Current health care systems are being pushed to use the capabilities of modern technology outside the hospital to increase efficiency and effectiveness of health delivery [1]. Transforming care processes by using digital platforms and remote monitoring may help address our increasingly older population and higher life expectancies [2]. Smart home and health monitoring technologies have been touted as the future for managing chronic diseases and allowing people to age in place [3-6].

With the technological advancements in the Internet of Things and the widespread use of machine learning and artificial intelligence, the application of smart home technology for aging in place has become more realistic and feasible. Numerous studies on technology development [3,7-10], clinical applicability [11-14], and user considerations [15-17] have been conducted to demonstrate that the technology is ready for mainstream use. There is a plethora of clinically evaluated activity and health recording devices readily available in the market, including wearables (eg, wrist bands, chest bands, and textiles) and ambient sensors (eg, motion sensors, cameras, and pressure sensors) [18-21]. However, the widespread adoption and development of health monitoring platforms remain limited for 2 reasons.

First, the evidence remains siloed within disease-specific reviews or secondary prevention, whether in heart failure [22,23], chronic obstructive pulmonary disease (COPD) [24], diabetes [25], or cardiometabolic health [26], making it difficult to compare effective monitoring models, delineate the overall evidence for home monitoring, and identify where gaps remain [27-29]. There is also a need to differentiate automated monitoring from user-based monitoring, which involves patients texting or phoning in their results.

Second, user and ethical concerns related to monitoring technologies remain major barriers to user adoption, particularly in research [30-33]. Much of the smart home research is focused on the technical development of devices rather than the reliability and usability of smart home systems [34]. Although the feasibility of smart home technology is high in most studies, the acceptability is a critical factor [35]. The benefits of the technology are touted without considering if, how, and at what cost a user may be willing to integrate the technology in their lives [17,36]. In addition, concerns related to data privacy and control, autonomy, and social connectivity are sometimes neglected when designing such systems [37-39].

This review aimed to map out the literature on two major research questions: (1) what is the evidence for the effectiveness of home-based patient monitoring technologies for improving the health and well-being of older adults and (2) what are the ethical concerns that older adults have with home-based patient monitoring technologies, and what frameworks have been proposed to address these concerns? By addressing both questions, we aimed to provide a tool for researchers in this field to understand what needs to be studied and how to study them while keeping in mind ethical and user-centered practices.

In this study, we defined home health monitoring as the use of technology, omitting telecommunications, to monitor the health of users over time (ie, remotely) [40]. This could include using a variety of technologies, including wearables and ambient sensors, to track physiological parameters, activity levels, and routines or to facilitate rehabilitation and treatment [40].

The first part of this review outlined the clinical evidence for using home health monitoring through a scoping review of systematic reviews on home monitoring interventions. We highlighted the domains being researched; performed quality assessments; and determined whether the evidence is positive, neutral, or negative for home monitoring.

The second part of this review considered ethical concerns when researching and developing smart home monitoring technology and user-centered frameworks that address issues of acceptance and adoption. We outlined the ethical and user-centered frameworks that are available to improve these trials and described the current level of adoption of these frameworks in smart home monitoring.

Two researchers (AC and RC) conducted a probing search on MEDLINE for studies using smart homes related to older adults. We decided upon a summary of search terms with consultation with an interdisciplinary team of clinicians and engineers. We completed a systematic scoping search on Scopus, MEDLINE, Embase, and CINAHL in July 2021. The search focused on systematic reviews and meta-analyses using smart homes and remote monitoring of older adults (Table S1 in Multimedia Appendix 2).

Three researchers (AC, RC, and KR) completed abstract screening and full-text screening. Inclusion criteria included systematic reviews written in English with >80% articles focused on older participants (aged >65 years) and published from 2010 to 2021. We excluded studies that focused on assistive technology, mobile health interventions, telephone- or videoconferencing-based interventions, mobile phones and apps, virtual reality, and robots because we focused on automated sensing technologies with clinician intervention. We also excluded studies in which >80% of the articles involved users texting or phoning in the results, rather than using automated monitoring systems. Narrative reviews and technical articles outlining the implementation of these technologies were excluded. We excluded studies that were not journal articles because our focus was on identifying gaps rather than estimating effect sizes.

