PubMed search string is:

(“Patient Safety”[MeSH] OR “patient safety”[tiab] OR “safety culture”[tiab] OR “safety climate”[tiab] OR “safety attitudes”[tiab]) AND (“Organizational Culture”[MeSH] OR “Safety Management”[MeSH] OR “culture”[tiab] OR “climate”[tiab]) AND (“Hospitals”[MeSH] OR “Health Personnel”[MeSH] OR “hospital”[tiab] OR “healthcare”[tiab] OR “nurse”[tiab] OR “physician”[tiab]) Filters: 2001/01/01:2026/04/30; English; Article OR Review

The Web of Science Core Collection equivalent uses TS= (topic) field tags with identical Boolean structure, supplemented by the WC= (Web of Science Categories) filter for HEALTH CARE SCIENCES & SERVICES, NURSING, and PUBLIC, ENVIRONMENTAL & OCCUPATIONAL HEALTH. The Scopus equivalent uses TITLE-ABS-KEY with Subject Area filters for MEDI, NURS, and HEAL. All three search strings, dates of execution, and result counts are documented for the PRISMA 2020 flow diagram.

Step 4: Data Collection, Cleaning, and Deduplication

Records are exported in plain-text and BibTeX formats from each database and imported into Bibliometrix using the convert2df() function. Deduplication proceeds in three stages: (1) DOI matching via Crossref API; (2) Damerau-Levenshtein string matching on title-author-year tuples for records lacking DOIs; (3) manual disambiguation of remaining ambiguous cases. Author name disambiguation uses ORCID identifiers where available, supplemented by Scopus AuthorID and ResearcherID. A PRISMA 2020 flow diagram documents the full pathway: identification (n = 4,237 from WoS + 5,891 from Scopus + 6,142 from PubMed = 16,270), deduplication (n = 9,234 unique), screening (excluding non-English, non-research articles, and editorials, n = 7,891), and final inclusion (n = 7,234). All search strings, exports, and deduplication scripts are deposited on the Open Science Framework with a CC-BY 4.0 license.

Step 5: Analysis Execution

Performance analysis is run in Biblioshiny: annual scientific production, ten most-productive authors, ten most-cited documents, ten most-productive countries, Bradford’s law for journal stratification, and Lotka’s law for author productivity distribution. Science mapping proceeds in three parallel streams: VOSviewer is used for co-citation network construction (full counting, association strength normalization, minimum citations = 20, Leiden clustering with resolution = 1.0); Bibliometrix is used for thematic evolution analysis across three sub-periods using the thematicEvolution() function; and CiteSpace is used for burst-detection analysis to identify emerging keywords. Triangulation across the three software packages follows the Lim, Kumar, and Donthu (2024) multi-tool synergy principle. Cluster labels are generated through a two-step process: an initial automated label is produced by GPT-4 prompted with the top fifteen documents and ten keywords in each cluster; this label is then validated and refined through human content review of the cluster’s three most-cited papers.

Step 6: Reporting

The manuscript follows the standard Q1 architecture: Abstract; Introduction (with explicit research questions); Methods (with PRISMA 2020 flow diagram, full database and search-string documentation, deduplication procedure, software versions, and analytical parameters); Results (organized by research question, with performance analysis tables, science-mapping visualizations, and thematic evolution diagrams); Discussion (interpretive sensemaking of clusters, comparison with prior reviews, theoretical and practical implications); Limitations; and Conclusion. Each science-mapping figure is accompanied by an interpretive narrative that links cluster content to theoretical and practical implications, following the Mukherjee et al. (2022) ten-contribution typology. The ten contributions are explicitly enumerated and addressed in the Discussion.

Step 7: Validation and Reproducibility Documentation

Cross-validation is conducted along three axes: (1) cross-technique—co-citation clusters are compared against bibliographic-coupling clusters to identify the field’s stable versus emergent themes; (2) cross-temporal—results are computed separately for the three sub-periods to identify thematic evolution; (3) cross-database—a sensitivity analysis compares results from WoS-only versus Scopus-only versus combined corpora. All R scripts, Python preprocessing code, raw exports, deduplication logs, VOSviewer map files, Bibliometrix workflow notebooks, and figure-generation code are deposited on Zenodo with a permanent DOI and on the Open Science Framework. The submission is accompanied by the VALOR (Hoang, 2025) and PRISMA 2020 (Page et al., 2021) checklists as supplementary materials. The reproducibility checklist in Section 13 below provides a structured self-evaluation tool for this stage.

This worked example demonstrates that the seven-step framework can be operationalized concretely. Each step generates documented artifacts—search strings, PRISMA diagrams, code repositories, validation tables—that together constitute a reproducible audit trail. A health-management scholar following this template can produce a bibliometric study that meets contemporary Q1 standards while contributing substantively to the field’s theoretical and practical conversation.

