Dementia affects more than 50 million people worldwide [1] and this number is expected to triple by 2050 [2]. To reduce the tension between an increasing demand for care and the growing shortage in residential care capacity [3], many countries have shifted their attention toward deinstitutionalization and to support people with dementia to live at home for as long as possible [4,5]. Although extended independent living is preferred by most people with dementia [6], this also puts more pressure on their informal caregivers [7] such as spouses, children, or other relatives providing unpaid care at home. The volume of informal care has already been relatively large in most European countries, making up most of the care received by those aged 50 years or older [8,9]. Informal caregivers of people with dementia can feel heavily burdened by their care responsibilities, often resulting in stress-related symptoms such as anxiety [10], mood, or sleep disorders [11]. The strain on informal caregivers and people with dementia has become even more present in light of the COVID-19 pandemic as routine professional care services were postponed or decreased [12], inevitably causing a greater reliance on home care as one pillar of the health care system [13].

The increasing need for support of both patients and their informal caregivers in the home setting has led to innovative solutions, including those from the field of eHealth technology [14]. In the Netherlands, the National Dementia Strategy 2021-2030 [15] promotes the utilization and further development of eHealth technologies to support both patients and caregivers. The goals of these technologies within dementia care are diverse. A systematic review performed by Ienca et al [16] distinguished several purposes of (smart) technologies for dementia care, including assistance for activities of daily living, cognitive and emotional assistance, health and behavioral monitoring, fostering social interaction, and remote communication and emergency systems. In this review, we consider eHealth as “the use of technology to support health, well-being, and healthcare” [17]. This rather inclusive and broad definition covers all (intelligent) assistive technologies and technology-based interventions that can be used to support people with dementia and their informal caregivers in the home setting.

Despite their promising potential and position within recent policy, the use of eHealth technologies in dementia care is unfortunately still limited [18]. Low adoption rates may signal problems during implementation. In fact, the sustainable implementation of eHealth technologies aimed at providing support in home-based settings is frequently unsuccessful in daily practice [19,20], resulting in these technologies falling into the “valley of death” [21] after the research projects have ended. This raises a question about what exactly facilitates or hinders the successful implementation of a broad range of eHealth technologies in the informal dementia care context, which so far has been given little attention in research [14].

A previous scoping review performed by Guisado-Fernández et al [22] investigated factors influencing the adoption of “smart health technologies” [22] for people with dementia and their informal caregivers. Their review identified attitudinal aspects, ethical issues, design-related issues, and dementia-related challenges as playing key roles. Another review by Christie et al [14] focused on the implementation of digital interventions for informal caregivers of people with dementia. Several determinants were identified, including perceived data security, psychological characteristics of caregivers, care policy, or financial constraints. Both studies illustrate simultaneously that knowledge about factors influencing the implementation of eHealth technologies in informal dementia care is currently fragmented across different studies in the literature. Previous reviews tended to either focus primarily on a specific part or outcome of implementation (eg, adoption or acceptance) [22,23] or zoomed in on a specific type of technology [14,24]. This makes it difficult to obtain a complete overview of supportive eHealth technologies in the context of informal dementia care and what factors facilitate or impede their implementation.

To the best of our knowledge, no review yet exists that aims to summarize the influential factors across the whole spectrum of implementation and related to a broad range of technologies studied in the specific context of informal dementia care. For such a review to deliver beneficial and complete results, we consider the guidance of a holistic view on implementation as essential. Numerous implementation frameworks exist that view implementation as a postdesign phase [25-27], although it is currently recommended to target aspects of implementation at an early stage of development [17]. By contrast, the Nonadoption, Abandonment, Scale-Up, Spread, and Sustainability (NASSS) framework [28] considers the implementation process in a multilevel and comprehensive fashion, which encompasses the influences on the adoption, nonadoption, abandonment, spread, scale-up, and sustainability of eHealth technologies. This evidence-based and theory-informed framework includes both adoption and acceptance from the viewpoint of the stakeholders but also considers aspects of implementation that relate to the wider context [28]. Because of its holistic view on implementation, the NASSS framework was used as a general guide in our study.

The aim of this review was to gain a more complete understanding of why it is often difficult to implement and integrate eHealth technologies in everyday life despite the numerous technologies that have been studied and designed to support people with dementia and their informal caregivers at home. In particular, the complementary study aims were to (1) provide an overview of the types of technologies that have been used and studied in the context of informal dementia care and (2) explore the factors influencing the implementation of those technologies.

