Mapping the global landscape of wearable devices in healthcare: A dual-database bibliometric analysis (2016–2026)

Abstract

Background

The escalating demand for proactive, personalized healthcare, heavily driven by the rising global burden of chronic diseases, necessitates a shift in modern medical management. While wearable devices are central to this transition, a comprehensive, multi-source quantitative mapping of the field’s evolutionary trajectory and application domains remains lacking.

Objective

This study systematically visualizes the knowledge structure, spatiotemporal distribution, and developmental trends of wearable healthcare technology over the past decade.

Methods

A bibliometric analysis was conducted using integrated data from Web of Science and Scopus (2016–2026). Utilizing CiteSpace and VOSviewer, we performed cooperation network analysis, co-citation clustering, and keyword burst detection to identify global collaboration patterns and research emerging trends.

Results

Analysis of 12,812 eligible articles identified China and the United States as leading contributors. Ten distinct research clusters emerged: (1) PM2.5; (2) breast milk; (3) chronic wounds; (4) COVID-19 pandemic; (5) Internet of medical things; (6) human activity recognition; (7) cancer survivors; (8) chronic disease; (9) water-soluble composite; (10) personalized health monitoring. Burst detection further highlighted an escalating shift toward telemedicine integration, cost-efficiency, and quality-of-life-focused rehabilitation.

Conclusion

This study presents a comprehensive bibliometric assessment of wearable health technology. By delineating established domains and emerging trajectories in telemedicine and personalized prevention, it provides evidence for researchers and policymakers to optimize resource allocation in digital health innovation.

1. Introduction

In recent years, the integration of wearable technology into healthcare has attracted considerable attention, reflecting a significant transition from reactive treatment to proactive health management. The escalating global burden of chronic diseases represents a critical healthcare challenge, with noncommunicable conditions accounting for approximately 74% of annual global deaths ¹ and consuming roughly 75% of healthcare expenditures in the United States alone. ² Consequently, while wearable technology spans a diverse spectrum of medical applications—ranging from infectious disease surveillance to maternal care—its evolutionary trajectory is fundamentally driven by this crisis. Driven by the urgent need to alleviate profound clinical and economic burdens, wearable devices facilitate a prominent conceptual shift toward proactive, personalized medicine through continuous physiological monitoring and timely remote interventions. ³

With the rapid advancement of sensor technology and the Internet of medical things (IoMT), ⁴ the volume of academic literature in this field has grown significantly. Previous studies have demonstrated the utility of wearables in specific domains, such as emergency response reliability ³ and disease prevention. To navigate this vast body of literature, researchers increasingly rely on bibliometric methods. These quantitative approaches leverage mathematical and statistical tools to objectively evaluate research performance, identifying structural relationships and thematic clusters that qualitative reviews may overlook.^5,6 Bibliometrics provides a transparent framework for mapping the intellectual landscape of scientific research and clarifying its integration with industrial practice.

However, despite the growing number of qualitative reviews and small-sample experiments, comprehensive quantitative analyses of this domain remain insufficient. While multi-database integration is standard practice in many bibliometric evaluations, several reviews specifically focusing on wearable technology remain limited to a single data source, such as the Web of Science. Neglecting the coverage provided by alternative databases like Scopus may result in an incomplete representation of cross-disciplinary collaborations. This limitation may result in a biased representation of global research trends and an incomplete understanding of cross-disciplinary collaborations. Furthermore, given the fast-moving nature of wearable technology, earlier analyses may fail to capture emerging hotspots such as COVID-19 applications or recent advancements in AI-driven rehabilitation.

Recently, several comprehensive bibliometric reviews have provided valuable foundational overviews of wearable technologies in healthcare, covering expansive timelines up to 2023. ⁷ While these seminal studies have successfully mapped the macroscopic technological themes, a critical methodological and analytical gap remains. First, the majority of these existing reviews rely heavily on single-database extractions (predominantly Web of Science), which inherently introduces coverage bias and often overlooks the engineering-focused research heavily represented in Scopus. Second, while prior works mapped structural themes, they frequently lack a dynamic predictive framework to distinguish between static research foundations and rapidly emerging socioeconomic clinical trends (e.g., healthcare cost reduction and telemedicine integration). To address this, our study not only integrates dual databases but uniquely employs Document Co-citation Analysis (DCA) to delineate deep structural clinical hotspots, ⁸ synergistically paired with keyword burst detection to dynamically forecast future trajectories. To address this gap, this study conducts a systematic bibliometric analysis of wearable device applications in healthcare over the past decade (2016–2026). To ensure a comprehensive dataset for this domain, this study integrates data from two major databases—the Web of Science Core Collection (WoSCC) and the Scopus. By utilizing visualization tools including CiteSpace and VOSviewer, this study aimed to: (1) map the global cooperation networks of authors, institutions, and countries; (2) identify current research hotspots through co-occurrence and co-citation analysis; and (3) predict emerging trends that will shape the future of healthcare. This study offers valuable references for researchers and practitioners seeking to understand the trajectory of this rapidly evolving field.

2. Materials and methods

2.1. Literature sources and retrieval strategies

This study utilized the WoSCC and Scopus as the primary data sources, covering a time span from January 1, 2016, to April 30, 2026. Subject terms were searched within the Medical Subject Headings (MESH) in PubMed, and search terms were refined based on expert knowledge. The search strategies employed have been formulated, with those for the Web of Science and Scopus being as follows (Table 1).

Table 1.

Search terms for wearable devices in the field of healthcare in WoSCC and Scopus.

Database	Search query
WoSCC	TS=(“wearable device*” OR “wearable technology” OR “wearable electronic*” OR biosensor*) AND TS=(healthcare OR “medical care” OR “health care”) NOT DT=(Retracted Publication) AND PY=(2016-2026) AND DT=(Article) AND LA=(English)
Scopus	(TITLE-ABS-KEY (“wearable devic*”) OR TITLE-ABS-KEY (“wearable technology”) OR TITLE-ABS-KEY (“wearable sensor*”) OR TITLE-ABS-KEY (“wearable electronic*”)) AND (TITLE-ABS-KEY (healthcare) OR TITLE-ABS-KEY (“medical care”) OR TITLE-ABS-KEY (“health care”)) AND PUBYEAR > 2015 AND PUBYEAR < 2027 AND LIMIT-TO (DOCTYPE, “ar”) AND LIMIT-TO (LANGUAGE, “English”)

2.1.1. Data collection and refinement

A systematic search across WoSCC and Scopus initially yielded 16,284 and 29,355 records, respectively (2016–2026). To minimize the risk of bias inherent in bibliometric literature searches, several systematic mitigation strategies were implemented during the data collection phase. Relying on a single database often skews bibliometric networks. By integrating data from both the WoSCC and Scopus, we significantly reduced coverage bias, capturing a more holistic global and interdisciplinary landscape. To ensure a comprehensive and unbiased mapping of the field, this study intentionally adopted a macro-level search strategy focusing purely on ‘wearables’ and ‘healthcare’. Specific clinical domains, such as chronic diseases or infectious diseases, were deliberately excluded from the initial search strings to avoid selection bias. Consequently, the highly specific medical applications identified in this study emerged organically as data-driven clusters rather than predefined targets.

A systematic search across the WoSCC and Scopus initially yielded 16,284 and 29,355 records, respectively. To ensure high evidentiary quality, we applied a strict screening protocol. During the screening and eligibility phase, 7,188 WoSCC records were excluded (1,292 outside the 2016–2026 timeframe; 5,838 non-article types; 58 non-English), yielding 9,096 eligible records. Concurrently, 20,691 Scopus records were excluded (2,696 outside the timeframe; 14,808 non-article types; 3,187 non-English), yielding 11,622 eligible records. The combined 20,718 eligible records were then subjected to our Python-based triple-verification deduplication algorithm, resulting in a final consolidated and deduplicated dataset of 12,812 unique articles for bibliometric analysis. All data retrieval across both databases was systematically conducted on May 24, 2026, applying a strict eligibility cutoff to exclusively include literature published up to April 30, 2026.

To ensure a comprehensive and unbiased mapping of the field, this study intentionally adopted a macro-level search strategy anchored specifically to the core concepts of ‘wearables’ and ‘healthcare’. This approach was deliberately chosen to facilitate bottom-up, data-driven mining rather than relying on top-down, subjective predefinition. Specifically, if researchers artificially predefine the search string with highly specific commercial devices (e.g., ‘smartwatch’, ‘fitness tracker’) or overarching digital health paradigms (e.g., ‘mobile health’, ‘remote patient monitoring’) based on prior knowledge, it inevitably introduces subjective selection bias and risks diluting the dataset with non-wearable hardware (such as standalone smartphone applications or traditional bedside monitors). By employing a focused macro-level query, we maintained a high signal-to-noise ratio. This unbiased baseline allowed highly specific clinical domains, novel device types, and shifting paradigms to organically emerge from the literature through objective co-citation clustering and keyword burst detection, thereby reflecting the true evolutionary trajectory of the field without predefined algorithmic interference.

2.1.2. Data integration and analysis

Data were exported in specialized formats (Plain Text for WoS; CSV for Scopus) containing full records and cited references. To ensure cross-database compatibility, data standardization and deduplication were conducted using CiteSpace (v6.4.R1) for WoS records and for Scopus records. The refined dataset was then subjected to bibliometric mapping via CiteSpace and VOSviewer, focusing on co-citation networks and emergent trend detection. The selection process is summarized in the PRISMA-compliant flow diagram (Figure 1).

Figure 1.

Literature screening process.

2.2. Research method

2.2.1. Data extraction

This study employs CiteSpace and VOSviewer, two visualization analysis software tools, in conjunction with Excel, to conduct a comprehensive bibliometric analysis of the acquired academic outputs. The visualization analysis constructs a multidimensional knowledge network, encompassing word frequency distribution, co-occurrence clustering, as well as citation network analyses across institutional, authorial, and terminological dimensions. This analytical framework systematically delineates the landscape of knowledge leadership while identifying temporal evolution patterns of wearable devices in the healthcare domain.

