Dementia is a progressive neurocognitive syndrome currently cited as the seventh leading cause of death worldwide [1]. This degenerative disease is also one of the leading causes of disability or dependency for aging populations today [2]. Dementia places a substantial burden on health care systems, national economies, and families, who often bear the responsibility of caregiving. Beyond the financial and logistical strain, individuals with dementia endure profound cognitive and neurological challenges, significantly affecting their quality of life [3]. In 2019, dementia affected over 55 million people globally and cost the world economy US $1.3 trillion. Nearly half of the financial burden was borne by informal caregivers [4]. With no currently available cure [5], efforts are focused on disease management, improving quality of life, and providing well-being support [6].

In recent years, the health care landscape has undergone a profound transformation with the integration of digital health technologies (DHTs), particularly the technologies applicable in the care of people living with dementia and their health care providers (formal and informal). Digital solutions offer novel avenues for enhancing medical care by supporting early diagnoses and implementing better preventive strategies [7]. Furthermore, DHTs can potentially minimize the burden of care felt by family members, who often devote over 5 hours daily to providing care and supervision [8]. Digital health interventions can offer a multifaceted approach to dementia treatment and prevention and revolutionize the modern approach to health care delivery and support in daily living. This comprehensive group of interventions covers multiple modalities of technologies, including artificial intelligence (AI), computer-based cognitive interventions and digital platforms, GPS, sensors for remote care, smart devices, mobile health, instant messaging applications, wearable devices, virtual reality, and telehealth [9].

Recognizing the value of multidisciplinary research initiatives, the World Health Organization (WHO) Regional Office for Europe has taken a leading role in advancing scientific research in this area. Key initiatives have made significant contributions to exploring the intersection between digital health and healthy aging, including the United Nations Decade of Healthy Ageing (2021-2030) [10], the WHO Global Digital Health Strategy 2020 to 2025, the Regional Digital Health action plan for the WHO European Region 2023 to 2030, and the WHO European Framework for Action on Mental Health 2021 to 2025. Acknowledging the importance of its role, the WHO is committed to supporting the achievement of the sustainable development goals, specifically goal 3 (good health and well-being) and goal 9 (industry, innovation, and infrastructure) [11]. The research questions in this study aligned with 4 particular targets: reduce mortality from noncommunicable diseases and promote mental health (3.4), achieve universal health coverage (3.8), enhance research and upgrade industrial technologies (9.5), and universal access to information and communications technology (9.8) [11].

A protocol was published on February 19, 2024, which provided an overview of systematic and scoping reviews detailing how the analysis of different modalities of DHTs could improve the prevention, diagnosis, treatment, care, and support of (older) people living with dementia, their families, wider support network, and care partners [12]. The research insights were drawn from a series of reviews on the topic and the numerous interventions for the holistic management of dementia, underscoring the necessity for compiling and summarizing the existing evidence in a solid and systematic document. As of April 2024, there has been a paucity of comprehensive, high-quality bibliometric analysis that examines the complete scope of research intersecting between the interface of DHTs and dementia care. Notably, compiling this bibliometric analysis provides valuable insights to identify areas for further investigation in granular detail, acting as a conduit for researchers working in areas of technology, gerontology, social science, gerontechnology, social research, and policy to initiate a route map based on the gaps in the literature.

This publication synthesizes the preliminary findings from the broad systematic search on February 21, 2024. This work will serve as a springboard for future investigations with a more detailed synthesis of specific modalities in digital health and how they relate to the 8 domains of the WHO’s Age-friendly Cities and Communities Framework. We anticipate that this series of documents will yield evidence beneficial to clinicians and policy makers alike working with people living with dementia.

Our study combines 2 methodologies within the field of evidence-based medicine as follows: (1) overview of systematic reviews and (2) bibliometric analysis. We combined both methodologies to leverage their complementary strengths in analyzing the complexity and evolution of DHTs in dementia care [13]. We initially used the principles of systematic reviews to obtain a broader number of studies, enabling the execution of the bibliometric analysis. The integration of both approaches provided a comprehensive and in-depth understanding of the current state of DHTs in dementia care—insights that neither method could achieve alone. This combined approach ensured rigorous and replicable evaluation of existing evidence from systematic reviews and also positioned our research within the broader context of global research trends.