The findings of this review are expected to be useful in determining directions for future research, and to help researchers and stakeholders in improving the development and implementation of eHealth technologies to support informal dementia care.

We used an umbrella review design to create an overview of the state of the art of technology for supporting informal dementia care and to identify determinants for its implementation. The methodology of an umbrella review can best be described as a systematic review of reviews, meaning that reviews are used as the analytical unit of the umbrella review [29]. Umbrella reviews are particularly fitting in a broad field of work and aim to provide a summary of the highest-quality studies into the state of the art of a certain domain. With this approach, gaps in the literature are highlighted and principal findings are presented in a tabular and concise manner that can be used in practice [29].

A systematic literature search of five databases (PubMed, PsycINFO, Medline, Scopus, and Cochrane) was performed in June 2020. The aims of the search were to identify reviews that describe (i) technologies that are in use or have been studied to support informal dementia care and (ii) information related to the implementation of those technologies. The search string included terms pertaining to four main categories of keywords: (i) eHealth, (ii) Implementation, (iii) Informal Care, and (iv) Dementia. To build the search string, thesaurus and nonthesaurus terms were used. The search string was then adapted to each database and approved by a library consultant at the University of Twente. In general, the four categories were separated by AND, and the single terms inside of each category were separated by OR. The complete search string and database adaptations can be found in Multimedia Appendix 1.

Included studies needed to be reviews containing information related to the implementation of eHealth technologies for people with dementia and/or their informal caregivers. Studies in English, German, Dutch, Italian, and Portuguese languages were searched. Nonreview articles such as primary studies, randomized controlled trials, books, dissertations, and grey literature were excluded, as recommended for umbrella reviews [30]. Studies that also included non-eHealth interventions or other types of diseases were excluded. Furthermore, studies published before 2010 were excluded as it was believed that

these would not add relevant information since those technologies are likely to have become outdated.

During title and abstract screening, each paper was independently evaluated by at least two reviewers (SB, MS, or CW) and conflicts were solved through discussion. For full-text screening, the same procedure was followed, but consensus was reached through unanimity. Studies that fit all inclusion criteria were examined using a data extraction form on Covidence. The form was used to extract data about study details (eg, title, year, author, type of review, number of included studies), information about included technologies (eg, technology types, purposes, and primary user groups), and any statements related to implementation, including potential barriers and facilitators. Each paper was randomly assigned to at least two of the three reviewers (SB, CW, MS) who performed the data extraction independently from each other. Subsequently, completed data extraction forms for each paper were reviewed and adapted until consensus among all reviewers was reached.

A qualitative thematic analysis was performed on the extracted data using Atlas.Ti9.

The NASSS framework was used to categorize the data. This framework consists of seven domains and can be used to evaluate the success of implementation of a health technology retrospectively and prospectively. Relevant fragments were selected and categorized into (1) one of the seven domains of the NASSS framework or (2) classified as miscellaneous information. Subsequently, selected fragments were further categorized inductively into overarching themes. To minimize single-researcher bias, the coded papers were checked independently by a second researcher. The final coding scheme was developed and defined based on consensus among the three researchers (SB, CW, MS).

Figure 1 illustrates the process of inclusion and exclusion of papers. The search strategy (available in Multimedia Appendix 1) produced 3109 results. After removing 904 duplicates and 2061 papers that were determined to be irrelevant according to our inclusion and exclusion criteria during title and abstract screening, 144 articles were considered for full-text screening. Finally, 21 papers were included for this review. Reasons for exclusions are detailed in Figure 1.

Table 1 summarizes details about the included studies such as author, year, title, type of review, and number of studies included per review. The studies were published between 2014 and 2020 and the majority were systematic reviews. The quality of the included studies was evaluated according to the Critical Appraisal Checklist for Systematic Reviews and Research Syntheses checklist [30], which ranged from 55% to 100% (for further details see Multimedia Appendix 2).

The included studies varied in their usage of implementation-related terminology and focus. For instance, 43% (9/21) of the included studies did explicitly use the term “implementation” in the text. Those that did not often used terminology related to certain subcomponents of implementation such as adoption, effectiveness, acceptability, or acceptance instead. Furthermore, although all studies contained information related to implementation, only 5 out of 21 reviews exclusively acknowledged implementation, or certain subcomponents of it, as the main focus of the study.