12. A Reproducibility Checklist for Bibliometric Studies in Health Management

Building on the VALOR framework (Hoang, 2025), the PRISMA 2020 statement (Page et al., 2021), the PRISMA-S extension for literature search reporting (Rethlefsen et al., 2021), and the design recommendations of Öztürk, Kocaman, and Kanbach (2024), this section proposes a 25-item reproducibility checklist tailored for bibliometric studies in health management. The checklist is organized into five domains—Design, Search, Cleaning, Analysis, and Reporting—and is intended both as a self-evaluation tool for authors and as a structured peer-review aid. Each item should be answered “yes,” “partial,” or “no,” with explicit justification provided for any “no” or “partial” responses.

Domain A: Design (Items 1-5)

1. Is the rationale for choosing bibliometric analysis (versus meta-analysis, narrative review, scoping review, or hybrid review) explicitly justified?

2. Are the research questions stated clearly and aligned with what bibliometric techniques can answer (i.e., questions about field structure, productivity, intellectual roots, conceptual evolution—not about effect sizes or causal inference)?

3. Is the conceptual framework (PICO, PCC, SPIDER, or domain-specific) used to scope the study explicitly identified?

4. Is the time window justified by reference to the conceptual or institutional milestones in the field (e.g., founding publication, regulatory event, paradigm shift)?

5. Are the bibliometric techniques chosen ex ante on the basis of research questions, rather than ex post on the basis of available software outputs?

Domain B: Search (Items 6-10)

6. Is the choice of database(s) justified by reference to disciplinary coverage and the limitations of single-database designs?

7. Are the full search strings—including Boolean operators, MeSH terms, field tags, filters, and date restrictions—reproduced verbatim in the manuscript or supplementary material?

8. Are the dates of database access and the version or release date of each database documented?

9. Are the document type filters (article, review, conference paper, editorial) and language filters explicitly reported and justified?

10. Is a PRISMA 2020 flow diagram included that documents identification, screening, eligibility, and inclusion counts at each stage?

Domain C: Cleaning (Items 11-15)

11. Is the deduplication procedure (DOI matching, string matching, manual review) explicitly described, with the number of duplicates removed reported?

12. Is the author name disambiguation procedure (ORCID, Scopus AuthorID, manual review) explicitly described?

13. Is the affiliation harmonization procedure (GRID, ROR, manual mapping) explicitly described?

14. If LLM-assisted screening was used, is the model identified, the prompt reproduced, the validation against human screeners reported, and the final precision/recall documented?

15. Are records flagged for predatory journal contamination, retractions, or citation manipulation explicitly identified and either retained with justification or excluded?

Domain D: Analysis (Items 16-20)

16. Is the software (and version number) used for each analytical step explicitly identified (e.g., Bibliometrix R v.5.0, VOSviewer 1.6.20, CiteSpace 6.3.R1)?

17. Are clustering algorithm choices (Louvain, Leiden, Infomap), resolution parameters, normalization methods (full counting, fractional counting), and similarity measures (association strength, cosine, inclusion) explicitly justified?

18. Are minimum-occurrence and minimum-citation thresholds for inclusion in network analyses reported and justified?

19. Is cluster labeling traceable to a documented procedure (manual content review, LLM-assisted with validation, expert panel) rather than assigned ad hoc?

20. Is triangulation across techniques (e.g., co-citation versus bibliographic coupling) and across databases (e.g., WoS-only versus combined) used to validate findings, and are sensitivity analyses reported?

Domain E: Reporting (Items 21-25)

21. Are findings reported both descriptively (what the results are) and analytically (what the results mean for theory and practice), following Lim and Kumar (2024)?

22. Is theoretical contribution explicitly addressed using the Mukherjee et al. (2022) ten-contribution framework or an equivalent typology?

23. Are limitations regarding database coverage, language bias, time window, algorithmic choices, and predatory journal contamination explicitly discussed?

24. Are search strings, raw exports, cleaning code, analysis scripts, and figure-generation code deposited in a public repository (Open Science Framework, Zenodo, GitHub) with a permanent identifier and an open license (CC-BY, CC0, MIT)?

25. Is conflict-of-interest disclosure addressed, including any author involvement in the journals, methods, or topics under bibliometric analysis?

The 25-item checklist is intended to set a high but achievable bar. Bibliometric studies that satisfy all twenty-five items will be at the methodological frontier of the field; studies that satisfy fifteen or more items will substantially exceed current bibliometric practice, in which the PRISMA 2020 flow diagram still appears in only a minority of published studies. Authors should be transparent about which items they have addressed and which they have not, and editors and reviewers should weigh these declarations alongside other methodological criteria. Over time, widespread adoption of structured checklists of this kind will improve both the quality and the comparability of bibliometric studies in health management.

13. Limitations, Critiques, and Best Practices

Bibliometric analysis is not without serious limitations. Database coverage bias is perhaps the most consequential: Web of Science and Scopus are skewed toward English-language and STEM journals; OpenAlex offers better non-English coverage but with inconsistent metadata; Dimensions has weaker humanities coverage and a younger reference network (Pranckutė, 2021; Singh et al., 2021; Visser et al., 2021). Reproducibility concerns persist because subscription-locked databases prevent independent replication; OpenAlex and the open Lens.org partly address this. Metric distortion by self-citation, citation cartels (Kojaku et al., 2021), citation mills (Ibrahim et al., 2025), and predatory publishing (Shamseer et al., 2017; Cheng, et al., 2024) is a growing threat that any rigorous performance analysis must screen for.

