In Canada, approximately 87% of people living in residential long-term care homes, or nursing homes, have dementia or another type of cognitive disability [1]. This percentage is expected to increase as the number of people older than 65 years is predicted to grow by 68% over the next 20 years [2] and the number of people living in North America with dementia is anticipated to increase by 63% [3]. The subsequent increase in the number of older adults living with cognitive disabilities is driving health care decision makers to consider the use of automation and novel technologies in an effort to decrease the cost of care while improving safety, quality of care, and efficiency in long-term care [4].

Real-time location systems (RTLS) consist of a software app and reference points that detect and synthesize positioning data from wireless tags or transmitters attached to objects or people. RTLS are being adopted in health care with the goal of improving health and safety. Uses of RTLS include monitoring handwashing [5], fall prevention [6], tracking of individuals [5,7] and assets [8], enhancing independence [9,10], and more recently for collecting health data [6,11-13]. RTLS use tags or badges attached to patients or assets, receiver devices, and software to continuously and inconspicuously monitor activity in real time over wireless networks.

The rising interest in the use of RTLS in long-term care homes is related not only to the ability to remotely track the movement of individuals and assets [5] but also to its potential to gather clinically significant objective streams of data to augment care providers’ subjective observation [6] and to increase safety and productivity in the face of an aging workforce [14,15]. RTLS are increasingly perceived as having the potential to enhance independence and improve the physical safety of residents living in long-term care while also reducing the widespread use of restrictive measures such as antipsychotic drugs, physical restraints, and secure units [16-18]. These first-generation smart systems [19] that use wireless geolocation have been successfully implemented in other industries such as insurance [20] and telecommunications [21] but present novel challenges when deployed in health care, particularly when they are also used to generate personal data from vulnerable people [22]. In fact, although the intent of care providers and organizations is to use RTLS to improve the lives of people living with dementia or other types of cognitive disabilities, their use may provide the opposite result by threatening residents' privacy and restricting their activity within the range of the technology [23].

Industry has leveraged the ability to collect, store, combine, and analyze large quantities of data about customers, despite data sets being anonymized. The disclosure of sensitive personal information, such as location data, poses reputational risk and risks to employability, insurability, and even civil liability. Moreover, the probability of harm increases with time and the frequency of data collection [24]. In health care, critical approaches to the implementation of monitoring technologies suggest the need for institutional accountability for any increased harm from their implementation, particularly for vulnerable populations [25]. Furthermore, there are recognized global standards around privacy for persons with disabilities, such as the United Nations Convention on the Rights of Persons with Disabilities, Article 22, that protects against privacy intrusions regarding the place of residence or living arrangements [26], which should also guide the implementation of monitoring technologies. Pursuant to recent criticism that the ethical standard of autonomy based on independence and noninterference is less relevant in long-term care [27], the introduction of RTLS into institutional settings requires a cautious approach that identifies all stakeholders’ perspectives (eg, residents, residents’ families, care providers, organizations, and society in general) and accounts for the benefits and risks to each that may result. Implementation of these technologies should only proceed when all risks have been accounted for and mitigated or when they are outweighed by agreed-upon aggregated benefits.

The question as to whether RTLS are valid, reliable, accurate, and adoptable in long-term care settings, particularly for the purpose of optimizing independence and safety for residents living with dementia or other cognitive disabilities is not well reported. In this paper, we explore the factors influencing the implementation, adoption, and use of such systems by conducting a systematic review to identify related literature from academic and peer-reviewed journals. The study of health care interventions, particularly focusing on the understanding and evaluation of their implementation, is important for researchers and health care managers. The interaction of agents such as organizations, care providers, and residents, their contexts, and their processes (both established and introduced by an intervention or novel technology) must be anticipated or, at the very least, accommodated in a successful implementation [28,29]. The authors are part of a large team of researchers exploring whether clinically relevant physical, cognitive, and mental well-being information about residents living with dementia can be derived from geolocation data. Before introducing RTLS software to various pilot locations that provide residential long-term care for adults with dementia or other cognitive disabilities, the authors seek to better understand the factors influencing the success of RTLS implementation, particularly in long-term care settings (eg, residential long-term care homes, long-stay psychiatric care facilities).

In collaboration with a Medical Library Information Specialist, we searched the electronic databases Embase (1974: week 25, 2019), CINAHL (1981: August 3, 2019), MEDLINE (1946: August 3, 2019), and PsycINFO (1967: August 3, 2019). Search strategies included subject headings and keywords related to 3 concepts that were combined using AND OR. The search concepts were (1) illnesses, diagnoses, and disorders associated with cognitive disability; (2) long-term care settings (eg, long-term care homes); and (3) RTLS (eg, GPS, sensor-based systems). The Ovid MEDLINE search strategy is given in Multimedia Appendix 1. We restricted searches to adult populations (18 years and older), and primary research studies and reviews published in English. The hand-searching process included reviewing relevant journals and references of included studies and searching Google Scholar to identify unindexed references; selected papers were included in the screening process. An updated search was conducted on all databases, and citations of selected papers, to April 16, 2020, resulting in the addition of one paper.