2.2.2. Data organization and analysis

The data retrieved from the WoSCC was exported in plain text file format, with filenames adopting the “download_***” format. The data obtained through searches in the Scopus database was exported in CSV format. Subsequently, CiteSpace 6.4.R1 was employed for visual analysis and processing.

2.2.3. Data integration and standardization

Adopting the BibexPy framework, ⁹ we implemented a systematic data integration pipeline involving multi-source format standardization via Python scripting and a triple-verification deduplication algorithm.

Specifically, the triple-verification deduplication algorithm prioritized unique identifiers sequentially. First, an exact match was performed using the Digital Object Identifier (DOI). For records lacking a DOI, a secondary fuzzy match was executed by concatenating the normalized article title (converted to lowercase with punctuation and stop words removed), publication year, and the first author’s last name. Finally, a tertiary contextual verification cross-referenced the journal title abbreviation and volume/issue/page numbers to confirm uniqueness. The complete scripts and deduplication protocols are accessible via the aforementioned framework publication, and the specific Python source code utilized for this procedure is publicly available at https://github.com/bcankara/BibexPy.

2.2.4. Bibliometric analysis settings

The parameter settings in CiteSpace were as follows: (i) Based on the time span of the 12,812 retrieved articles, the time span in CiteSpace was set from January 2016 to April 2026; (ii) For the term source, the title, abstract, author keywords, and Keyword Plus were selected for analysis; (iii) CiteSpace conducts analysis based on various node selections. Nodes for network analysis include keywords, subject terms, and categories. Nodes for collaboration network analysis include authors, institutions, and countries. Nodes for co-occurrence network analysis include keywords, subject terms, and categories. Node types include authors, institutions, countries, keywords, categories, references, cited authors, and cited journals; (iv) To observe emerging trends over a longer time span and minimize the impact of short-term fluctuations, the Years Per Slice setting was set to 2 years, reducing the fragmentation of data caused by overly fine slicing; (v) For the top N selection criteria, the top 50 most frequently cited or occurring items from each time slice were selected; (vi) Pruning methods included Minimum Spanning Tree and Pruned Sliced Network. Other settings were left as default. Since keywords better reflect the main focus of the articles, this study selected co-occurrence network analysis based solely on keywords. Node types included authors, institutions, countries, keywords, references.

It is important to distinguish between the two primary analytical dimensions used in this study: DCA was employed to delineate static, structural thematic domains, encompassing specific clinical contexts (e.g., COVID-19) and environmental health factors (e.g., fine particulate matter, PM_2.5), whereas Kleinberg’s burst detection algorithm was applied to keywords to track the sudden shifts and emerging trends in research focus (e.g., surges in telemedicine integration) that often drive these structural domains. ¹⁰

3. Results

3.1. Publication trajectories and growth dynamics

As illustrated in Figure 2, the annual output of publications from 2016 to 2025 exhibits a robust and consistent upward trajectory, demarcated into three evolutionary phases. The Incipient Phase (2016–2017) was characterized by a modest volume (<500 articles/year), reflecting early-stage exploration. This was followed by a Rapid Growth Phase (2018–2020), where annual publications surged from 647 to 992, likely catalyzed by technological maturation and academic integration. The Steady Expansion Phase (2021–2025) demonstrated sustained momentum, rising from 1,192 in 2021 to a peak of 2,498 in 2025, signaling the field’s progression toward maturity. This decade-long escalation underscores the enduring academic priority of this domain. Based on the current trajectory, research productivity is projected to maintain its upward momentum, reinforcing its status as a central pillar of innovation. This optimistic projection is already being validated by preliminary data from 2026. With 1,076 publications recorded in just the first half of the year, the annual output is on track to significantly surpass previous records, indicating that the field is entering an accelerated phase of profound development.

Figure 2.

Annual publication trend of research on current status.

Note. The x-axis maps the developmental timeline of the field (2016–2026), while the y-axis quantifies the annual publication output as a metric of academic engagement and research scale. The solid line traces the empirical growth trajectory (2016–2025), with its steepening slope highlighting exponential expansion and sustained academic momentum. The trendline projects future trajectories based on recent historical patterns, indicating continued robust productivity and demonstrating that the field remains in a pre-plateau phase of rapid expansion.

3.2. Analysis of authors

Using authors as nodes, this study analyzes bibliographic data from the WoSCC and Scopus to construct visual representations of prolific authors’ networks (Figure 3 and Table 2). The author network in wearable-healthcare research (2016-2026) demonstrates a fragmented yet structured topology, featuring multiple densely connected micro-clusters that collectively drive methodological innovation and clinical translation. These visualizations quantitatively capture publication volume, citation influence, and collaborative patterns among key researchers, providing systematic insights into the field’s developmental trajectory. Based on the updated bibliometric data, Wang Wei and Wan Peng Bo exhibit outstanding performance in the dataset. Wang Wei is identified as the most prolific author, while Wan Peng Bo ranks among the top scholars in both publication volume and citation impact, reflecting their significant contributions to this field. To address the deeper technological aspects of wearable devices, an in-depth analysis of the top three most-cited authors reveals the specific types of wearable technologies, their targeted diseases, clinical roles, as well as their advantages and limitations.

Figure 3.

Visualization of authors in the application of wearable devices in the healthcare field.

Note. The visualization was constructed using VOSviewer. Each node represents an individual author. The connecting lines (edges) denote co-authorship relationships, illustrating the presence and pathways of academic collaboration. Nodes sharing the same color belong to a unified collaborative cluster or research community, indicating tightly-knit scholarly networks.

Table 2.

Top 10 most productive and highly cited authors, institutions, and countries.

Rank	Authors	Counts	Authors	Citations	Institution	Counts	Institution	Citations	Country	Counts
1	Wang, Wei	45	Wan, Peng Bo	5452	Chinese Academy of Sciences	344	Chinese Academy of Sciences	20302	China	3446
2	Li, Yang	31	Wang, Zhong Lin	4386	University of California	194	University of California	8435	USA	2765
3	Wang, Zhong Lin	29	Gao, Wei	4029	Zhejiang University	136	Tsinghua University	7441	India	1712
4	Yeo, Woon-Hong	25	Chen, Jun	4011	Tsinghua University	127	Stanford University	7074	South Korea	1071
5	Chen, Jun	24	Zhang, Li Qun	3169	University of Electronic Science and Technology of China	110	Georgia Institute of Technology	6506	United Kingdom	1024
6	Wan, Peng Bo	23	Javey, Ali	2587	Imperial College London	105	National University of Singapore	6411	Italy	562
7	Li, Wei	23	Lee, Nae-eung	2560	King Saud University	103	Beijing University of Chemical Technology	6284	Saudi Arabia	545
8	Huang, Wei	23	Takei, Kuniharu	2533	Stanford University	102	Sungkyunkwan University	4755	Australia	509
9	Zhang, Yan	22	Lee, Cheng Kuo	2261	Vellore Institute of Technology	101	Northwestern University	4711	Canada	452
10	Wang, Yan	22	Xu, Feng	2190	National University of Singapore	99	University of Electronic Science and Technology of China	4690	Germany	407

Representing flexible epidermic sensors and on-demand therapy, Wan Peng Bo utilizes nanomaterials such as MXene to predict early-stage Parkinson’s disease by monitoring tremors and manages chronic wounds using thermoresponsive nanomesh for triggered antibiotic release. ¹¹ These flexible sensors overcome the impermeability and mechanical mismatch of conventional substrates by providing ultrahigh sensitivity and breathability. Advancing the field of self-powered active motion sensors, Wang Zhong Lin integrates triboelectric nanogenerators into smart textiles to prevent muscular disorders through posture correction and support elderly care with fall alarms. ¹² In the visual knowledge map of co-authors depicted in Figure 3, the size of the author nodes is proportional to their publication output, while the connecting lines represent collaborations between authors. Significant collaborative networks among several researchers have been identified. Prominent collaboration networks, such as those linking Gao, Wei, Rogers, John A., and Wang, Joseph, alongside the active partnership between Chen, Jun, Kumar, Santosh, and Wang, Zhong Lin, underscore the robust synergies within this academic community. While these nanogenerators sustainably harvest biomechanical energy to eliminate battery dependence, Wang’s team further resolved previous limitations of low power and bacterial contamination by developing 3D orthogonal woven structures and antibacterial composite films. ¹³ Furthermore, Gao Wei drives the transition toward biochemical body-fluid analysis by developing wearable sweat sensors for the noninvasive diagnosis of cystic fibrosis and chip-less neuromorphic systems for rapid sepsis detection. ¹⁴ Although these platforms drive a significant transition toward continuous precision healthcare, they still face substantial challenges regarding reliable continuous sweat sampling and ensuring accurate correlations between sweat and blood analyte concentrations. ¹⁵

3.3. Analysis of institution

The top five most frequently top productive institutions in the field concerning wearable devices’ applications in healthcare are identified as follows: Chinese Academy of Sciences, University of California, Zhejiang University, Tsinghua University, University of Electronic Science and Technology of China (Table 2). These institutions serve as crucial nodes within the global cooperation network, actively engaging in extensive collaborative endeavors not only among themselves but also with other international counterparts. They have been at the forefront of research and development efforts focused on the application of wearable devices in healthcare.

To move beyond descriptive bibliometric statistics and delve into the technological core of wearable devices, an in-depth analysis of the leading institutions reveals distinct technological focuses, targeted diseases, clinical roles, and their respective advantages and limitations. Representing the advancement of biochemical monitoring, Zhejiang University focuses on developing minimally invasive continuous glucose monitoring (CGM) systems specifically targeted at closed-loop diabetes management. By synergizing organic electrochemical transistors (OECT) and microneedle arrays, these devices effectively amplify biochemical signals with high signal-to-noise ratios while minimizing pain during continuous subcutaneous sampling 1116 and 17; however, long-term continuous usage still faces limitations regarding biofouling and the stability of the skin-device hydrogel interface.