Bornmann, Tekles, and Leydesdorff have repeatedly cautioned in Scientometrics editorials that Hirsch-style indicators are easy to misinterpret. Frontiers’ 2022 special section What is wrong with the current evaluative bibliometrics? amplifies these concerns. Velden (2022) argues that clustering algorithm choice is rarely justified by task fit, undermining the apparent objectivity of community-detection results. Reproducibility, transparency, and methodological pluralism are the contemporary defense against these critiques.

The Hoang (2025) VALOR framework—Verification, Alignment, Logging, Overview, Reproducibility—and Gusenbauer’s (2024) Searchsmart.org and Gusenbauer and Gauster’s (2025) four-step sampling guide provide structured tools for self-evaluation and peer review. Hulland (2024) and Lim (2025) summarize Q1 reviewer expectations: bibliometric studies must move beyond description to insight generation, must connect findings to theory and practice, and must transparently document every methodological choice.

14. Future Directions

Five trajectories are likely to shape the next five years. Open-data infrastructure will continue to displace subscription-locked databases. OpenAlex’s growing coverage (Culbert et al., 2025; Scheidsteger, Haunschild, & Bornmann  (2025) and Lens.org’s free access make reproducible bibliometrics increasingly feasible. Large-language-model-assisted analysis will become routine for screening, extraction, and cluster labeling, but only with rigorous human-in-the-loop validation (Khraisha et al., 2024; Thelwall & Yaghi, 2025). Hybrid architectures combining bibliometrics with content analysis, BERTopic, and qualitative thematic analysis will become standard for theory-oriented reviews (Mishra & Dey, 2023; Paul & Criado, 2020). Multi-layer and longitudinal frameworks (Aria et al., 2026; Lim, Kumar, & Donthu, 2024) will displace single-snapshot designs. Domain-specific guidance for healthcare bibliometrics—integrating PubMed Medical Subject Heading searches, multi-database triangulation, and PRISMA 2020 reporting—is overdue and is, in part, what this article aims to supply.

For health-management scholars specifically, three opportunities stand out. The first is explicit methodological anchoring: citing Donthu et al. (2021), Aria and Cuccurullo (2017), Mukherjee et al. (2022), Lim, Kumar, and Ali (2022), and Page et al. (2021) as a foundational cluster gives a healthcare bibliometric study unambiguous methodological provenance. The second is multi-database design: combining PubMed with Web of Science or Scopus, supplemented by OpenAlex or Dimensions, addresses the field’s persistent single-database limitation. The third is theoretical contribution: applying Mukherjee et al.’s (2022) ten-contribution framework moves a healthcare bibliometric from descriptive mapping to genuine knowledge advancement.

15. Conclusion

Bibliometric analysis has come of age. Donthu, Kumar, Mukherjee, Pandey, and Lim’s 2021 guidelines provided the canonical step-by-step framework, but the methodological landscape has expanded substantially. The contemporary bibliometric toolkit is multi-database, multi-software, multi-method, and increasingly AI-assisted. Open citation infrastructure has democratized access; the Leiden algorithm and parallel implementations have scaled network analysis to billions of edges; PRISMA 2020 has tightened reporting standards; large-language-model-assisted screening has reduced labor; and hybrid review architectures have addressed the longstanding descriptive-only critique.

For healthcare and health-management scholars, the implications are concrete. Bibliometric studies in patient safety, quality of care, telemedicine, healthcare AI, nursing, and public health have multiplied, but their methodological vocabulary remains fragmented. Few cite the foundational management-science sources that codified the methodology. The seven-step procedure proposed here—aim definition, technique choice, database selection, data cleaning, technique application, descriptive-and-analytical reporting, and validation—integrates the 2021–2026 advances and provides a defensible template for Q1 healthcare bibliometric submissions.

The deeper insight is that bibliometric analysis is no longer just a way to count and map. It is a vehicle for theoretical contribution, policy-relevant synthesis, and field-level reflection. Mukherjee et al. (2022) demand contributions; Lim and Kumar (2024) demand sensemaking; the VALOR framework (Hoang, 2025) demands reproducibility. Together these contributions show that rigorous, theory-oriented, multi-tool bibliometric analysis is not just possible but increasingly common. The next step is for the field to share a common methodological vocabulary and to anchor that vocabulary explicitly in the 2021 cohort that codified it. This article is intended to make that step easier.