We included primary research studies and reviews that (1) used and/or described issues relating to RTLS, (2) focused on adults 18 years or older with cognitive disability, and (3) focused on long-term care homes or other types of residential facility settings. In the screening process, we excluded studies that (1) did not satisfy the inclusion criteria; (2) involved, or referred to, children or youth (17 years or younger); (3) focused on individuals without cognitive disabilities; or (4) were conference abstracts, editorials, and commentaries.

Before commencing the screening process, a calibration exercise was conducted to ensure reliability in correctly selecting paper for inclusion. This process entailed 4 researchers (YK, AG, JM, and AB) independently screening a random sample of the references. This process was repeated until unanimity was achieved; any conflicts were resolved through discussion. Titles and abstracts of all references were then screened independently for inclusion by at least 2 reviewers (YK and AG or JM). Conflicts were resolved via discussion or by a third reviewer. All references that were found to be eligible or potentially eligible underwent full-text screening to confirm eligibility, papers were screened independently by 3 reviewers (YK, AG, and JM), and conflicts were resolved by discussion.

Data and information from publications were systematically extracted using a prepiloted data extraction form created in Microsoft Excel. Extracted information included publication information and type, methodology, theoretical or conceptual frameworks used, target population characteristics, study participant characteristics, institution information, and nature of RTLS. One reviewer independently extracted this information (YK), and the extracted data were reviewed by at least 1 other reviewer (JM or AG). All included publications were critically appraised for quality using the corresponding JBI quality assessment tool for the publication type and/or study design [33]. One reviewer independently assessed quality, and assessments were confirmed by at least 1 other reviewer (JM or AG).

Selected publications were imported into NVivo 12.0 software (QSR International) for data management and analysis. An initial set of categories were identified a priori during the scope determination phase of the review; these provided the basis for the initial round of descriptive coding (eg, resident, outcomes, risk). We progressively developed new concepts, and codes were then added, updated, or deleted and then later combined into larger themes. By applying constant comparison techniques, we iteratively revised the coding framework and simultaneously coded and analyzed the data [34,35]. This began with 3 authors (JM, AG, YK) independently reading each paper several times to identify links between them and conducting an initial round of coding using the initial set of codes and identifying new ones (eg, open coding). We then discussed and further refined the coding framework [36]. This was followed by several rounds of axial coding and selective coding comparing and contrasting findings across studies, identifying patterns, and conceptualizing broader themes that were then iteratively discussed among the authors (JM, AG, YK), with conflicts resolved via consensus [37].

Implementation barriers are those factors that, if present, will actively impede the adoption and use of an RTLS. They are aggregated under 3 categories: lack of motivation for engagement; technology ecosystem and infrastructure; and myths, stories, and shared understanding. All are relevant to the resident, frontline care provider, and/or organizational contexts.

Implementation enablers are those factors that, if present, will actively facilitate the adoption and use of an RTLS. They are classified into 4 categories: understanding local workflows, policies, and technologies; usability and user-centered design; communication with care providers; and establishing policies, frameworks, governance, and evaluation.

The implementation of an RTLS in any setting is influenced by what people do (agentic contributions) and the relationship between agents (individuals and groups) and their contexts (spatial, organizational, and normative). In long-term care homes, this is made more complex by the tensions and opportunities for actions that arise as a result of the goals and values of different stakeholders or groups of agents (eg, residents, family members, care providers, and organizations), their relationship with each other, and their relative decision-making power with respect to the adoption and use of the RTLS. Despite this complexity, the collected papers rarely acknowledge these tensions, and for the most part, explore the motivations and actions of care providers and organizations in isolation, largely divorced from a consideration of the relationship between them, with others, or with the context. This is problematic, as research in other congregate settings for older adults on monitoring technologies demonstrates that the failure to account for the mismatch in priorities between stakeholders may lead to resistance and discontinuation [57].