In the realm of physical monitoring, the Chinese Academy of Sciences focuses on developing 3D porous flexible piezoresistive sensors utilizing carbon nanomaterials (such as carbon black and multi-walled carbon nanotubes). These flexible tactile sensors are applied in artificial skin and long-term healthcare systems to monitor vital physical signs and detect subtle physiological anomalies. ¹⁸ While they overcome the rigidity of conventional sensors by providing high sensitivity and a broad sensing range, ensuring stable electrical performance under extreme or repetitive mechanical deformation remains a primary challenge.

Furthermore, reflecting the clinical implementation of wearables, institutions like the University of California, emphasize the integration of wearable information and communication technologies (ICTs) for mental health and chronic disease management. These devices track lifestyle modifications, such as sleep quality and physical activity, serving as crucial adjuvant treatments for mental disorders like depression in primary care settings. ¹⁹ Although these ICT-based wearables offer the advantage of objective, continuous data collection for non-pharmacological interventions, they continue to encounter real-world limitations concerning long-term patient adherence, data privacy, and algorithmic biases.

In VOSviewer, institution clustering fundamentally aggregates entities characterized by intensive collaboration, thematic proximity, and robust regional or academic linkages into cohesive groups distinguished by uniform color coding. These clusters delineate academic communities or collaborative blocs within the research landscape. ²⁰ The visualization analysis diagram of these institutions (Figure 4) vividly illustrates the extensive cooperative relationships they have forged both domestically and internationally. Within the institutional collaboration network map generated by VOSviewer, the red cluster represents the largest cooperative consortium, centered on Chinese Academy of Sciences. Comprising 188 collaborating institutions, this cluster constitutes the most extensive and densely interconnected unit in the network, both in terms of institutional count and partnership intensity. This widespread collaboration highlights the global recognition and attention that the research on wearable devices’ applications in healthcare has garnered, indicating a promising future for this innovative field.

Figure 4.

Visualization of institutions in the application of wearable devices in the healthcare field.

Note. The visualization was generated using VOSviewer to map the global institutional collaboration landscape. Each node represents an individual research institution. The connecting lines (edges) indicate co-authorship relationships between institutions, with shorter distances and denser connections reflecting stronger collaborative ties. Institutions sharing the same color are grouped into a unified collaborative cluster, denoting a closely-knit international or regional research community.

3.4. Analysis of country

In the field of healthcare research on wearable devices, researchers from various countries (Figure 5) have demonstrated an active spirit of exploration. Table 2 lists the number of published papers from the top ten countries. In research based on wearable devices within the healthcare sector, scholars worldwide have demonstrated robust research vitality. This methodological framework enables systematic identification of key contributors, institutional leadership patterns, and emerging research trajectories while optimizing collaborative opportunities. ²⁰ Although our initial literature retrieval strategy was globally inclusive and executed without any geographic restrictions, the bibliometric results empirically indicate that specific regions—most notably China and the United States—account for a disproportionately large share of the publications. Nevertheless, scholars worldwide have demonstrated robust research vitality, actively participating in the extensive collaborative networks (e.g., Cluster 1 in Figure 5). China exhibits leading research output (3,446 publications), reflecting robust institutional capacity and strategic R&D investment. China predominantly leads in the hardware innovation of self-powered advanced materials, specifically developing biofuel-induced electricity generators (BEGs) and flexible e-skins for continuous physiological and metabolic biomarker tracking. These biosensors offer the significant advantage of eliminating external batteries through metabolic energy harvesting and ensuring high biocompatibility; however, their clinical translation remains limited by the requirement for continuous sweat production to maintain stable power outputs. ²¹ The United States maintains second place with 2,756 publications. Conversely, the United States demonstrates a strong clinical and translational orientation, extensively utilizing wrist-worn actigraphy and smart accelerometry for the continuous management of neurological and neuropsychiatric disorders, such as focal epilepsy. These devices play a crucial role in monitoring 24-hour rest-activity rhythms (RARs) to objectively assess seizure severity and associated comorbidities like depression and sleepiness. While they provide the advantage of continuous, real-world biophysical tracking, their limitation lies in capturing only motion data rather than direct underlying neurological activity. ²² Notably, India, South Korea, and the United Kingdom demonstrate surpassing the 1000-publication threshold, collectively contributing significantly to global technological advancement in wearable healthcare solutions. Furthermore, research from India heavily emphasizes the public health and epidemiological integration of digital health wearables, particularly accelerated by the COVID-19 pandemic. In this context, wearables serve a pivotal role in remote patient monitoring and decentralized healthcare delivery to alleviate overwhelmed clinical systems. Although these technologies offer scalable solutions for infectious disease management and rapid triage, their widespread implementation in such regions is still significantly hindered by systemic socioeconomic inequities, digital literacy gaps, and infrastructure limitations. ²³

Figure 5.

Global distribution of the application of wearable devices in the healthcare field.

Note. Figure 5 Global collaboration network of countries/regions contributing to wearable healthcare technology research. Node size is proportional to the total publication yield of each respective country. The thickness of the connecting lines reflects the strength of international collaborative partnerships, while distinct node colors delineate specific collaborative clusters.

3.5. Analysis of co-citation reference

Co-citation reference analysis provides a rigorous quantitative framework for delineating the intellectual architecture and evolutionary trajectory of a scientific domain by evaluating the co-occurrence frequency of pivotal publications, authors, and journals. ²⁴ And research hotspots can be identified through co-citation analysis. Clustering is a practical and powerful data analysis method that can reflect the research themes within a specific field. It has a wide range of applications, and continuous research efforts are constantly enhancing its efficiency and effectiveness in handling complex high-dimensional data. ²⁵ From the data presented in the figures, Figure 6 shows a Q value of 0.8194 and an S value of 0.9195. These values indicate significant clustering with a highly reliable structure. The Q and S values from CiteSpace’s bibliometric analysis assist researchers in assessing the structure and clustering validity of knowledge graphs, thereby enhancing the understanding of knowledge structure and trends within the research area. Q values range from 0 to 1, with higher values indicating stronger community formation among nodes in a network. S values close to 1 suggest that the clustering results are robust, meaning that data points within each cluster are cohesive and distinct from those in other clusters. Through visualization analysis, we identified a total of 10 distinct clusters. The cluster labels in Figure 6 are as follows: # 0 pm2.5, #1 breast milk, #2 chronic wound, #3 COVID-19 pandemic, #4 Internet of medical things, #5 human activity recognition, #6 cancer survivors, #7 chronic disease, #8 water-soluble composite, #9 personalized health monitoring.

Figure 6.

Co-citation map of research literature on wearable devices in the healthcare field.

Note. Figure 6 Document co-citation network and clustering analysis. Distinct clusters delineate cohesive thematic sub-domains within the research field. The connecting lines (edges) represent co-citation relationships, illustrating the strength of intellectual affinity between publications. Nodes highlighted in red possess high betweenness centrality, acting as pivotal intellectual bridges that facilitate knowledge diffusion across disparate thematic clusters. The network evaluation metrics (upper-left corner) validate the structural robustness of the clustering analysis. A Modularity Q value (Q)> 0.3 indicates a significant community structure, while a Mean Silhouette score (S) > 0.7 corroborates the high internal homogeneity and reliability of the delineated clusters.

3.6. Analysis of keywords burstness

3.6.1. Column definitions

‘Year’ indicates the initial year the keyword first appeared within the analyzed dataset; ‘Begin’ denotes the specific year when the sudden surge (burst) in the keyword’s usage frequency commenced; ‘End’ represents the year this burst phase concluded (a value of 2026 indicates the burst remains ongoing up to the end of the analyzed timeframe). ‘Strength’ represents the algorithmically calculated intensity of the burst.

3.6.2. Visual elements

The blue line represents the overall time interval of the study period. The solid red blocks visually denote the specific duration of the citation burst, reflecting a sudden and intense surge in academic attention for that particular keyword. Keywords are sorted chronologically by their ‘Begin’ year.

4. Discussion

4.1. General information

While previous bibliometric analyses have laid important groundwork utilizing dual-database approaches, the current study builds upon these foundations by extending the temporal scope to 2026, thereby capturing the latest technological convergences.^26,27

The past decade has witnessed a burgeoning body of literature regarding wearable devices in healthcare, reflecting sustained global interest. A pivotal surge occurred in 2020, with publications reaching 992—a trend likely catalyzed by the urgent demand for remote physiological monitoring during the COVID-19 pandemic. This inflection point underscores the clinical significance of wearables. Furthermore, the integration of mobile health technology has enhanced the scalability and accessibility of these devices, significantly broadening the research landscape. Notably, research productivity has maintained a distinct upward trajectory since 2021 (1,192 papers), ultimately peaking at 2,498 publications in 2025, with current trajectories indicating an ongoing and robust expansion.

Leading contributors include Wang Wei (n=45) and Li Yang (n=31). Wang Wei’s work predominantly delineates the technological infrastructure requisite for transitioning from “passive treatment” to “proactive health management.” Conversely, Li Yang’s principal research focuses on the development of novel flexible, wearable, multimodal biosensors and advanced functional materials, aiming to advance their frontier applications in personalized continuous health monitoring, adjunctive diagnosis of major pathologies, and point-of-care testing. To elucidate the technological paradigms of wearable devices, an analysis of the top-cited authors—Wan Peng Bo (n = 5,452) and Wang Zhong Lin (n = 4,386)—delineates key modalities and clinical utilities. Wan primarily develops MXene-based flexible epidermic sensors that overcome conventional impermeability and mechanical mismatch, highlighting Parkinson’s prediction via tremor monitoring and thermoresponsive chronic wound management. Conversely, Wang drives advancements in self-powered sensors by integrating triboelectric nanogenerators into smart textiles. By sustainably harvesting biomechanical energy and utilizing antibacterial architectures, this approach circumvents battery dependence to support geriatric care and muscular disorder prevention.