We created an extraction table and completed extractions independently. The extracted items included population characteristics (number of articles, included diseases or disorders, and percentage of articles including older adults), study type (meta-analysis and quantitative or qualitative study inclusion), and monitoring methods (automated monitoring, mixed automated monitoring, and user-reported monitoring).

We extracted outcome measures, including physiological outcomes (vital signs and blood tests), symptoms or health events (falls, exacerbations, or mortality), health care use, cognitive decline, functional status, adherence to rehabilitation, and activity levels. Studies on human factors included user perspectives on the ethics of home monitoring, the acceptability and usability of devices, and changes in quality of life (QoL) or social connections through technology. We coded outcomes as positive, negative, mixed, or “not enough evidence” according to the author’s assessment and summarized in tables. Finally, we extracted the challenges that the author listed on health monitoring research, ranging from privacy and security concerns and technology acceptability to technical challenges and lack of clinical evidence. The challenges were coded based on what the authors reported. We developed a spreadsheet with a separate column for each challenge that each author would raise. If a new challenge was listed, it was added to a separate column. After we added all the articles and challenges, the 3 authors reduced the number of challenges by grouping together similar challenges by consensus.

Four authors (AC, RC, KR, and DB) conducted a quality appraisal using the AMSTAR-2 checklist and appraised independently after deciding upon critical, important, and unimportant categories. Categories rated as “Critical” included a comprehensive literature search strategy, describing included studies in adequate detail, and accounting for the risk of bias on evidence synthesis. Categories rated as “unimportant” in quality assessment included explaining the selection of study designs, providing a list of excluded studies, and reporting sources of funding for studies included in the review. All the other items were considered important. Studies with ≥2 “No” ratings on critical categories were considered critically low quality, whereas studies with one “No” rating were considered low quality. Moderate-quality studies had ≥2 “important” categories scored as “No,” whereas high-quality studies had <2 “important” categories scored as “No.” The entire list can be found in Multimedia Appendix 3.

To answer the second research question, we conducted 2 searches. The first search focused on the ethical challenges and concerns related to smart home technology. We conducted a systematic search from April to May 2021 using Scopus, Web of Science, and Dimensions AI. The search terms are listed in Table S2 in Multimedia Appendix 2. Once we collected all the papers, we screened the titles and abstracts to select the papers to be used for this review. Studies published before 2015 were excluded.

The second search focused on studies involving stakeholder participation while developing health care services and apps. We used Google Scholar to gather related chapters, journals, and articles with keywords (Multimedia Appendix 2). We excluded studies with health care frameworks that did not focus on user centeredness (ie, stakeholder involvement).

The search yielded 1022 articles, which was reduced to 822 after deduplication (Figure 1). Screening abstracts yielded 480 articles, whereas full-text and additional screening during extractions yielded 94 systematic reviews. Most articles were excluded because they did not focus on telemonitoring technologies (325/728, 44.6%), followed by articles that did not focus on older adult users (120/728, 16.5%) and that were not systematic reviews (104/728, 14.3%). The extraction procedures are presented in Multimedia Appendix 4.

Figure 2 shows an increasing trend for systematic reviews on home-based patient monitoring technologies, particularly starting in 2018. Live and automonitoring has only become more commonly reviewed since 2018. The number of systematic reviews showed an upward trend starting in 2018.

Figure 3 displays the study designs used in each systematic review, organized according to the outcomes discussed. Many articles have presented >1 outcome category. Explanations of the outcomes are described in the methodology.