Health care organizations are bound by workplace safety and data privacy laws and regulations that struggle to keep up with novel technologies [68]. The literature suggests that care providers and residents must rely on the culture and values of each individual organization to guide the adoption of RTLS. Our analysis suggests that there is interest among providers and organizations to better integrate nascent RTLS into work routines that improve safety and quality of care. However, there are numerous barriers to their effective use, including lack of engagement, trust, shared understanding of goals, and a reliable and appropriate RTLS that meets the needs of residents and care settings. A variety of factors to improve implementation of RTLS in the long-term care setting should be considered, including integration with local workflows, policies, and technologies and ensuring that care providers, residents, and families are involved, and if possible, leading that process [65,69]. Given the financial constraints in this sector, policy makers should consider the creation of a variety of incentives to encourage the use of technology implementation best practices, such as improved staff training and technology infrastructure development, to optimize outcomes and impact on care quality [70].

Decision makers must also be aware that daily interactions with care providers are often the primary source of social contact for many residents, and thus, the replacement of human contact with surveillance technology may unintentionally increase social isolation [45]. Where possible, institutions should select a flexible, interoperable RTLS that allows customization to suit individual residents, settings, and technological ecosystems [43], understanding that resident and operational needs change over time [27]. Technology developers need to be sensitive to the financial constraints that dominate this sector [43] and aim to build technologies that are elegant in their simplicity and easy to use and focus on solving issues at a lower total cost of ownership (eg, financial, infrastructure, effort, privacy) [40] to institutions, providers, and residents. Issues such as poor battery life, lost tags, and interoperability are solvable with current technology [15]. Researchers are well advised to devise RTLS research protocols and pilot studies that prompt all relevant stakeholders to engage meaningfully with the technology.

The ethical use of RTLS in long-term care settings is similar to other decisions made by providers and management, in that it requires an assessment of the potential benefits and harms, both direct and indirect, to all stakeholders involved. However, this is complicated, given that persons living with dementia or other cognitive disabilities are not assessing the possible risks and benefits independently. Many preferred to defer to family members and other proxies rather than support residents in making independent choices about the use of monitoring technology [4]. Although substitute decision makers are legally mandated to make choices on behalf of residents who lack capacity, this does not absolve them, researchers, and care providers from asking whether the use of an RTLS is ethical, whether it is something residents would want, and what constitutes an acceptable risk as a result of their implementation. It is clear that the use of monitoring technologies with individuals who are nonverbal or have cognitive disabilities, to address issues such as mobility and independence and improve predictive clinical diagnostic capabilities, has preceded our understanding of the rights, risks, and unintended consequences of their use in long-term care. A complete assessment of these issues is recommended before any implementation, drawing on a growing body of work exploring ethical design and implementation of technologies for vulnerable populations [71,72]. Furthermore, it is optimal to anticipate conflicts of interest, such as incidental monitoring of care providers, before they become barriers to a successful implementation [45], and to explore other mechanisms for workforce management [43]. Although this study focused primarily on monitoring residents with RTLS, care providers were also monitored either directly with tags or indirectly through the imputation of activity related to residents, such as the time it took them to respond to an alarm. Although RTLS may offer data to help assess and optimize clinical workflows, automatically open and lock doors, and improve operations [15], it also has the potential to reduce employee privacy and may negatively affect care provider recruitment and retention [43] in a sector that already struggles to find qualified staff [73].

Although the quality of selected papers was acceptable, the limited number of studies and their small sample sizes reflective of qualitative research methodologies and exploratory research suggest that generalization of our results and recommendations for future research should be limited to similar population samples and sectors. Furthermore, 3 of the studies that were included in this review [38-40] contributed a limited amount of data and information to the main research objective of examining factors influencing the implementation, adoption, and use of RTLS.

There is a striking lack of evidence to support the justification and implementation of RTLS to improve the quality of life of residents and work of care providers in long-term care settings. More research related to RTLS use with individuals with cognitive disabilities is required and should include longitudinal evaluations of end-to-end implementations that are theoretically informed and include rigorous analysis of functionality, efficiency, and effectiveness in improving outcomes that are important to all stakeholders involved [4,42]. Empirical studies that rigorously evaluate the practical utility and adoption of RTLS and their related processes into a controlled environment, the value of customization to the requirements of individual residents, and technology infrastructure versus one-size-fits-all adoptions are also required to advance our understanding of their utility.

The use of RTLS to support workflow efficiencies, manage person-to-person contact, and collect clinical data for use in diagnosis and therapeutics is largely unexplored and offers opportunities for future research and use. The workload and care provider capacity for real-time monitoring and data management and analysis for optimal use and outcomes must be made explicit and included in cost-benefit analyses that precede the purchase and adoption of these systems. Furthermore, training for the operation and use of RTLS will require incremental skills training and increased staffing levels, at least in the short term, in a sector where availability of resources and high workloads are already problematic issues. Finally, ethical considerations related to monitoring residents with RTLS, and also directly or indirectly their care providers, are acknowledged but not settled and require further empirical research.