Geographically, China and the United States dominate the research landscape, complemented by a robust tier of nations—such as India, South Korea, and the UK, each exceeding 1,000 publications—whose leading institutions are delineated by network analysis into five major, highly interconnected clusters. Such extensive transnational partnerships actively facilitate the cross-pollination of multidisciplinary expertise, establishing a resilient global research ecosystem that effectively accelerates the clinical translation of wearable innovations.

The emergence of chronic disease management and infectious disease control (notably COVID-19) as dominant clusters validates the primary clinical utility of wearable devices. Importantly, these specific domains emerged organically from our broad search strategy, objectively reflecting the scientific community’s natural pivot toward addressing high-burden global health challenges without the interference of predefined search bias.

4.2. Research hotspots

4.2.1. Wearable devices in healthcare: Applications in COVID-19 management

As detailed in the co-citation network analysis (Table 3 and Figure 6), Cluster #3 (“COVID-19 pandemic”) emerges as a pivotal specialized research domain, comprising a size of 63 core nodes with a highly cohesive Silhouette score of 0.941 and an average publication year of 2021. Representing a critical intellectual turning point, the advent of the pandemic catalyzed a profound transition in healthcare delivery, accelerating the integration of wearable devices for continuous, remote patient monitoring. As outlined in Table 4, which details clinical applications across key domains, wearable medical devices are integral to disease management and intervention. During the COVID-19 pandemic, for example, the effective integration of core sensors like smartwatches and fitness trackers enabled the creation of robust remote patient monitoring systems. This approach not only mitigated viral transmission risks but also leveraged synergistic artificial intelligence analysis to achieve early prediction of clinical deterioration.

Table 3.

Summary of key bibliometric metrics for major clusters.

Cluster ID	Size	Silhouette	Label (LSI)	Label (LLR)	Label (MI)	Average year
#0 pm2.5	93	0.889	pm2.5	fine particulate matter (528.03, 1.0E-4)	air pollution (0.3)	2016
#1 breast milk	70	0.914	breast milk	human milk (515.37, 1.0E-4)	lactation (0.42)	2016
#2 chronic wound	68	0.952	chronic wound	wound healing (557.26, 1.0E-4)	diabetic ulcer (0.34)	2017
#3 COVID-19 pandemic	63	0.941	COVID-19 pandemic	SARS-CoV-2 (638.55, 1.0E-4)	coronavirus (0.82)	2021
#4 Internet of medical things	61	0.909	Internet of medical things	healthcare IoT (697.02, 1.0E-4)	smart healthcare (0.42)	2021
#5 human activity recognition	57	0.885	human activity recognition	motion tracking (567.77, 1.0E-4)	action recognition (0.37)	2019
#6 cancer survivors	55	0.915	cancer survivors	quality of life (394.22, 1.0E-4)	oncology (0.52)	2018
#7 chronic disease	52	0.925	chronic disease	disease management (433.51, 1.0E-4)	comorbidities (0.4)	2020
#8 water-soluble composite	48	0.913	water-soluble composite	biodegradable polymer (429.36, 1.0E-4)	water solubility (0.43)	2021
#9 personalized health monitoring	44	0.88	personalized health monitoring	point-of-care testing (288.14, 1.0E-4)	precision medicine (0.49)	2016

Table 4.

Core clusters and clinical applications of wearable healthcare devices.

Cluster domain	Target population/disease	Core sensor types	Clinical & intervention roles	Current technical/Clinical limitations
Cluster #0: Environmental Exposure & PM2.5 Monitoring (PM2.5)	Patients with chronic lung and respiratory conditions.	Wearable PM2.5 environmental sensors; integrated with physiological tracking tools (e.g., heart rate and oxygen saturation monitors).	Real-time assessment of individual particulate exposure in dynamic environments; generation of fine-grained geospatial data to proactively minimize exposure to ambient pollutants; empowerment of multimodal remote patient care.	Requires rigorous calibration against research-grade instruments to ensure diagnostic data reliability; inherent power constraints necessitate innovative engineering solutions to extend battery life.
Cluster #1: Breastfeeding Support (Breast Milk)	Postpartum nursing mothers and breastfed infants.	Smart nursing garments integrated with accelerometers and flexible bending sensors; lactation pads with microfluidic channels and laser-induced graphene sensors.	Real-time postural monitoring to mitigate musculoskeletal pain and improve postural alignment; rapid quantitative detection of pharmacological residues (e.g., acetaminophen) in breast milk to prevent toxicity.	Systemic barriers persist regarding data accuracy, patient privacy, device usability, and seamless integration with mobile applications.
Cluster #2: Chronic Wound Management (Chronic Wound)	Patients with non-healing chronic wounds.	Smart wound dressings embedded with microelectronic and electrochemical sensors, alongside localized temperature and moisture detectors.	Continuous, real-time monitoring of physiological parameters within the wound microenvironment; autonomous triggering of on-demand pharmacological release or targeted electrical stimulation via closed-loop bioelectronic systems to accelerate tissue regeneration.	Sensitive biosensors lack long-term biochemical stability in highly corrosive, enzyme-rich exudates; critical need for standardized datasets and robust clinical validation frameworks.
Cluster #3: COVID-19 & Infectious Diseases (COVID-19 Pandemic)	Infected individuals, isolated patients, and individuals with Long COVID.	Smartwatches, fitness trackers, and smart masks (monitoring heart rate, SpO2, body temperature, heart rate variability, and sleep patterns).	Facilitates robust remote patient monitoring to minimize viral transmission risks while ensuring high-quality care; synergizes with AI to analyze voluminous physiological datasets for early prediction of clinical deterioration.	Profound challenges pertaining to data security, privacy protection, and structural interoperability across healthcare systems; optimal efficacy relies on overcoming regulatory hurdles and improving patient engagement.
Cluster #4: IoMT	Patients requiring remote management for chronic morbidities (e.g., cardiovascular/respiratory diseases) and general consumer health trackers.	Clinical-grade biomonitors (ECG, precise blood pressure, continuous glucose monitors); consumer-grade lifestyle trackers.	Shifts the paradigm from passive data collection to predictive analytics and preemptive clinical interventions via AI; manages massive monitoring datasets and mitigates energy demands through edge computing.	Ubiquitous deployment is frequently impeded by data privacy vulnerabilities, structural interoperability, and stringent regulatory compliance mandates (often necessitating blockchain integration).
Cluster #5: Human Activity Recognition	General populations requiring continuous tracking of physical activity patterns and functional mobility.	Inertial measurement units (IMUs) and flexible stretch sensors, concurrently integrated with wireless electrocardiograms.	Translates raw sensor data into actionable insights regarding physical activity and postural deviations; automates deep feature extraction using deep learning architectures to enhance motion analysis accuracy.	Computationally intensive algorithms present substantial energy bottlenecks on resource-constrained wearables; overall generalizability is constrained by user variability, requiring user-centered data acquisition methodologies.
Cluster #6: Oncology Survivorship Care (Cancer Survivors)	Post-treatment oncology populations (cancer survivors).	Pedometers, fitness trackers, and sophisticated cardiovascular monitors.	Elevates moderate-to-vigorous physical activity to improve physical function and quality of life; captures nuanced cardiovascular data for early detection of comorbidities (e.g., atrial fibrillation); longitudinally monitors psychosocial symptoms like anxiety and depression.	Complex challenges regarding long-term patient adherence; necessitates rigorous regulatory standardization and robust data security protocols to ethically protect patient privacy.
Cluster #7: Chronic Disease Management (Chronic Disease)	Patients with chronic conditions such as diabetes, cardiovascular disorders, and hypertension.	Sensors providing continuous, real-time monitoring of vital physiological parameters (blood pressure, glucose levels, heart rate).	Functions as a proactive early-warning system via AI and physiological tracking; empowers clinicians to make dynamic, data-driven therapeutic adjustments; intrinsically motivates patients to adhere to rehabilitation protocols and cultivate healthier lifestyles.	Full-scale integration is impeded by sociotechnical barriers, including pervasive concerns regarding data privacy and the initial economic burden of the devices.
Cluster #9: Personalized Health Monitoring (Personalized Health Monitoring)	General populations requiring precise medical diagnostics, alongside patients with chronic cardiovascular and metabolic diseases.	Implantable devices and miniature biosensors (monitoring physiological vital signs such as heart rate, blood pressure, and blood glucose).	Facilitates continuous, real-time monitoring of multiparameter vital signs for early anomaly detection and preventive intervention; leverages deep AI and machine learning integration to offer predictive analytics and customized therapeutic recommendations.	Substantial sociotechnical bottlenecks regarding data privacy and security vulnerabilities, questionable accuracy/reliability of sensor data, and interoperability barriers in seamlessly integrating with Electronic Health Record (EHR) systems.

Catalyzed by the COVID-19 pandemic, the profound synergy between wearable sensors and telemedicine platforms has driven a significant transition in healthcare delivery, establishing robust remote patient monitoring systems that enable continuous tracking and early detection of critical vital signs, thereby successfully reconciling rigorous infection control with high-quality isolated clinical care. ²⁸ Beyond the acute phase of infection, wearables equipped with physiological sensors play a pivotal role in understanding and managing Long COVID by tracking nuanced metrics like sleep patterns and heart rate variability. ²⁹ The integration of advanced artificial intelligence has transformed wearables from passive monitors into proactive diagnostic platforms that analyze multidimensional physiological data to predict clinical deterioration and facilitate personalized, preemptive interventions. ³⁰ However, the ubiquitous deployment of these digital health tools is not without significant impediments. The medical community must navigate profound challenges pertaining to data security, privacy protection, and the interoperability of disparate healthcare systems. ³¹ Consequently, while AI-enhanced wearables hold substantial potential for mitigating future public health crises, their optimal efficacy is contingent upon resolving complex regulatory hurdles and improving overall patient engagement.