This bibliometric analysis originated as part of the protocol published before the start of the reviewing process under the PROSPERO tracking locator CRD42024511241 [12]. For this bibliometric analysis, we adhered to the guidelines for bibliometric studies, BIBLIO statement [14] and report the associated checklist as Multimedia Appendix 1. This analysis began with a large-scale review of 4839 articles retrieved from a comprehensive overview of systematic reviews, examining the current literature landscape at the interface of digital health and dementia. According to the international guidelines in evidence-based medicine, overviews of reviews are intended to evaluate the existing evidence from 2 or more reviews of different health interventions for a similar population or condition; they also assess these interventions across varying outcomes, settings, or time points [15].

This publication is a part of the project protocol using 2 primary techniques—systematic review and bibliometric analysis. In the first phase, we systematically searched 5 leading medical databases (MEDLINE, Embase, Scopus, Epistemonikos, and Cochrane Database of Systematic Reviews) for eligible systematic and scoping reviews. In collaboration with an expert committee, our information specialist designed a search strategy based on Medical Subject Headings and nonstandardized technical terms. In addition, the search was complemented by obtaining the first 300 hits from the Scholar (Google LLC) platform (source for gray literature) for eligible reviews from database inception until February 21, 2024, regardless of publication language. Search terms involved, but were not limited to, “dementia,” “Alzheimer disease,” “Lewy body dementia,” “mixed dementias,” “Huntington disease,” “multi-infarct dementia,” “vascular dementia,” and “major neurocognitive disorder.” Full access to the search strategy used in these preliminary findings is available in Multimedia Appendix 2.

Most investigators participated in the study selection phase, which was conducted independently by at least 2 investigators using the Covidence Systematic Review Platform (Veritas Health Innovation). The screening process was conducted in two stages as follows: (1) paper title and abstract and (2) full-text screening. Disagreements were resolved by a third investigator not involved in the initial screening process by evaluating the raised discrepancies compared to the inclusion criteria in the published protocol. Interrater reliability was noted to be average with a mean (SD) of 0.66 (0.31) and 0.57 (0.28) and median (IQR) of 0.58 (0.48-0.79) and 0.66 (0.54-0.77) for the screening stages, respectively, indicating moderate to strong reliability.

Using the PICOS framework, the search criteria for review analysis are described below.

Systematic reviews that included more than one database search, in either narrative or quantitative format, were eligible. Moreover, scoping and rapid reviews were considered, categorized as “a review aiming to identify and map the available evidence, but that utilized a systematic methodological approach for including studies” [16]. We excluded

narrative and integrative reviews (reviews without systematization of search and inclusion approaches) because they are more likely to report data in a highly biased pattern. Regarding the exclusion of narrative and integrative reviews, we would like to highlight several reasons [16,17]. First, narrative reviews are highly prone to selection bias, as they typically do not systematically cover all relevant studies. Similarly, integrative reviews often lack the rigorous, standardized search and inclusion criteria that are fundamental to systematic and scoping reviews [16,17]. Therefore, we excluded these types of publications to minimize potential inconsistencies and biases that could compromise the validity of our bibliometric analysis. [16,17].

Following the initial identification of eligible reviews, a bibliometric analysis was conducted on the records that met our inclusion criterion. First, we downloaded eligible data from Covidence in a cvs format file to create our “primary library.” Raw cvs data were transformed into a relational Research Information System database, with the further exclusion of duplicates not primarily identified by Covidence (“secondary library”). After identifying unique records using their digital object identifier (DOIs), we created a matching “secondary library.” We then downloaded a comprehensive cvs file, which was further used for the bibliometric analyses.

This review used VOSviewer (version 1.6.20; Leiden University) to analyze the type of documents, years, authors, countries, institutions, journal sources, references, and keywords, allowing the creation of social network maps. VOSviewer uses probabilistic-driven normalization to ensure that the strengthen of relationships between nodes is accurately represented, accounting for differences in publication volumes and citations practices across various fields [18]. The projected clusters, each represented by a different color, suggest groups of related items that frequently cooccur in the literature, indicating underlying thematic connections [18]. The importance of these clusters lies in their ability to identify trending research topics, emerging trends, and potential gaps in the literature [18]. The total number of citations each year was obtained from the Scopus citation report. Social network analysis was set as the primary method used to analyze co-occurrence. Data associated with the journal’s impact factor (IF) were obtained from the Clarivate Journal Citation Reports (2022) and matched to the highlighted journals in our results section. From all variables prioritized in our study, corresponding data obtained from VOSviewer visualization analysis and calculations were also exported and tabulated on Microsoft Excel. VOSviewer uses probabilistic-based data normalization to create maps in fields like keywords, countries, and authors. These data were subsequently displayed in descriptive tables.