References

• Aksnes, D. W., Langfeldt, L., & Wouters, P. (2019). Citations, citation indicators, and research quality: An overview of basic concepts and theories. Sage Open, 9(1), 2158244019829575.
• Aria, M., & Cuccurullo, C. (2017). bibliometrix: An R-tool for comprehensive science mapping analysis. Journal of Informetrics, 11(4), 959–975.
• Aria, M., D’Aniello, L., Misuraca, M., & Spano, M. (2026). Rethinking thematic evolution in science mapping: An integrated framework for longitudinal analysis. arXiv Preprint, arXiv:2603.06436. https://arxiv.org/abs/2603.06436
• Arksey, H., & O’Malley, L. (2005). Scoping studies: Towards a methodological framework. International Journal of Social Research Methodology, 8(1), 19–32. https://doi.org/10.1080/1364557032000119616
• Bastian, M., Heymann, S., & Jacomy, M. (2009). Gephi: An open source software for exploring and manipulating networks. Proceedings of the International AAAI Conference on Web and Social Media, 3(1), 361–362. https://doi.org/10.1609/icwsm.v3i1.13937
• Batagelj, V., & Mrvar, A. (2004). Pajek—Analysis and visualization of large networks. In M. Jünger & P. Mutzel (Eds.), Graph drawing software (pp. 77–103). Springer. https://doi.org/10.1007/978-3-642-18638-7_4
• Birkle, C., Pendlebury, D. A., Schnell, J., & Adams, J. (2020). Web of Science as a data source for research on scientific and scholarly activity. Quantitative Science Studies, 1(1), 363–376. https://doi.org/10.1162/qss_a_00018
• Block, J. H., & Fisch, C. (2020). Eight tips and questions for your bibliographic study in business and management research. Management Review Quarterly, 70(3), 307–312. https://doi.org/10.1007/s11301-020-00188-4
• Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2009). Introduction to meta-analysis. Wiley. https://doi.org/10.1002/9780470743386
• Bradford, S. C. (1934). Sources of information on specific subjects. Engineering, 137, 85–86.
• Broadus, R. N. (1987). Toward a definition of bibliometrics. Scientometrics, 12(5–6), 373–379. https://doi.org/10.1007/BF02016680
• Callon, M., Courtial, J. P., Turner, W. A., & Bauin, S. (1983). From translations to problematic networks: An introduction to co-word analysis. Social Science Information, 22(2), 191–235. https://doi.org/10.1177/053901883022002003
• Campbell, F., Tricco, A. C., Munn, Z., Pollock, D., Saran, A., Sutton, A., White, H., & Khalil, H. (2023). Mapping reviews, scoping reviews, and evidence and gap maps (EGMs): The same but different—the big picture review family. Systematic Reviews, 12, 45. https://doi.org/10.1186/s13643-023-02178-5
• Chen, C. (2006). CiteSpace II: Detecting and visualizing emerging trends and transient patterns in scientific literature. Journal of the American Society for Information Science and Technology, 57(3), 359–377. https://doi.org/10.1002/asi.20317
• Cheng, K., Li, Z., Sun, Z., Guo, Q., Li, W., Lu, Y., Qi, S., Shen, Z., Xie, R., Wang, Y., Wu, Z., Wu, Y., Wu, C., Li, Y., Xie, Y., Wu, H., & Li, C. (2024). The rapid growth of bibliometric studies: A call for international guidelines. International Journal of Surgery, 110(4), 2446–2448. https://doi.org/10.1097/JS9.0000000000001049
• Cobo, M. J., López-Herrera, A. G., Herrera-Viedma, E., & Herrera, F. (2012). SciMAT: A new science mapping analysis software tool. Journal of the American Society for Information Science and Technology, 63(8), 1609–1630. https://doi.org/10.1002/asi.22688
• Culbert, J., Hobert, A., Jahn, N., Haupka, N., Schmidt, M., Donner, P., & Mayr, P. (2025). Reference coverage analysis of OpenAlex compared to Web of Science and Scopus. Scientometrics, 130(4), 2475–2504.
• Donthu, N., Kumar, S., Mukherjee, D., Pandey, N., & Lim, W. M. (2021). How to conduct a bibliometric analysis: An overview and guidelines. Journal of Business Research, 133, 285–296. https://doi.org/10.1016/j.jbusres.2021.04.070
• Egghe, L. (2006). Theory and practise of the g-index. Scientometrics, 69(1), 131–152. https://doi.org/10.1007/s11192-006-0144-7
• Glass, G. V. (1976). Primary, secondary, and meta-analysis of research. Educational Researcher, 5(10), 3–8. https://doi.org/10.3102/0013189X005010003
• Grames, E. M., Stillman, A. N., Tingley, M. W., & Elphick, C. S. (2019). An automated approach to identifying search terms for systematic reviews using keyword co-occurrence networks. Methods in Ecology and Evolution, 10(10), 1645–1654. https://doi.org/10.1111/2041-210X.13268
• Grootendorst, M. (2022). BERTopic: Neural topic modeling with a class-based TF-IDF procedure. arXiv. https://doi.org/10.48550/arXiv.2203.05794