4.2.2. Wearable devices in healthcare: Integration with the IoMT

As detailed in the co-citation network analysis (Table 3 and Figure 6), Cluster #4 “IoMT” emerges as a pivotal specialized research domain, comprising a size of 61 core nodes with a highly cohesive Silhouette score of 0.909 and an average publication year of 2021. As outlined in Table 4, wearable devices within IoMT framework have established a patient-centric care continuum by integrating clinical-grade biomonitors with consumer health trackers. Powered by artificial intelligence and edge computing, this synergy efficiently manages vast continuous data streams, reflecting a significant transition from passive data collection toward predictive analytics and preemptive clinical interventions for chronic disease management. Although literature frequently groups IoMT-integrated wearable devices together based on shared technological infrastructures, it is critical to delineate this broad ecosystem into two distinct functional sub-domains: clinical-grade biomonitoring and consumer-grade lifestyle tracking.

The first distinct domain within the IoMT involves clinical-grade biomonitoring. Equipped with advanced, highly calibrated bio-sensors, these systems capture continuous, high-fidelity physiological metrics—such as electrocardiograms (ECG), precise blood pressure, and continuous blood glucose levels. ³² These devices are subject to stringent regulatory compliance and are deployed specifically for the rigorous remote management of chronic morbidities (e.g., cardiovascular and respiratory diseases) and preemptive clinical interventions. ³³

Conversely, the second domain encompasses consumer-grade lifestyle and health trackers. These wearables focus on longitudinal behavioral metrics, such as sleep patterns, daily step counts, and basic heart rate trends. ³⁴ While they may not meet the diagnostic thresholds of clinical biomonitors, they serve a fundamentally different, yet equally vital, clinical purpose: patient empowerment, behavior modification, and preventative wellness outside traditional clinical settings. ³⁵ Ultimately, the true value of the IoMT ecosystem lies in its ability to act as a bridge—securely aggregating and processing both streams of data (clinical and lifestyle) via AI to create a holistic, patient-centric healthcare continuum.

Concurrently, by embedding AI within these networks, modern platforms transcend rudimentary tracking to facilitate predictive analytics and preemptive clinical interventions. ³⁶ To manage the large-scale datasets generated by constant monitoring and mitigate the intensive energy demands of off-device processing, contemporary IoMT architectures are rapidly shifting toward edge computing. ³⁷ Despite these profound architectural advancements, the ubiquitous deployment of intelligent healthcare systems is frequently impeded by systemic vulnerabilities regarding data privacy, structural interoperability, and stringent regulatory compliance mandates. ³⁸ In direct response to these critical security imperatives, decentralized ledger technologies, notably blockchain, are being rigorously integrated to establish secure, immutable communication protocols across heterogeneous IoMT networks. ³⁹ Ultimately, as these synergistic innovations mature alongside next-generation 5G and 6G telecommunications, they promise to significantly augment data transfer capabilities, reflecting the definitive realization of highly resilient smart hospital environments and interconnected global care ecosystems.

4.2.3. Wearable devices in healthcare: Innovations in chronic disease management

As evidenced by network metrics (Table 3 and Figure 6), Cluster #7 (‘chronic disease’) forms a highly cohesive research domain (size = 52, Silhouette = 0.925, average year = 2020). While Figure 6 illustrates the robust temporal growth of literature concerning chronic diseases, a deeper mechanistic analysis reveals how wearable technologies drive this significant clinical transition across three critical pillars: prevention, treatment, and longitudinal management. As demonstrated by the clinical applications in chronic disease management outlined in Table 4, wearable devices have emerged as a cornerstone for disease prevention and long-term follow-up by facilitating the continuous, real-time monitoring of vital physiological parameters, including blood pressure, glucose levels, and heart rate. By integrating artificial intelligence, these advanced sensing systems not only establish highly efficient clinical early-warning mechanisms but also empower healthcare teams to overcome the latency inherent in traditional intermittent clinical visits, thereby highlighting dynamic, data-driven, and precise therapeutic adjustments.

Crucially, these devices provide continuous, real-time monitoring of vital physiological parameters—such as blood pressure, glucose levels, and heart rate—which are essential for managing chronic conditions like diabetes and cardiovascular disorders. ⁴⁰ By utilizing advanced sensors and artificial intelligence, wearables can detect subtle physiological anomalies and provide personalized health recommendations, thereby reducing the reliance on traditional intermittent clinical assessments.^41,42 Furthermore, these digital tools actively empower patient self-management by fostering healthier lifestyle behaviors; for instance, high-frequency wearable use has been shown to significantly improve physical activity levels, especially in patients with hypertension. ⁴³ The delivery of timely, non-judgmental feedback and personalized advice through these platforms fundamentally enhances patients’ self-efficacy and their adherence to long-term rehabilitation regimens.^44,45 From a clinical perspective, the integration of wearable data with IoMT and telemedicine frameworks enables robust remote patient monitoring, allowing healthcare providers to intervene proactively from a distance.^46,47 This patient-centered, data-driven approach not only optimizes personalized care plans but also effectively reduces emergency room visits and hospital readmissions, ultimately lowering the total economic burden on healthcare systems while substantially enhancing the health-related quality of life. ^41,47,48

4.2.4. Wearable devices in healthcare: Enhancing cancer survivorship care

As detailed in the co-citation network analysis (Table 3 and Figure 6), Cluster #6 (“cancer survivors”) emerges as a pivotal specialized research domain, comprising a size of 55 core nodes with a highly cohesive Silhouette score of 0.915 and an average publication year of 2018. It is important to note that while Cluster #6 (‘Cancer Survivors’) designates a user demographic rather than a disease pathology, it emerged algorithmically as a distinct structural domain. This highlights that the application of wearables in this cohort extends far beyond general consumer fitness tracking. As detailed by the clinical applications for cancer survivorship care outlined in Table 4, the deployment of wearable telemetry has successfully shifted the oncological clinical paradigm from episodic follow-up evaluations to continuous physiological and behavioral monitoring. By leveraging pedometers and advanced cardiovascular monitors integrated with artificial intelligence, this approach not only effectively enhances survivors’ moderate-to-vigorous physical activity and overall quality of life, but also enables the early detection of high-risk post-treatment comorbidities, such as atrial fibrillation, while facilitating the long-term, digitally precise management of psychosocial symptoms like anxiety and depression.

The integration of consumer wearable devices into oncological care represents a significant advancement in enhancing cancer survivorship, transitioning care models from episodic clinical evaluations to continuous physiological monitoring. Recent meta-analyses underscore the significant efficacy of these devices, such as pedometers and fitness trackers, in substantially elevating moderate-to-vigorous physical activity and subsequently improving overarching metrics of physical function, mood states, and health-related quality of life among survivors. ⁴⁹ Beyond rudimentary kinetic tracking, wearables are increasingly instrumental in capturing nuanced cardiovascular data, offering a critical mechanism for the early detection and proactive management of comorbidities like atrial fibrillation, which uniquely threaten post-treatment oncology populations. ⁵⁰ Furthermore, the scope of these technologies extends deeply into psychosocial oncology; by coalescing longitudinal sensor data with self-reported outcomes, clinicians can dynamically monitor symptoms of anxiety and depression to provide a comprehensive view of mental health that expedites targeted interventions. 5551 Notably, the trajectory of this field is currently being propelled by the convergence of wearable telemetry and AI, highlighting predictive analytics that catalyze the synthesis of biometric data to optimize personalized symptom management and tailored care plans. ⁵² Conversely, while this technological integration shows immense promise, it introduces complex challenges regarding long-term patient adherence, necessitating rigorous regulatory standardization and robust data security protocols to ethically protect patient privacy across widespread adoption. ⁵²

4.2.5. Wearable devices in healthcare: Advancements in human activity recognition

As detailed in the co-citation network analysis (Table 3 and Figure 6), Cluster #5 “human activity recognition” emerges as a pivotal specialized research domain, structurally defined by 57 core nodes with a highly cohesive Silhouette score of 0.885 and an average publication year of 2019. As detailed by the clinical applications in Human Activity Recognition outlined in Table 4, wearable medical devices leverage the deep integration of inertial measurement units (IMUs) and advanced deep learning algorithms to accurately translate continuous raw sensor data into highly actionable clinical insights regarding physical activity and postural deviations. This application not only catalyzes the transition from basic activity quantification to the remote evaluation of complex chronic diseases, but also serves an indispensable role in empowering post-acute precision rehabilitation and behavioral health interventions.

Wearable medical devices have fundamentally transformed the paradigm of human activity tracking and contemporary healthcare delivery by seamlessly highlighting the continuous, real-time monitoring of critical physiological parameters, thereby reducing the necessity for frequent clinical visits. ⁵³ Moving beyond basic physical activity quantification, these intelligent sensors are now extensively deployed in complex chronic disease management frameworks, where they facilitate the early detection of pathophysiological anomalies and allow practitioners to remotely evaluate longitudinal disease progression and treatment efficacy. ³² Furthermore, the strategic integration of such biomechanical and physiological monitoring systems significantly augments post-acute rehabilitation protocols, systematically empowering individuals to regain functional independence through the use of smart prosthetics and sophisticated human augmentation technologies. ⁴² Notably, the convergence of ubiquitous computing and advanced artificial intelligence within these wearable platforms is expanding their clinical utility into behavioral health domains, meticulously capturing subtle physiological shifts that serve as highly reliable indicators of a user’s underlying affective state. ⁵⁴ Despite these profound advancements in facilitating personalized medical interventions and motivating positive behavioral modifications, the widespread, standardized clinical adoption of wearable medical technologies remains heavily constrained by multifaceted challenges encompassing data accuracy, stringent privacy protocols, clinical validation, and comprehensive ethical considerations. ⁵⁵

4.2.6. Wearable devices in healthcare: Current trends in breastfeeding support

As detailed in the co-citation network analysis (Table 3 and Figure 6), Cluster #1 (“breast milk”) emerges as a specialized research domain, comprising a size of 70 core nodes with a highly cohesive Silhouette score of 0.914 and an average publication year of 2016. As summarized in Table 4 regarding breastfeeding support, wearable technology is fundamentally revolutionizing postpartum maternal care and neonatal nutrition management. By integrating accelerometers and flexible bending sensors into smart nursing garments, these systems enable real-time postural monitoring to optimize maternal alignment and alleviate musculoskeletal pain. Furthermore, the incorporation of microfluidic channels and laser-induced graphene sensors into lactation pads allows for the rapid, quantitative detection of pharmaceutical residues (e.g., acetaminophen) in breast milk—successfully achieving a leap from traditional physical kinematic tracking to complex, precision biochemical diagnostics.