Table 2 presents an overview of the number of reviews describing factors related to implementation of technology supporting informal dementia care per the domains of the NASSS framework. Most of the reviews described factors relating to the technology and the condition of dementia, followed by reviews describing factors related to adopters (informal caregivers) and the technology value proposition. Factors relating to the implementing organization, the wider system, and embedding and adaptation of technology over time were the least represented. In conclusion, the most identified factors provide information about how the condition of dementia, the technology itself, its expected and perceived value (demand side), and the informal caregiver might influence successful implementation, whereas considerably less has been reported on factors relating to the implementing organization, the wider institutional and sociocultural of context of policy and regulations, and the continuing adaptation of technology over time.

Our study aimed at gaining a more complete understanding of why it is often difficult to implement eHealth technologies that have been specifically designed to support people with dementia and their informal caregivers in everyday life. According to Bauer and colleagues [49], the main goals of implementation science should be (1) identifying barriers and facilitators to the uptake of innovations and (2) developing and applying strategies to promote the successful implementation of innovations. Our umbrella review integrates knowledge that has been fragmented across different reviews until now by (1) providing an overview of the types of technologies that have been used and studied in the context of informal dementia care, and (2) exploring the factors influencing the implementation of those technologies.

Our review found that regardless of the difficulties that come with implementing supportive technologies, a broad range of existing or to-be-developed technologies are studied in the context of informal dementia care. Similar to Gibson et al [47], we generally identified technologies that aim to be used by different primary user groups, characterized by varying levels of involvement of informal caregivers, ranging from technologies used by informal caregivers to technologies used between caregivers and people with dementia to technologies used by people with dementia. It was possible to identify a certain degree of overlap between the categorization by user groups that emerged in our study and the structure of the NASSS framework, specifically with respect to how certain determinants of implementation refer to the patients (Condition and Adopters) and others to the caregivers (Adopters). However, we find the distinction between the domain of Condition and the subdomain “Patients” (embedded in the Adopter domain) to be less practical.

One of the largest groups of technologies found in our review were monitoring devices, including in-home monitoring of health and safety, and outdoor location identification of people with dementia, showing that this technology domain has developed rapidly and is seen as promising. We found that privacy and ethical issues were frequently mentioned as a barrier in relation to this type of technology; however, ways to overcome this barrier have mostly been unaddressed. In a previous study among potential users, we found that artificial intelligence–driven monitoring systems particularly require introduction in a way that prevents caregivers from feeling undervalued [50]. We have published a set of requirements that can benefit the development and introduction of in-home monitoring technologies aimed at supporting home-based dementia care [50].

An important finding of our study was the uneven distribution of references identified within the 7 domains of the NASSS framework. Although most reviews contained information on how the condition of dementia, the technology itself, its expected and perceived value, and the informal caregiver might influence successful implementation, considerably less has been reported on factors related to the implementing organization, wider institutional and sociocultural context, and continued adaptation of technology over time.

Interestingly, two included reviews came to a similar observation. The review from Christie et al [14], which focused on digital interventions for caregivers of people with dementia, found contextual factors related to implementing organizations and the wider context to be underrepresented in the included studies. Similarly, the implementation factors identified by Thordardottir et al [41] often related to a “micro level” (the individual user), whereas factors on the “meso level” (organizational processes) and “macro level” (national policy context) were less frequently found.

An additional blind spot that emerged from our study was the lack of information on factors related to the supply-side value, which was surprising as business modeling is crucial for the success of an eHealth technology and can serve on a strategic level to guide sustainable implementation [51].

Overall, there seems to be a mismatch between the focus of research performed on supportive technologies for people with dementia and their informal caregivers, and existing implementation frameworks. In our view, this might indicate (1) a misconception or partial mental model of implementation within researchers in the context of informal dementia care or (2) a lower interest in research about the wider contextual factors. In fact, researchers probably prefer to focus more on concrete, well-known, and measurable aspects of implementation instead of focusing more on abstract concepts. Nevertheless, these results could constitute a possible explanation to implementation failures that are very diffused in this (and other) contexts [52].