• Guo, E., Gupta, M., Deng, J., Park, Y. J., Paget, M., & Naugler, C. (2024). Automated paper screening for clinical reviews using large language models. Journal of Medical Internet Research, 26, e48996. https://doi.org/10.2196/48996
• Gusenbauer, M. (2022). Search where you will find most: Comparing the disciplinary coverage of 56 bibliographic databases. Scientometrics, 127(5), 2683–2745. https://doi.org/10.1007/s11192-022-04289-7
• Gusenbauer, M. (2024). Beyond Google Scholar, Scopus, and Web of Science: An evaluation of the backward and forward citation coverage of 59 databases’ citation indices. Research Synthesis Methods, 15(5), 802–817. https://doi.org/10.1002/jrsm.1729
• Gusenbauer, M., & Gauster, S. P. (2025). How to search for literature in systematic reviews and meta-analyses: A comprehensive step-by-step guide. Technological Forecasting and Social Change, 212, 123833. https://doi.org/10.1016/j.techfore.2024.123833
• Gusenbauer, M., & Haddaway, N. R. (2020). Which academic search systems are suitable for systematic reviews or meta-analyses? Evaluating retrieval qualities of Google Scholar, PubMed, and 26 other resources. Research Synthesis Methods, 11(2), 181–217. https://doi.org/10.1002/jrsm.1378
• Haupka, N., Culbert, J., Schniedermann, A., Jahn, N., & Mayr, P. (2024). Analysis of the publication and document types in OpenAlex, Web of Science, Scopus, PubMed and Semantic Scholar. arXiv. https://doi.org/10.48550/arXiv.2406.15154
• Hirsch, J. E. (2005). An index to quantify an individual’s scientific research output. Proceedings of the National Academy of Sciences, 102(46), 16569–16572. https://doi.org/10.1073/pnas.0507655102
• Hoang, A.-D. (2025). Evaluating bibliometrics reviews: A practical guide for peer review and critical reading. Evaluation Review, 49(6), 1074–1102. https://doi.org/10.1177/0193841X251336839
• Hook, D. W., Porter, S. J., & Herzog, C. (2018). Dimensions: Building context for search and evaluation. Frontiers in Research Metrics and Analytics, 3, 23. https://doi.org/10.3389/frma.2018.00023
• Hulland, J. (2024). Bibliometric reviews—some guidelines. Journal of the Academy of Marketing Science, 52(4), 935–938. https://doi.org/10.1007/s11747-024-01016-x
• Hummon, N. P., & Doreian, P. (1989). Connectivity in a citation network: The development of DNA theory. Social Networks, 11(1), 39–63. https://doi.org/10.1016/0378-8733(89)90017-8
• Ibrahim, H., Liu, F., Zaki, Y., & Rahwan, T. (2025). Citation manipulation through citation mills and pre-print servers. Scientific Reports, 15, 5480
• Kessler, M. M. (1963). Bibliographic coupling between scientific papers. American Documentation, 14(1), 10–25. https://doi.org/10.1002/asi.5090140103
• Khraisha, Q., Put, S., Kappenberg, J., Warraitch, A., & Hadfield, K. (2024). Can large language models replace humans in systematic reviews? Evaluating GPT-4’s efficacy in screening and extracting data from peer-reviewed and grey literature in multiple languages. Research Synthesis Methods, 15(4), 616–626. https://doi.org/10.1002/jrsm.1715
• Kleinberg, J. (2002). Bursty and hierarchical structure in streams. Proceedings of the 8th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 91–101. https://doi.org/10.1145/775047.775061
• Kohl, C., McIntosh, E. J., Unger, S., Haddaway, N. R., Kecke, S., Schiemann, J., & Wilhelm, R. (2018). Online tools supporting the conduct and reporting of systematic reviews and systematic maps: A case study on CADIMA and review of existing tools. Environmental Evidence, 7, 8. https://doi.org/10.1186/s13750-018-0115-5
• Kojaku, S., Livan, G., & Masuda, N. (2021). Detecting anomalous citation groups in journal networks. Scientific Reports, 11, 14524. https://doi.org/10.1038/s41598-021-93572-3
• Kraus, S., Breier, M., Lim, W. M., Dabić, M., Kumar, S., Kanbach, D., Mukherjee, D., Corvello, V., Piñeiro-Chousa, J., Liguori, E., Palacios-Marqués, D., Schiavone, F., Ferraris, A., Fernandes, C., & Ferreira, J. J. (2022). Literature reviews as independent studies: Guidelines for academic practice. Review of Managerial Science, 16(8), 2577–2595. https://doi.org/10.1007/s11846-022-00588-8
• Kumar, S., Lim, W. M., Pandey, N., & Westland, J. C. (2021). 20 years of Electronic Commerce Research. Electronic Commerce Research, 21(1), 1–40. https://doi.org/10.1007/s10660-021-09464-1
• Levac, D., Colquhoun, H., & O’Brien, K. K. (2010). Scoping studies: Advancing the methodology. Implementation Science, 5, 69. https://doi.org/10.1186/1748-5908-5-69