Recently, this technological evolution has expanded to address the specific, yet historically underrepresented, domain of postpartum maternal care and breastfeeding support. A prominent advancement in this specialized field involves ergonomic interventions designed to mitigate musculoskeletal pain, a prevalent deterrent to exclusive breastfeeding. Notably, the development of smart nursing garments integrated with accelerometers and flexible bending sensors allows for real-time postural monitoring and feedback, demonstrating significant improvements in maternal postural alignment and relaxation. ⁵⁶ Furthermore, wearable innovations have transcended physical monitoring to encompass complex biochemical analysis. A novel application includes the integration of microfluidic channels and laser-induced graphene sensors into lactation pads, highlighting the rapid, quantitative detection of acetaminophen levels directly within breast milk. ⁵⁷ This diagnostic capability is critical for preventing pharmacological toxicity in both nursing mothers and their breastfed infants. Despite these promising clinical trajectories, optimizing wearable efficacy requires overcoming persistent systemic barriers. Challenges pertaining to data accuracy, patient privacy, device usability, and seamless mobile application integration remain substantial impediments to widespread clinical adoption.^58,59 Addressing these multifaceted limitations is imperative to fully harness the potential of wearable therapeutics in advancing comprehensive, patient-centered postpartum care.

4.2.7. Wearable devices in healthcare: Medical applications of PM_2.5 environmental sensors

Structurally defined by 93 core nodes, a robust Silhouette score of 0.889, and a mean publication year of 2016 (Table 3 and Figure 6), Cluster #0 (“pm_2.5”) encapsulates a rapidly burgeoning research trajectory. This domain specifically focuses on the strategic deployment of wearable sensing architectures to facilitate high-resolution, PM_2.5-associated environmental health monitoring. As summarized in Table 4 regarding PM_2.5 environmental exposure monitoring, wearable sensors are fundamentally transforming chronic respiratory disease management. Synergizing real-time ambient particulate tracking with physiological vital sign monitors (e.g., heart rate and SpO₂), this approach generates fine-grained geospatial data that empowers users to proactively minimize pollution exposure, thereby constructing a robust, multimodal remote patient care framework conducive to prompt clinical interventions.

Wearable devices equipped with PM_2.5 environmental sensors have emerged as pivotal tools in modern healthcare, fundamentally reshaping the continuous monitoring of respiratory conditions and environmental exposures. Unlike traditional fixed monitoring stations, these advanced wearables facilitate the real-time assessment of individual exposure to particulate matter across dynamic indoor and outdoor environments, a capability that is critical for managing chronic lung diseases. ⁶⁰ By generating fine-grained geospatial data, novel portable monitors empower users to visualize immediate air quality and proactively minimize their exposure to harmful ambient pollutants. ⁶¹ Furthermore, integrating PM_2.5 sensors with sophisticated physiological tracking tools-such as heart rate and oxygen saturation monitors-enables a comprehensive, multimodal approach to remote patient care that actively supports timely clinical interventions. ⁶² Despite these substantial clinical advancements, the widespread integration of wearable air quality monitors into decentralized healthcare frameworks necessitates overcoming significant technical hurdles. Specifically, ensuring the rigorous calibration of these devices against research-grade instruments remains an imperative prerequisite for maintaining diagnostic data reliability. ⁶³ Additionally, addressing inherent power constraints through innovative engineering solutions, such as battery-aware algorithms and solar energy harvesting, is essential to extend operational runtime and ensure sustained, uninterrupted health monitoring. ⁶⁴

4.2.8. Wearable devices in healthcare: Smart dressings for chronic wound management

Structurally defined by 68 core nodes, a robust Silhouette score of 0.952, and a mean publication year of 2017 (Table 3 and Figure 6), Cluster #2 (“chronic wound”) encapsulates a rapidly burgeoning research trajectory dedicated to smart wound dressings. Driven by the severe morbidity of non-healing wounds, this domain has evolved into a cornerstone of precision medicine. Mechanistically, the research within this cluster is delineated by four foundational pillars: continuous longitudinal monitoring, the strategic mitigation of surgical frequency, the implementation of autonomous drug delivery, and the critical evaluation of prevailing technical constraints. As summarized in Table 4 regarding chronic wound management, smart dressings represent a fundamental paradigm shift from passive observation to active, closed-loop bioelectronic interventions. Integrating microelectronic and electrochemical sensors allows for the continuous, real-time tracking of microenvironmental metrics like temperature and moisture, which dynamically drives autonomous, on-demand drug delivery or electrical stimulation to catalyze tissue regeneration.

While traditional passive wound dressings provide a basic protective barrier, their inability to dynamically address the complex intricacies of the healing process remains a formidable clinical obstacle. ⁶⁵ To overcome these fundamental limitations, advanced smart wound dressings have emerged as a transformative solution by integrating microelectronic and electrochemical sensors that facilitate the continuous, real-time monitoring of critical physiological parameters within the wound microenvironment. ⁶⁶ Furthermore, modern engineering paradigms have evolved beyond mere diagnostic observation, coupling these sensing capabilities with active therapeutic interventions to establish sophisticated, closed-loop bioelectronic systems. By leveraging continuous data feedback from embedded sensors-such as localized temperature and moisture detectors-these wireless, wearable architectures can autonomously trigger on-demand pharmacological release or deliver targeted electrical stimulation, thereby accelerating tissue regeneration and synergistically managing excessive wound exudate. ⁶⁷ Conversely, despite the profound potential of these integrated platforms to enable highly personalized wound care, their successful translation from bench to bedside is currently impeded by significant technical hurdles. Notably, ensuring the long-term biochemical stability of sensitive biosensors in highly corrosive, enzyme-rich exudates, alongside the pressing need to establish standardized data sets and robust clinical validation frameworks, represent critical imperatives for future research. ⁶⁸ Ultimately, resolving these integration and stability challenges will be essential for cementing intelligent biomaterial systems as the gold standard in chronic wound management.

4.2.9. Wearable devices in healthcare: Advancements in personalized health monitoring

As detailed in the co-citation network analysis (Table 3 and Figure 6), Cluster #9 (“personalized health monitoring”) emerges as a specialized research domain, comprising 44 core nodes with a highly cohesive Silhouette score of 0.88 and an average publication year of 2016. This cluster represents a sophisticated research trajectory focused on the deployment of wearable technologies to continuously measure critical biomarkers for customized medical diagnostics and physiological evaluation. Over the past few decades, the tremendous development of electronics, the tremendous development of electronics, biocompatible materials, and nanomaterials has resulted in the development of implantable devices and small sensors, making real-time personalized precision medicine possible. ⁶⁹ As summarized in Table 4 regarding personalized health monitoring, wearable biosensors are reflecting a significant transition toward proactive precision medicine through the continuous, real-time tracking of multiparameter vital signs. Synergizing with artificial intelligence and machine learning, these systems transcend basic observation to achieve early anomaly detection and preventive intervention, ultimately deciphering complex biometric data to deliver predictive analytics and highly tailored therapeutic strategies.

The application of wearable devices in the healthcare sector is profoundly accelerating the evolution of personalized medicine by facilitating continuous, real-time monitoring of multiparameter physiological vital signs, such as heart rate, blood pressure, and blood glucose, thereby providing patients with highly precise health management solutions. ⁴⁰ In clinical practice, these devices not only enable the early detection and preventive intervention of physiological anomalies but also play a critical role in the longitudinal management of chronic conditions, including cardiovascular diseases and diabetes. ⁴¹ By integrating with the IoMT to support robust telemedicine frameworks, wearables significantly enhance patient engagement and self-management, effectively overcoming the inherent limitations of traditional, intermittent clinical assessments.^70,71 Furthermore, the deep integration of artificial intelligence (AI) and machine learning algorithms enables these systems to decipher complex health data patterns, offering proactive predictive analytics and customized therapeutic recommendations.^72,73 However, the widespread clinical translation of these technologies is currently impeded by substantial sociotechnical challenges, primarily concerning data privacy and security vulnerabilities, the accuracy and reliability of sensor data, and the interoperability barriers hindering seamless integration with existing electronic health record systems. ⁷⁴ To surmount these bottlenecks, future developments must leverage the synergy of next-generation biosensors and 5G telecommunications to optimize data transmission speeds, while concurrently exploring the convergence of AI and blockchain technologies to fortify privacy protection protocols.^70,75 Ultimately, the establishment of unified global policies and industry standards will be imperative to resolve system interoperability issues and fully actualize a highly efficient, ubiquitous landscape of personalized preventive healthcare. ⁴¹

4.3. Emerging trends

While the co-citation clusters (Figure 6) delineate ‘what’ specific clinical applications (e.g., COVID-19, breast milk) the field has structurally focused on, keyword burst analysis reveals ‘when and how’ specific methodological concepts rapidly gained traction. The bursts identified below provide the dynamic technological context that catalyzed the growth of the aforementioned clinical clusters.

4.3.1. Wearable devices in healthcare: Economic impact on healthcare expenditures

CiteSpace keyword burst (Figure 7) analysis reveals a critical intersection between digital health and healthcare economics. “Wearables” emerges as the leading research frontier (burst strength: 12.64), closely aligned with “telemedicine” and “health monitoring.” Concurrently, a sustained burst in “medical costs” (strength: 10.66) since 2024 highlights a pivotal clinical focus: leveraging wearables for continuous health monitoring. By facilitating early pathology detection and proactive interventions, these technologies offer a strategic solution to the escalating economic burdens of modern clinical practice.

Figure 7.

Keywords with the strongest citation bursts related to the application of wearable devices in healthcare.

Note. This updated burstness map reflects literature data spanning up to April 2026.