The identified mismatch between theory and research practice was also visible in the fact that most of the included reviews did not generally identify the use of implementation frameworks in their included studies nor did they employ such a framework to systematize the results. However, the latter could be explained by the fact that many of the reviews did not focus on implementation “as a whole” but rather focused on specific subcomponents of implementation such as acceptance or adoption. One review we included produced an ad hoc framework to guide the design of “smart health technology” [22]. Interestingly, this DemDesCon framework [22] also covers considerations to be made at the user, social, and development levels.

To the best of our knowledge, this review is the first of its kind to explore factors influencing the implementation of eHealth technologies to support informal dementia care at this level of abstraction. By analyzing reviews (that included 840 studies in total) instead of primary studies, we were able to (indirectly) include a large knowledge base. According to the methodology employed, our results could lay the grounds to provide practical insights for decision making in the context of implementation of eHealth to support informal dementia care [30]. An additional strength of this review also lies in the rigor of the data extraction and analysis, with multiple researchers independently screening and analyzing the information.

However, some limitations must be considered. First, our results specifically refer to the context of informal dementia care and therefore are not necessarily generalizable to other implementation contexts. Second, due to the employed umbrella review design, we had to rely on the way review authors have summarized their findings. The level of detail provided within the included reviews varied; not all reviews provided a detailed description of the types and purposes of their included technologies, with some only providing a shorter summary. Lastly, although the authors kept close track of recent literature, due to technical time for finalization and publication of the manuscript, papers that would have otherwise been included might have been overlooked.

Our results suggest that more research is needed to understand how implementing organizations, the wider institutional and sociocultural context, and business modeling influence the successful implementation of technologies to support people with dementia and their caregivers, as many of the included reviews failed to address these aspects. In addition, future implementation research within this target group should increase its focus on continued adaptation and embedding of technology over time. As eHealth technologies to support informal dementia care develop rapidly, it seems essential for implementers not to fall behind the technological progress or eventual changes in context and care standards.

Lastly, our review provides an overview of factors influencing implementation, but it does not differentiate between different types of technologies in this regard. Future research should investigate what is needed for successful implementation of specific kinds of technologies to support people with dementia and their informal caregivers at home.

To help future developers in creating and successfully implementing meaningful technologies for both informal caregivers and people with dementia, we generally recommend use of the NASSS framework in combination with a holistic and iterative development approach, which views implementation not as a postdesign phase but rather intertwined with development right from the start. In light of our results, the CeHRes Roadmap [17,53]—a toolkit to guide holistic eHealth development—seems especially suitable in several ways. First, this roadmap pays special attention to the characteristics of and interrelation between relevant stakeholders, the (wider) context, and the technology. Second, it incorporates evidence-based methods from participatory development and human-centered design. Third, it focuses on cocreation of a business model even before a prototype of a technology is being made. In this way, possible implementation barriers such as those identified in our study can be addressed and accounted for at an early stage of development.

In addition, we present a synthesis of our most important results in the form of checklist (see Multimedia Appendix 3) aimed at promoting reflections and providing insights for readers interested in the field of technologies to support informal dementia care. The present checklist is intended for researchers, policymakers, practitioners, experts, and any other stakeholder interested in technologies for informal dementia care who want to gain specific insights on the implementation process and determinants. Specifically, readers are provided with (i) a concise overview of relevant aspects and domains of implementation identified in this review in light of the NASSS framework and (ii) best practices and recommendations. The first step to use this simple tool is to define the technology (or type of technology) that needs to be implemented, and its primary user group. Readers can make use of Figure 2 to navigate the different options. By filtering the most and least important domains, researchers can concentrate on the most relevant aspects. Moreover, the reader can rate their relevance, the extent to which they were addressed in design or implementation, and make use of the practical insights that directly derive from our review and CeHRes Roadmap [17,53].

The increasing number of community-dwelling people with dementia worldwide underscores the need for innovative eHealth solutions that can provide support to both patients and their caregivers in the home setting. However, sustainable implementation of supportive technologies within this target group can be difficult. Our umbrella review has provided a comprehensive overview of eHealth technologies studied in the context of informal dementia care and contributes to a better understanding of a broad range of factors influencing their implementation. These findings can help researchers and stakeholders improve the development and implementation of eHealth technologies to support informal dementia care. More research is needed to identify the specific factors determining successful implementation related to the wider institutional and sociocultural context, the implementing organization and technology supplier, and continued adaptation and embedding of technology over time.