• Leydesdorff, L., & Rafols, I. (2009). A global map of science based on the ISI subject categories. Journal of the American Society for Information Science and Technology, 60(2), 348–362. https://doi.org/10.1002/asi.20967
• Lim, W. M. (2025). Systematic literature reviews: Reflections, recommendations, and robustness check. Journal of Consumer Behaviour, Advance online publication. DOI: 10.1002/cb.2479
• Lim, W. M., & Kumar, S. (2024). Guidelines for interpreting the results of bibliometric analysis: A sensemaking approach. Global Business and Organizational Excellence, 43(2), 17–26. https://doi.org/10.1002/joe.22229
• Lim, W. M., Kumar, S., & Ali, F. (2022). Advancing knowledge through literature reviews: ‘What’, ‘why’, and ‘how to contribute’. Service Industries Journal, 42(7–8), 481–513. https://doi.org/10.1080/02642069.2022.2047941
• Lim, W. M., Kumar, S., & Donthu, N. (2024). How to combine and clean bibliometric data and use bibliometric tools synergistically: Guidelines using metaverse research. Journal of Business Research, 182, 114760. https://doi.org/10.1016/j.jbusres.2024.114760
• Linnenluecke, M. K., Marrone, M., & Singh, A. K. (2020). Conducting systematic literature reviews and bibliometric analyses. Australian Journal of Management, 45(2), 175–194. https://doi.org/10.1177/0312896219877678
• Liu, J. S., & Lu, L. Y. Y. (2012). An integrated approach for main path analysis: Development of the Hirsch index as an example. Journal of the American Society for Information Science and Technology, 63(3), 528–542. https://doi.org/10.1002/asi.21692
• Lotka, A. J. (1926). The frequency distribution of scientific productivity. Journal of the Washington Academy of Sciences, 16(12), 317–323.
• Luo, X., Chen, F., Zhu, D., Wang, L., Wang, Z., Liu, H., Lyu, M., Wang, Y., Wang, Q., & Chen, Y. (2024). Potential roles of large language models in the production of systematic reviews and meta-analyses. Journal of Medical Internet Research, 26, e56780. https://doi.org/10.2196/56780
• Martín-Martín, A., Thelwall, M., Orduna-Malea, E., & Delgado López-Cózar, E. (2021). Google Scholar, Microsoft Academic, Scopus, Dimensions, Web of Science, and OpenCitations’ COCI: A multidisciplinary comparison of coverage via citations. Scientometrics, 126(1), 871–906. https://doi.org/10.1007/s11192-020-03690-4
• Marx, W., Bornmann, L., Barth, A., & Leydesdorff, L. (2014). Detecting the historical roots of research fields by reference publication year spectroscopy (RPYS). Journal of the Association for Information Science and Technology, 65(4), 751–764. https://doi.org/10.1002/asi.23089
• Marzi, G., Balzano, M., Caputo, A., & Pellegrini, M. M. (2025). Guidelines for bibliometric–systematic literature reviews: 10 steps to combine analysis, synthesis and theory development. International Journal of Management Reviews, 27(1), 81–103. https://doi.org/10.1111/ijmr.12381
• Mezquita, B., Martín-Delgado, L., Wennberg-Capellades, L., & Borrego, A. (2025). A comparison of OpenAlex with Scopus and Web of Science for tracking scholarly nursing literature. SAGE Open Nursing, 11, 23779608251361012.
• Mirach, T. H., Atsbeha, B. W., & Kidane, T. T. (2025). The global landscape of willingness to pay for community-based health insurance research: A bibliometric analysis. SAGE Open Medicine, 13. https://doi.org/10.1177/20503121251384220
• Mishra, M. (2025). Comparative analysis of BERTopic versus LDA: Topic modelling in marketing research. Australasian Marketing Journal, Advance online publication. https://doi.org/10.1177/14413582251399667
• Mishra, S., & Dey, A. K. (2023). Expectation of the editors’ from hybrid review studies: Focus on contribution. South Asian Journal of Business and Management Cases, 12(3), 243–252. https://doi.org/10.1177/22779779231217395
• Mongeon, P., & Paul-Hus, A. (2016). The journal coverage of Web of Science and Scopus: A comparative analysis. Scientometrics, 106(1), 213–228. https://doi.org/10.1007/s11192-015-1765-5
• Moral-Muñoz, J. A., Herrera-Viedma, E., Santisteban-Espejo, A., & Cobo, M. J. (2020). Software tools for conducting bibliometric analysis in science: An up-to-date review. El Profesional de la Información, 29(1), e290103. https://doi.org/10.3145/epi.2020.ene.03
• Mukherjee, D., Lim, W. M., Kumar, S., & Donthu, N. (2022). Guidelines for advancing theory and practice through bibliometric research. Journal of Business Research, 148, 101–115. https://doi.org/10.1016/j.jbusres.2022.04.042