Wearable medical technologies are increasingly recognized for their capacity to alleviate the escalating economic burden on global healthcare systems through proactive monitoring and decentralized disease management. For instance, continuous glucose monitoring systems have revolutionized diabetes care by delivering real-time physiological data that minimizes costly complications, thereby optimizing long-term therapeutic expenditures. ⁷⁶ Similarly, cardiovascular and neurodegenerative conditions have emerged as prominent focal points within the recent literature. By examining the citation context of the burst keywords associated with these domains, we observed that researchers are increasingly focusing on the potential of wearables for early detection and tailored interventions. For instance, within this core citing literature, individual primary studies have reported highly promising outcomes; notably, Noci et al. 2025 observed that devices monitoring heart failure could decrease 30-day readmission rates by up to 89%, suggesting profound systemic cost savings. ⁷⁷ It is important to emphasize that while our bibliometric analysis highlights a surging academic interest in these clinical and economic benefits, the validation of such effectiveness claims rests entirely on these individual clinical studies rather than our bibliometric mapping. Furthermore, inertial sensors deployed for advanced Parkinson’s disease allow for precise medication optimization, yielding projected national healthcare savings of up to €137.8 million in countries such as Germany. ⁷⁸ The financial benefits also extend to diagnostic pathways, where wearable tools for obstructive sleep apnea enhance diagnostic capabilities while significantly lowering the cost per test. ⁷⁹ Despite these compelling economic advantages, the widespread integration of wearables is contingent upon overcoming substantial sociotechnical barriers. Notably, while the prospect of utilizing health data to reduce insurance premiums appeals to consumers, pervasive concerns regarding data privacy and the initial economic burden of the devices themselves continue to impede universal adoption. ⁸⁰ Consequently, realizing the optimal return on investment necessitates comprehensively addressing these privacy frameworks and regulatory challenges to foster broader clinical and consumer acceptance.

4.3.2. Wearable devices in healthcare: Integration with telemedicine systems

CiteSpace keyword burst analysis (Figure 7) reveals a profound paradigm shift within the digital health literature. Since 2023, the term ‘wearables’ (burst strength: 12.64) has exhibited a sustained, high-intensity citation burst, followed closely in 2024 by similarly robust surges in ‘telemedicine’ (36.85) and the ‘IoMT’ (33.87). This convergence highlights a critical research frontier: the synergy of remote medical frameworks and interconnected devices. By facilitating the continuous capture of “health monitoring” (23.56), these technologies bridge traditional clinical settings and decentralized care, establishing the foundation for advanced healthcare delivery.

The integration of wearable devices with telemedicine infrastructures is fundamentally transforming modern healthcare delivery by facilitating continuous, real-time patient monitoring and remote disease management. Contemporary wearable technologies, such as clinical-grade smartwatches and continuous glucose monitors, enable the seamless extraction of dynamic physiological data-including electrocardiogram readings and glycemic fluctuations-which are critical for the proactive management of chronic conditions like cardiovascular and respiratory diseases. ⁸¹ Integrating wearable devices into active clinical tools is far from a simple ‘plug-and-play’ endeavor; rather, it necessitates dedicated system architectures leveraging standardized protocols like Health Level 7 Fast Healthcare Interoperability Resources (HL7 FHIR) to establish secure, bidirectional data pipelines. ⁸² To mitigate clinical alert fatigue, the optimal integration framework should incorporate edge computing and artificial intelligence intermediaries as a triage mechanism, thereby precisely embedding high-fidelity, actionable physiological alerts into routine EHR workflows without compromising data security. ⁸³ Furthermore, the architectural convergence of these sensor ecosystems with advanced computational frameworks, notably cloud computing and AI, has significantly augmented diagnostic and prognostic capabilities. ⁸⁴ AI-driven algorithms applied to wearable-derived metrics not only enhance predictive analytics and clinical decision-making but also expedite early interventions, thereby optimizing patient outcomes and minimizing hospital readmission rates. ⁸⁵ Conversely, while this technological synergy promotes operational efficiency and patient empowerment, its widespread clinical assimilation is impeded by substantial systemic barriers. Chief among these are the imperative need for robust data encryption protocols to safeguard patient privacy and the lack of standardized interoperability across disparate device architectures. ⁷⁵

4.3.3. Wearable devices in healthcare: Enhancing health-related quality of life

CiteSpace keyword burst analysis (Figure 7) underscores a distinct evolution toward patient-centric digital health. Since 2022, “quality of life” has emerged as a prominent research frontier (burst strength:20.98), paralleling the rapid growth of “wearables” (12.64) and “health monitoring” (25.36). This convergence highlights a critical clinical evolution: leveraging advanced monitoring technologies not merely for disease tracking, but to actively enhance the patient’s daily lived experience.

The integration of wearable devices into healthcare systems has fundamentally transformed patient care, significantly enhancing health-related quality of life through continuous physiological monitoring and proactive health management. By facilitating the real-time collection of vital metrics-such as physical activity, sleep patterns, and vital signs-wearables empower healthcare providers to implement timely interventions outside traditional clinical settings. ⁸⁶ Furthermore, this continuous stream of health data, when coupled with AI and big data analytics, enables the development of highly customized treatment plans that improve overall patient outcomes. ⁸⁷ Beyond clinical oversight, these devices foster profound personal health awareness, intrinsically motivating individuals to adopt healthier lifestyle behaviors and strictly adhere to medical regimens. Conversely, despite these transformative benefits, the widespread clinical adoption of wearable technologies remains impeded by critical challenges. Notably, vulnerabilities regarding data privacy and security necessitate rigorous safeguarding protocols to establish and maintain patient trust. ⁸⁸ Moreover, ensuring the absolute accuracy and reliability of sensor data is paramount, as discrepancies could precipitate erroneous medical interventions. ⁸⁹ Ultimately, while wearable healthcare technologies possess immense potential to shift the paradigm from reactive to preventive medicine, addressing these technological and ethical barriers is essential for their seamless integration into modern healthcare infrastructures.

4.3.4. Wearable devices in healthcare: Applications in physical rehabilitation

CiteSpace burst analysis (Figure 7) reveals a pronounced shift in the digital health literature toward technology-driven physical recovery. Since 2024, “training” has emerged as a key research frontier (burst strength: 34.29), closely paralleling the sustained prominence of “wearables” (12.64) and “health monitoring” (23.56). This convergence highlights a pivotal clinical evolution: the transition from generalized physical therapy to digitally augmented precision rehabilitation. By integrating these interconnected devices, modern training protocols can effectively address clinical vulnerabilities and redefine patient-centric recovery.

The integration of wearable devices into physical rehabilitation has precipitated a critical transition from episodic clinical encounters to continuous, patient-centric care models. By leveraging advanced hardware such as accelerometers, gyroscopes, and electromyography, these technologies facilitate the real-time, objective quantification of movement biomechanics and postural deviations. ⁹⁰ This real-time continuous monitoring fundamentally enhances the efficacy of remote rehabilitation architectures, highlighting practitioners to evaluate patient progress and minimize the logistical constraints of frequent hospital visits. ⁹¹ Furthermore, the continuous feedback loop provided by wearables significantly augments patient engagement and exercise adherence, a dynamic particularly critical for optimizing outcomes in cardiovascular rehabilitation and orthopedic recovery protocols.^92,93 By dynamically synthesizing patient-specific kinetic data, clinicians can construct highly personalized therapeutic regimens tailored to individualized recovery trajectories, moving beyond generalized care. ⁹⁰ Notably, the convergence of wearable hardware with big data analytics and AI promises to deepen clinical insights, transitioning rehabilitation from a reactive to a predictive discipline capable of identifying early markers of musculoskeletal decline. ⁹⁴ Conversely, the widespread clinical assimilation of these systems remains encumbered by systemic challenges, most notably data privacy vulnerabilities, variable sensor fidelity, user acceptability, and the imperative for rigorous clinical validation.^90,91,94 Resolving these bottlenecks is essential to fully manifest the transformative potential of wearable therapeutics in modern musculoskeletal and cardiovascular care.

4.3.5. Wearable devices in healthcare: Personalized medicine and AI integration

CiteSpace keyword burst analysis (Figure 7) reveals a critical intersection between AI (burst strength: 16.98) and personalized healthcare services. Within the broader paradigm of “digital health” (burst strength: 21.35), “wearables” emerges as a leading research frontier (12.64), closely aligned with “health monitoring” (23.56). This technological convergence highlights a core clinical focus: leveraging advanced algorithms to transition generalized care into highly customized, predictive medicine.^95,96

As pivotal components of the broader digital health ecosystem, a diverse array of wearable devices—spanning smartwatches, fitness trackers, smart clothing, and portable biosensors—enables the continuous, real-time acquisition of voluminous physiological and behavioral datasets, including heart rate metrics, blood pressure, ECG, oxygen saturation levels, and sleep patterns.^87,97 The integration of AI and machine learning architectures—such as deep learning, convolutional neural networks, and reinforcement learning—into these devices empowers digital health systems to rapidly identify potential health risks through multi-source data fusion, anomaly detection, and predictive analytics.^87,95,96 For instance, embedded AI engines can scrutinize streaming biophysical data in real time to detect outliers, thereby generating customized health alerts and providing early warnings for cardiovascular disorders such as arrhythmia and hypertension, which effectively mitigates the risk of irreversible clinical damage.^98,99 This profound technological synthesis represents a milestone transformation in healthcare delivery, unlocking unprecedented capabilities for digital health platforms to offer personalized monitoring, automated diagnosis, and targeted therapeutic interventions.^95,96 By dynamically analyzing individualized health profiles, these intelligent systems can not only optimize personalized rehabilitation protocols and tailored training regimens but also effectively empower patients to engage in proactive self-health management outside traditional clinical settings.^97,100,101 Furthermore, transcending conventional physical sign tracking, AI-assisted wearables are expanding the emerging trends of digital health into the domain of psychological and mental health care, capturing subtle behavioral modifications and autonomic physiological responses to reliably assess affective states and cognitive stress levels.^54,97 Concurrently, the convergence of these sensor ecosystems with the IoMT framework, coupled with decentralized digital health architectures like edge computing and federated learning, vastly enhances remote patient care capabilities, optimizing data processing efficiency while ensuring the fidelity of real-time clinical interventions.^89,95,101 Nevertheless, despite demonstrating immense clinical potential in disease forecasting and functional recovery, the widespread assimilation of fully personalized, AI-driven wearable systems into mainstream digital health infrastructures remains encumbered by significant socio-technical and ethical bottlenecks.^102,103 The primary challenges impeding complete clinical translation include stringent data privacy regulations, the imperative for robust data encryption protocols, the lack of algorithmic transparency (the “black-box” nature of complex models), and interoperability hurdles in seamlessly integrating heterogeneous wearable sensor streams into existing EHR systems.^75,104 To fully manifest this revolutionary potential and cultivate enduring user trust, future research trajectories must prioritize explainable AI (XAI), foster multi-centric interdisciplinary collaborations among engineers and clinicians, and establish standardized data governance frameworks alongside global digital health regulatory guidelines.^46,95,101