In bibliometrics, social network assessments are widely used to identify research hot spots and trends within specific scientific fields. The cluster findings were evaluated through VOSviewer to develop social network visualizations, demonstrating the relevance of the node size and line thickness. The nodes represent the number of occurrences or frequencies, while the observed lines between the nodes suggest relationships between nodes. A thicker line denotes a more substantial relationship between the nodal components, while a slimmer line denotes a weaker relationship and connection between nodes. The reported network analyses afford the visualization of trends in reviews on the interface between digital health and dementia.

Data management was performed predominantly in VOSviewer, which uses machine learning algorithms, processed input data related to study identification (title, authors, publication year, journal, and DOI), the number of citations and impact (citation count and IF of the journal in which the review was published), reported geographic regions attributed during indexation, departments and institutions declared during indexation, as well as main descriptors (standardized or nonstandardized indexation nomenclature).

Our bibliometric analysis is based on an overview of systematic and scoping reviews. Bibliometric analyses focus on quantifying publication patterns within a field, including the number of citations, authorships, coauthorship networks, institutional contributions, and keyword trends. Consequently, readers should not expect quality assessments of individual reviews, such as the use of AMSTAR tool.

In addition to our bibliometric assessment, a further classification of included reviews using a 3-step data assessment plan was conducted independently using Covidence. Each review was tagged for the number of primary studies reported in the manuscript (for instance, 0-10, 11-30, 31-60, 61-90, and 91-infinity), the modality of DHT being reported within the review, and if the record was a “systematic” or “scoping” review. We designed a focused analytic step of shortlisted studies to identify overviews intersecting between digital health and dementia. An additional filtering step was applied to identify records that had already used an overview methodology in their assessment. This step was introduced to avoid duplicating research efforts and to prioritize a more detailed exploration of modalities that have historically been less examined in the literature. This resulted in the identification of 14 records with the identifier “overview” or “umbrella,” as shown in Multimedia Appendix 3.

Ethics approval was waived as we used secondary, publicly available data.

A review flowchart diagram is shown in Figure 1. Complete details regarding the justification for the exclusion of the shortlisted studies is available in Multimedia Appendix 4. Reasons for excluding shortlisted studies were primarily associated with conference papers (not published as full reports), protocols, reviews not focusing on DHTs, reviews of bioinformatics concepts, literature reviews, and studies not enrolling patients with dementia or cognitive impairment.

Our results show the categorization of eligible reviews based on the number of primary studies, the modalities of DHTs reported within reviews, and the type of review (scoping or systematic review). Most of the included records were systematic reviews (537/663, 80.9%), while some scoping reviews were also registered (126/663, 19%). In addition, most reviews (346/663, 52.1%) included 11 to 30 primary studies (regardless of the study design). Notably, a lower number of records were allocated to the 31 to 60, 61 to 90, and ≥91 primary studies.

At the time of conducting the literature search, we ascertained that there were no published protocols of an overview of reviews (systematic or nonsystematic reviews), collating the available evidence of multiple modalities of DHTs pertaining to the integrated care of people living with dementia or any other related neurocognitive disorders or cognitive impairment disorders. This literature search is significant because of the volume of studies involved, including a substantial number of published overviews analyzing the relevance of DHTs in dementia care. While we assessed DHT modalities already covered in existing literature reviews, we placed particular emphasis on thoroughly examined less-explored modalities using our overview methodology. In doing so, we ensured a focused and detailed exploration of underrepresented areas. As part of this process, 14 records containing the term “overview” or “umbrella review” were identified in the tracked records (Multimedia Appendix 3). These overviews, published after 2022, have been undertaken on people diagnosed with dementia to understand the effect of DHTs, such as wearable sensors, AI, virtual reality, eHealth, and web-based interventions. We did not include studies without a clear clinical, radiological, or any other deterministic form of diagnosing dementia on the patients under investigation in our analysis.

A total of 755 references were included, and 52 reviews were duplicated (not primarily identified through automatic exclusion using the reference manager software), resulting in 704 references eligible for the bibliometric analysis. Furthermore, 12 references did not have a DOI identifier (Multimedia Appendix 5), and 28 were not indexed in Scopus (Multimedia Appendix 6). Therefore, a total of 663 articles were downloaded into the cvs extension from Scopus, and visualization analysis was conducted.