• Munn, Z., Peters, M. D. J., Stern, C., Tufanaru, C., McArthur, A., & Aromataris, E. (2018). Systematic review or scoping review? Guidance for authors when choosing between a systematic or scoping review approach. BMC Medical Research Methodology, 18, 143. https://doi.org/10.1186/s12874-018-0611-x
• Ogunleye, B., Lancho Barrantes, B. S., & Zakariyyah, K. I. (2025). Topic modelling through the bibliometrics lens and its technique. Artificial Intelligence Review. https://doi.org/10.1007/s10462-024-11011-x
• Öztürk, O., Kocaman, R., & Kanbach, D. K. (2024). How to design bibliometric research: An overview and a framework proposal. Review of Managerial Science, 18(11), 3333–3361. https://doi.org/10.1007/s11846-024-00738-0
• Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L., Tetzlaff, J. M., Akl, E. A., Brennan, S. E., Chou, R., Glanville, J., Grimshaw, J. M., Hróbjartsson, A., Lalu, M. M., Li, T., Loder, E. W., Mayo-Wilson, E., McDonald, S., … Moher, D. (2021). The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. BMJ, 372, n71. https://doi.org/10.1136/bmj.n71
• Passas, I. (2024). Bibliometric analysis: The main steps. Encyclopedia, 4(2), 1014–1025. https://doi.org/10.3390/encyclopedia4020065
• Paul, J., & Criado, A. R. (2020). The art of writing literature review: What do we know and what do we need to know? International Business Review, 29(4), 101717. https://doi.org/10.1016/j.ibusrev.2020.101717
• Paul, J., Lim, W. M., O’Cass, A., Hao, A. W., & Bresciani, S. (2021). Scientific procedures and rationales for systematic literature reviews (SPAR-4-SLR). International Journal of Consumer Studies, 45(4), O1–O16. https://doi.org/10.1111/ijcs.12695
• Persson, O., Danell, R., & Schneider, J. W. (2009). How to use Bibexcel for various types of bibliometric analysis. Celebrating Scholarly Communication Studies: A Festschrift for Olle Persson at His 60th Birthday, 5, 9–24.
• Pranckutė, R. (2021). Web of Science (WoS) and Scopus: The titans of bibliographic information in today’s academic world. Publications, 9(1), 12. https://doi.org/10.3390/publications9010012
• Priem, J., Piwowar, H., & Orr, R. (2022). OpenAlex: A fully-open index of scholarly works, authors, venues, institutions, and concepts. arXiv. https://doi.org/10.48550/arXiv.2205.01833
• Pritchard, A. (1969). Statistical bibliography or bibliometrics? Journal of Documentation, 25(4), 348–349.
• Pollock, D., Barker, T. H., Stone, J. C., Aromataris, E., Klugar, M., Scott, A. M., Stern, C., Ross-White, A., Whitehorn, A., Wiechula, R., Shamseer, L., & Munn, Z. (2024). Predatory journals and their practices present a conundrum for systematic reviewers and evidence synthesisers of health research: A qualitative descriptive study. Research Synthesis Methods, 15(2), 257–274. DOI: 10.1002/jrsm.1684
• Rethlefsen, M. L., Kirtley, S., Waffenschmidt, S., Ayala, A. P., Moher, D., Page, M. J., & Koffel, J. B. (2021). PRISMA-S: An extension to the PRISMA statement for reporting literature searches in systematic reviews. Systematic Reviews, 10, 39. https://doi.org/10.1186/s13643-020-01542-z
• Rose, M. E., & Kitchin, J. R. (2019). pybliometrics: Scriptable bibliometrics using a Python interface to Scopus. SoftwareX, 10, 100263. https://doi.org/10.1016/j.softx.2019.100263
• Sahu, S., Kothapalli, K., & Banerjee, D. S. (2024). Fast Leiden algorithm for community detection in shared memory setting. Proceedings of the 53rd International Conference on Parallel Processing, 11–20. https://doi.org/10.1145/3673038.3673146
• Scheidsteger, T., Haunschild, R., & Bornmann, L. (2025). How similar are field-normalized citation impact scores obtained from OpenAlex and three popular commercial databases? An empirical comparison based on large German universities. Scientometrics, 130(7), 3537–3569. https://doi.org/10.1007/s11192-025-05338-7
• Scherbakov, D., Hubig, N., Jansari, V., Bakumenko, A., & Lenert, L. A. (2025). The emergence of large language models as tools in literature reviews: a large language model-assisted systematic review. Journal of the American Medical Informatics Association, 32(6), 1071–1086. https://doi.org/10.1093/jamia/ocaf063
• Shamseer, L., Moher, D., Maduekwe, O., Turner, L., Barbour, V., Burch, R., Clark, J., Galipeau, J., Roberts, J., & Shea, B. J. (2017). Potential predatory and legitimate biomedical journals: Can you tell the difference? A cross-sectional comparison. BMC Medicine, 15(1), 28. https://doi.org/10.1186/s12916-017-0785-9