4.3.6. Wearable devices in healthcare: Personalized medicine and integration with EHR

CiteSpace keyword burst (Figure 7) analysis reveals a critical intersection between digital health ecosystems and personalized medicine. “EHR” emerges as the preeminent research frontier (exhibiting the highest burst strength of 40.28), closely aligned with “personalized medicine”. Concurrently, a sustained burst in “telemedicine” since 2024 highlights a pivotal clinical focus: leveraging wearable-generated data for continuous, individualized health monitoring. By facilitating seamless data transmission into clinical workflows, these technologies offer a strategic solution to transitioning from intermittent clinical assessments to proactive, personalized patient care.

Wearable devices, ranging from consumer-grade fitness trackers to highly calibrated clinical-grade biosensors, are profoundly reshaping contemporary healthcare paradigms by transitioning reactive treatment into proactive, personalized disease management.^41,70 By tracking the continuous, real-time capture of physiologic, behavioral, and environmental patient-generated health data (PGHD) between traditional clinic visits, these digital health tools effectively overcome the inherent limitations of intermittent clinical assessments, highlighting early anomaly detection and comprehensive physiological profiling.^41,105,106 When seamlessly integrated into EHR and electronic medical records (EMRs), wearables provide clinical teams with accurate, up-to-date health metrics that support prompt, personalized care and well-informed, deliberate clinical decision-making.^105,107 This integrated digital architecture has demonstrated profound clinical utility across a wide spectrum of medical contexts, facilitating robust remote patient monitoring for chronic conditions—such as diabetes, cardiovascular diseases, and respiratory disorders—while concurrently optimizing post-surgery recovery tracking, elderly care, and mental health surveillance.^70,108,109 Furthermore, the widespread deployment of these smart biosensors has fundamentally reshaped health management systems, especially in the post-COVID-19 era, by minimizing unnecessary hospital visits and establishing high-quality data exchange portals at a lower relative socioeconomic cost.^53,108 Despite these clear clinical benefits, the full-scale implementation of integrated wearable systems within current administrative workflows remains substantially encumbered by persistent challenges in data interoperability, privacy protection, and data management complexity.^53,108 Achieving seamless interoperability requires the universal standardization of method communication and data formats, which is increasingly facilitated by integrating standardized data models such as the HL7 FHIR protocols. ¹¹⁰ Concurrently, because wearables harvest massive streams of highly sensitive user data, prioritizing robust cybersecurity measures, compliance with regulatory standards like Health Insurance Portability and Accountability Act, and establishing transparent consent practices remain critical prerequisites to safeguard patient confidentiality and maintain system trust.^111,112 Moreover, the continuous stream of high-volume biometric data threatens to cause severe data overload, creating severe workflow challenges that can overwhelm clinical staff and existing EHR infrastructures.^105,112 To resolve this computational strain and mitigate clinical alert fatigue, contemporary paradigms are rapidly integrating advanced AI, machine learning (ML) algorithms, and predictive analytics to automate features extraction and distill actionable, real-time insights for preventive and personalized medicine.^41,109,113 Ultimately, as these advanced software architectures mature alongside high-speed 5G connectivity, the IoMT, and robust cloud computing environments, they will seamlessly bridge remote patient tracking with centralized medical institutions, fostering a highly responsive, proactive, and patient-centric global care ecosystem.^41,113,114

5. Conclusion

This study provides a comprehensive dual-database (WoSCC and Scopus) bibliometric assessment of wearable device applications in healthcare from 2016 to 2026. The global knowledge base demonstrates a robust and continuous upward trajectory, mirroring the escalating global demand for digital health innovations. Geographically, China and the United States dominate the research landscape, while at the institutional level, the Chinese Academy of Sciences leads in global publication output. Furthermore, the strategic integration of Scopus effectively mitigated single-source bias, revealing a substantial volume of engineering-focused research from institutions in India and South Korea. Professor Wang Wei possesses the highest publication output, whereas Professor Wan Peng Bo achieved the highest citation impact, reflecting their distinct yet profound contributions to the development of wearable healthcare technologies.

Based on the updated bibliometric mapping, the global knowledge structure of wearable healthcare devices is distinctly anchored around core clinical and technical domains, most notably chronic disease management, COVID-19 pandemic response, IoMT, and personalized health monitoring. Furthermore, keyword burst detection reveals a dynamic trajectory shifting from isolated hardware development toward integrated, patient-centric digital health ecosystems. Current research emerging trends are heavily driven by the integration of telemedicine, AI, and EHR, with an escalating clinical focus on reducing medical costs, optimizing physical rehabilitation (training), and enhancing the overall quality of life. These objective bibliometric outputs provide a clear roadmap of the field’s evolution from passive monitoring to proactive, AI-driven personalized medicine.

In conclusion, while this bibliometric analysis maps the structural and dynamic evolution of wearable technologies in healthcare, its broader implications extend far beyond academic publication trends. The trajectory of this field signifies a profound transition from reactive clinical interventions to continuous, proactive, and personalized health management. However, the true realization of an AI-driven, decentralized digital health ecosystem requires transcending narrow disciplinary boundaries. Future advancements depend heavily on robust multidisciplinary synergy: engineers must resolve critical hardware constraints and data interoperability issues; clinicians must establish standardized protocols for integrating patient-generated health data into existing electronic workflows without causing alert fatigue; and policymakers must urgently architect comprehensive data governance frameworks to safeguard patient privacy. Ultimately, navigating these complex sociotechnical challenges is imperative to translate wearable innovations into sustainable, equitable, and globally accessible healthcare solutions.

6. Limitation and future perspectives

While this study provides a comprehensive landscape of wearable technologies in healthcare, its findings must be interpreted within the context of five major bibliometric challenges, adapting the framework recently proposed by Ref. 115. First, regarding Quality, our metrics inherently reflect academic attention (e.g., citation frequency) rather than the true methodological rigor or clinical efficacy of the wearable devices. Second, concerning Impact, high citation parameters do not automatically equate to practical utility in clinical settings, as they can be influenced by field-specific citation behaviors. Third, in mapping collaborative networks, our analysis is constrained by issues of Co-authorship, as bibliometric data cannot differentiate true intellectual contributions from honorary authorship. Fourth, despite our comprehensive approach, we acknowledge the limitations of commercial Databases, particularly their inherent English-language bias which may underrepresent impactful localized studies. Finally, automated bibliometrics is vulnerable to academic Fraud; the software takes metadata at face value and may inadvertently encompass manipulated citations or retracted literature that evades initial screening. Future research must combine these bibliometric insights with systematic clinical reviews to overcome these inherent constraints.

Although methodologically rigorous, this study is constrained by publication, language, and temporal citation biases inherent to English-centric academic databases and retrospective bibliometric tools, highlighting the need for future research to leverage multi-lingual platforms, patent datasets, and AI-driven topic modeling. Furthermore, the bibliometric mapping reveals a geographical “silo effect,” with research predominantly driven by China and the United States. To address this, future initiatives must actively encourage multi-center, cross-border collaborations to ensure that wearable innovations are equitable and generalizable across diverse socioeconomic and ethnic populations. Additionally, to bridge the translational gap between engineering prototypes and real-world clinical adoption, future efforts must prioritize large-scale, longitudinal randomized controlled trials (RCTs). These trials are essential to rigorously evaluate not only device accuracy and cost-effectiveness but also long-term user compliance and the psychological impacts of continuous physiological monitoring.

While this study provides a robust snapshot and trend prediction based on bibliometric algorithms, the rapid evolution of the field suggests that future research could benefit from integrating ‘living’ online monitoring systems—similar to paradigms like Stress in Action—to achieve real-time tracking of technological breakthroughs. By synergizing such dynamic empirical databases with macroscopic bibliometric roadmaps, the academic community can construct a more agile, comprehensive, and highly responsive knowledge infrastructure for digital health.

ORCID iDs

Yixuan Ding https://orcid.org/0009-0006-2446-1735

Siyi Wen https://orcid.org/0009-0008-0238-1456

Wenbing Yu https://orcid.org/0000-0001-6183-6942

Ethical considerations

As this study exclusively utilized publicly accessible, secondary bibliometric data and involved no direct contact with human participants or animal subjects, ethical approval and informed consent were waived.

Consent for publication

This manuscript does not contain any individual patient’s data, personal identifiers, images, or videos.

Data Availability Statement

To ensure full methodological transparency and reproducibility, a Complete Reproducibility Package for this study has been deposited in a permanent public repository Zenodo and is openly accessible at: https://zenodo.org/records/20841921. This repository comprehensively includes: (1) the raw search export files from the WoSCC and Scopus databases; (2) the specific Python execution scripts and precise deduplication decision logs generated during the BibexPy framework integration; (3) the customized term-cleaning thesaurus files used for node standardization; and (4) the final, cleaned dataset comprising the 12,812 eligible records utilized for the bibliometric visualizations. The foundational architecture of the BibexPy framework utilized in this study remains publicly accessible via its original publication (https://doi.org/10.1016/j.softx.2025.102098). ⁹