Table 1 and Figure 2 present our results on the authors with the most publications and the journals with the highest citations in the field of dementia and DHTs based on our included reviews. The Journal of Medical Internet Research had the highest number of reviews (34/663, 5.12%), with the highest citation count from all journals provided in the ranking (1052). Similarly, the Journal of Alzheimer’s Disease and the Cochrane Database of Systematic Reviews showed substantial metrics regarding published reviews and citation count. Note that the average publication year widely varied across ranked journals. The Ageing Research Reviews (IF_2022=13.1) was the journal with highest IF, while the lowest was the Geriatric Nursing (IF=2.7). Six distinct clusters were identified represented in red, green, blue, yellow, purple, orange, and cyan, each corresponding to a unique collaborative or thematic pattern. A complete analysis of the identified clusters is reported as a legend to Table 1.

As seen in Figure 2, the analysis of clusters revealed that cluster 1 (in red) included journals like Dementia, Disability and Rehabilitation: Assistive Technology, and Frontiers in Psychology, which focus on psychological, social, and rehabilitation features of dementia care, particularly the integration of dementia to assistive technologies. Cluster 2 (in green) contained journals like Healthcare (Switzerland) and Frontiers in Aging Neuroscience, which were likely to be associated with health care delivery and neuroscience aspects of dementia care, providing a focused and tailored analysis using medical and biological components attributed to DHTs. In cluster 3 (in blue), we observed the presence of highly influential journals like the Journal of Medical Internet Research and International Psychogeriatrics, whose primary focus relates to internet-based research, psychogeriatrics, and perhaps the broader implications of DHTs in dementia care. Cluster 4 (in yellow) featured journals like the Cochrane Database of Systematic Reviews and Ageing Research Reviews, representing evidence-based reviews and systematic approaches to aging and dementia, emphasizing the fundamental role of rigorous evidence synthesis in the development and validation of DHTs in dementia care. In cluster 5 (in purple), journals like International Journal of Nursing Studies and Journal of Clinical Medicine, were found, which mainly publish research focused on the clinical and nursing aspects of dementia care as well as their integration with DHTs into routine clinical practice. Cluster 6 (in cyan) contained journals like the Journal of Alzheimer’s Disease and Sensors, with a particular focus on the interface between dementia research and sensor technologies, potentially emphasizing the role of wearable and remote monitoring devices in dementia care. Finally, the small cluster 7 (in orange) represented a distinguished group of journals, highlighting specific thematic focus within the field of dementia care and DHTs.

Researchers from 66 countries published reviews on the intersection between dementia and DHTs, and Table 2 details the number of citations, documents, and the average number of publications per year, etc. The top 5 countries with evidence-based research in this arena are the United Kingdom (139/663, 20.9%), United States (105/663, 15.8%), Australia (66/663, 9.9%), Canada (61/663, 9.2%), and the Netherlands (53/663, 7.9%). Column 3 of Table 2 shows the country coauthorship networks, particularly evidencing the chord diagram of country cooperation networks ranked by total link strength (Figure 3). The United Kingdom had the highest number of total link strengths, followed by the United States, the Netherlands, Australia, and Switzerland (335, 155, 145, 142, 128, respectively). Countries with the highest number of citations included the United Kingdom (n=6004), United States (n=4911), Australia (n=3491), the Netherlands (n=2647), and Canada (n=1758). Cluster analysis identified 6 different research clusters within the intersection between DHTs and dementia care. Explanation related to identified clusters is available in Table 2.

As seen in Figure 3, cluster 1 (in red) shows that North America and East Asian countries are powerhouses on the research in the interface between DHTs and dementia care, cluster 2 (in green), represents diverse global contributors with a focus on European and South Asian countries, cluster 3 (in blue) suggests leaders in the European and Middle Eastern research hubs, cluster 4 (in yellow) flags Northern European and Nordic countries as research hubs, cluster 5 (in purple) represents Southern Europe and Middle Eastern countries as emerging players in research with growing interest in advancing research in the intersection of DHTs and dementia care, and cluster 6 (in cyan), represents smaller and more focused contributors with specific expertise (in terms of citations).

Between 2002 and 2024, a total of 3076 authors have published reviews relating to the intersection between dementia and DHTs (Table 3), including authors who have published ≥3 papers. Seven (0.22%) researchers had published 5 documents based on the identified reviews (Rose-Marie Dröes, Julian Hirt, Minmin Leng, Meyer Gabriele, Ita Daryanti Saragih, Jing Wang, and Zhiwen Wang). Column 4 of Table 3 pertains to the author coauthorship networks, particularly related to the density visualization of the number of publications by the author (Figure 4). Some of the 63 (2.04%) authors in the network were not connected to each other, and the largest set of connected authors consisted of 31 (1%) authors. In Figure 4, the redder the color, the higher the number of publications from the determined investigator, and the closer the distance between authors, the greater the collaboration intensity. With regards to the total link strength between identified authors, the most substantial co-occurrence network was performed by Minmin Leng (n=11), Mingyue Hu (n=11), Julian Hirt (n=10), Li Chen (n=9), and Huiru Yin (n=9).