• Singh, V. K., Singh, P., Karmakar, M., Leta, J., & Mayr, P. (2021). The journal coverage of Web of Science, Scopus and Dimensions: A comparative analysis. Scientometrics, 126(6), 5113–5142. https://doi.org/10.1007/s11192-021-03948-5
• Small, H. (1973). Co-citation in the scientific literature: A new measure of the relationship between two documents. Journal of the American Society for Information Science, 24(4), 265–269. https://doi.org/10.1002/asi.4630240406
• Snyder, H. (2019). Literature review as a research methodology: An overview and guidelines. Journal of Business Research, 104, 333–339. https://doi.org/10.1016/j.jbusres.2019.07.039
• Thelwall, M., & Yaghi, A. (2025) Evaluating the predictive capacity of ChatGPT for academic peer review outcomes across multiple platforms. Scientometrics, 130, 5285–5307. https://doi.org/**10.1007/s11192-025-05287-1
• Torraco, R. J. (2005). Writing integrative literature reviews: Guidelines and examples. Human Resource Development Review, 4(3), 356–367. https://doi.org/10.1177/1534484305278283
• Torraco, R. J. (2016). Writing integrative literature reviews: Using the past and present to explore the future. Human Resource Development Review, 15(4), 404–428. https://doi.org/10.1177/1534484316671606
• Traag, V. A., Waltman, L., & van Eck, N. J. (2019). From Louvain to Leiden: Guaranteeing well-connected communities. Scientific Reports, 9, 5233. https://doi.org/10.1038/s41598-019-41695-z
• Tranfield, D., Denyer, D., & Smart, P. (2003). Towards a methodology for developing evidence-informed management knowledge by means of systematic review. British Journal of Management, 14(3), 207–222. https://doi.org/10.1111/1467-8551.00375
• Tricco, A. C., Lillie, E., Zarin, W., O’Brien, K. K., Colquhoun, H., Levac, D., Moher, D., Peters, M. D. J., Horsley, T., Weeks, L., Hempel, S., Akl, E. A., Chang, C., McGowan, J., Stewart, L., Hartling, L., Aldcroft, A., Wilson, M. G., Garritty, C., … Straus, S. E. (2018). PRISMA Extension for Scoping Reviews (PRISMA-ScR): Checklist and explanation. Annals of Internal Medicine, 169(7), 467–473. https://doi.org/10.7326/M18-0850
• van de Schoot, R., de Bruin, J., Schram, R., Zahedi, P., de Boer, J., Weijdema, F., Kramer, B., Huijts, M., Hoogerwerf, M., Ferdinands, G., Harkema, A., Willemsen, J., Ma, Y., Fang, Q., Hindriks, S., Tummers, L., & Oberski, D. L. (2021). An open source machine learning framework for efficient and transparent systematic reviews. Nature Machine Intelligence, 3(2), 125–133. https://doi.org/10.1038/s42256-020-00287-7
• van Eck, N. J., & Waltman, L. (2010). Software survey: VOSviewer, a computer program for bibliometric mapping. Scientometrics, 84(2), 523–538. https://doi.org/10.1007/s11192-009-0146-3
• van Eck, N. J., & Waltman, L. (2014). Visualizing bibliometric networks. In Y. Ding, R. Rousseau, & D. Wolfram (Eds.), Measuring scholarly impact: Methods and practice (pp. 285–320). Springer. https://doi.org/10.1007/978-3-319-10377-8_13
• Velden, T. (2022). Know thy tools! Limits of popular algorithms used for topic reconstruction. Quantitative Science Studies, 3(4), 1054–1086. https://doi.org/10.1162/qss_a_00217
• Verdonschot, R. G., Dilles, T., Cahir, C., De Graef, M., Holis, R. V., Fryden, J., … & Schor, M. (2026). To include or not to include? A prescription from the pharmacy on how to use active learning-assisted screening in systematic reviews. Systematic Reviews.
• Visser, M., van Eck, N. J., & Waltman, L. (2021). Large-scale comparison of bibliographic data sources: Scopus, Web of Science, Dimensions, Crossref, and Microsoft Academic. Quantitative Science Studies, 2(1), 20–41. https://doi.org/10.1162/qss_a_00112
• Westgate, M. J. (2019). revtools: An R package to support article screening for evidence synthesis. Research Synthesis Methods, 10(4), 606–614. https://doi.org/10.1002/jrsm.1374
• White, H., Albers, B., Gaarder, M., Kornør, H., Littell, J., Marshall, Z., Mathew, C., Pigott, T., Snilstveit, B., Waddington, H., & Welch, V. (2020). Guidance for producing a Campbell evidence and gap map. Campbell Systematic Reviews, 16(4), e1125. https://doi.org/10.1002/cl2.1125
• Whittemore, R., & Knafl, K. (2005). The integrative review: Updated methodology. Journal of Advanced Nursing, 52(5), 546–553. https://doi.org/10.1111/j.1365-2648.2005.03621.x
• Zipf, G. K. (1949). Human behavior and the principle of least effort: An introduction to human ecology. Addison-Wesley.
• Zupic, I., & Čater, T. (2015). Bibliometric methods in management and organization. Organizational Research Methods, 18(3), 429–472. https://doi.org/10.1177/1094428114562629