A total of 2218 institutions or academic departments (organizations) have produced reviews on the interface between dementia and DHTs (Table 4). The analysis evidenced that the College of Nursing, Kaohsiung Medical University, Kaohsiung in Taiwan (5/2218, 0.22% documents) and the Dementia Services Development Centre Wales, Bangor University, Bangor in the United Kingdom (4/2218, 0.18% documents) were the most productive centers in the field of digital health and telemedicine. In relation to the highest total number of citations from each organization, 3 (0.13%) institutions from the Netherlands had the highest citation number (n=396). These institutions are Delft University of Technology, Delft Bio-Robotics Lab, in Delft, Maastricht University, Care and Public Health Research Institute, in Maastricht, and Zuyd University of Applied Sciences, Research Centre for Technology in Care, in Heerlen. Column 5 presents the institution coauthorship networks and the chord diagram of the institution cooperation network. In Figure 5, it is worth reiterating that the size of the nodes represents the number of publications per institution. The distance between each node as well as the thickness of the link represents the strength of cooperation between institutions. Cluster analysis is fully described in Table 4. However, the identified clusters represent 1 (in red) institution leading in technological innovation and multidisciplinary research in DHTs; 2 (in green) institutions that focuses on health care implementation, gerontology, and ethics; and 3 (in blue) institutions at the intersection of nursing and applied psychology.

On the basis of the co-occurrence network, a total of 2808 indexed keywords were found with each keyword being used at least 20 times. Table 5 illustrates the keyword co-occurrence network of reviews focusing on dementia and DHTs. The network map contains 5 clusters of 101 items, 4159 links, and 43,365 total link strengths. “Human” was the most frequently occurring keyword (523/2808, 18.63%), followed by “dementia” (385/2808, 13.71%), “systematic review” (327/2808, 11.65%), “review” (327/2808, 11.65%), and “aged” (181/2808, 6.45%). Table 5 presents the top 97 indexed keywords. These keywords are divided into 5 clusters (Figure 6), and its highlighted topographical zones.

As seen in Figure 6, cluster 1 (in red) represents all technological innovation and multidisciplinary research groups, characterized by groundbreaking institutions at the forefront of technological innovation in DHTs. Cluster 2 (in green) includes institutions that represent health care implementation, gerontology, and ethical considerations, while cluster 3 (in blue) represents evidence groups working on the interface of nursing and applied psychology groups.

In 663 reviews analyzing the interface between dementia and DHTs, 42,758 references were cited (Figure 7). On the basis of the statistical assessments performed on VOSviewer to create our co-citation network (parameters: 3 for a minimum number of citations of cited references), 439 (1.02%) cited references were identified. Some of the 439 cited references in the obtained network were not connected. The most extensive set of connected references was 433 (1.01%) (Figure 7). Table 6 presents the top 15 most cited references among the 663 included reviews. The most cited reference was by the WHO, titled “Dementia: a public health priority,” published in 2012 [19].

This review indicates that there has been an increase in both dementia- and DHT-related publications since the first review in 2002, with a more prominent growth observed after 2017. Despite the diverse approach of DHT modalities for dementia care, authors have focused on mobile health, virtual reality, AI, and sensor technologies. Analysis revealed prominent authors and institutions studying the intersection between digital health and dementia. High-income countries were identified as the leaders in publishing and maintaining research collaborations on this topic. Overall, this analysis provides an initial point of reference for policy makers and academics who wish to conduct research or collaborate on funding calls to understand the contemporary landscape in a bid to undertake innovative and applied research that will primarily benefit people living with dementia, their caregivers, and service provision and resources. We particularly recommend prioritizing the exploration of underrepresented DHTs for research questions and expanding research in low- and middle-income countries to ensure alignment with global equity principles. In addition, it is essential for the scientific and governmental communities to develop frameworks that effectively integrate DHTs into dementia care, guided by the principles of accessibility, affordability, and user friendliness. This review has illustrated emerging global trends and identified areas where future research should focus, especially considering the current emphasis on high-income countries and the remaining knowledge gaps.