Meta-analyses were most common when reporting physiological evidence, health events, and health use (9/18, 50.0% for physiological; 18/31, 58.1% for health events; and 20/30, 66.7% for health use). Studies focusing on cognition, safety, and activities of daily living (ADLs) were mostly mixed methods or nonpatient studies (13/18, 72.2%), whereas studies focusing on exercise were mostly meta-analyses or quantitative clinical studies (10/15, 66.7%). Most systematic reviews on ethics, acceptability, and usability used mixed methods or nonpatient studies (4/5, 80% for ethics; 20/32, 62.5% for acceptability; and 19/24, 79.2% for usability), and no single study design dominated QoL or social-focused studies.

Figure 4 displays the disease processes studied according to the category of disease monitoring. Acute prevention studies focused on reducing heart failure and COPD exacerbations and fall prevention (24/61, 39.3% reviews). Chronic management studies focused on blood pressure reduction, blood glucose control, and metabolic disease management (13/61, 21.3% reviews). Home monitoring studies focused on monitoring aging and status of patients with dementia (13/61, 21.3% reviews). Studies on physical activity monitoring focused on rehabilitation or increasing physical activity in the older population (11/61, 18.0% reviews).

The challenges related to current evidence and implementation of home monitoring were also analyzed. Many studies listed multiple challenges, whereas 9.3% (9/94) studies listed no challenge. The most common challenges in the literature included a lack of strong clinical evidence for monitoring (510/94, 53.1%), poor descriptions of methodologies of how patients were monitored (27/94, 28.7%), and applicability to broader patient populations (10/94, 10.6%). On the implementation side related to human factors, the acceptance of technology (33/94, 35.1%), usability of devices (18/94, 19.1%), privacy concerns (17/94, 18%), cost-effectiveness (23/94, 24.2%), and safety concerns with devices (7/94, 7.6%) were listed. Technical challenges were not as commonly reported, although concerns with accuracy (10/94, 10.7%) and connectivity of devices (9/94, 9.7%) were more common than the others.

Figure 5 shows the AMSTAR-2 quality assessments organized according to the monitoring category. Most studies were either of critically low or low quality across all systematic reviews (36/94, 38.3%, and 35/94, 37.2%, respectively). Most studies did not discuss the impact of risk of bias on results (57/94, 60.6%), did not discuss heterogeneity (55/94, 58.5%), or did not include an explicit statement on following a protocol (50/94, 53.2%).

Table 1 reports the outcomes from articles that focused on clinical outcomes, as shown in Figure 4. The disease-specific details are presented in Multimedia Appendix 5. In total, 51 of 64 (80%) or more systematic reviews reported positive results across all categories. Chronic disease management and physical activity were among the categories with the largest number of systematic reviews with positive evidence. Evidence was most limited in managing degenerative diseases, health events, and use in rehabilitation interventions.

The first theme involved detecting acute events including exacerbations of heart failure, COPD, or falls (Table 1 and Table S1 in Multimedia Appendix 5). Most reviews reported on exacerbations of heart failure and COPD. Monitoring reduced hospitalizations for patients with COPD (5/7, 71.4% studies) and for patients with heart failure (10/14, 71.4% studies). Mortality was unchanged for COPD in a few reviews (3/7, 42.8%) but for most reviews for heart failure (11/14, 78.6%). Only 1 study had reviewed fall interventions in an older population [46]. No clinical studies have focused on atrial fibrillation, although devices showed high specificity and sensitivity in detecting atrial fibrillation.

Management of chronic diseases included managing diabetes, blood pressure, kidney function, or multiple diseases simultaneously (Table 1 and Table S2 in Multimedia Appendix 5). In most studies, home health monitoring helped to reduce hemoglobin A1c (8/10, 80%) and blood pressure (5/9, 55.6%), although the results were more mixed when only considering high- or medium-quality studies. In total, (2/4 (50%) of studies reported reduced mortality, reduced health events, and fewer hospitalizations when monitoring chronic cardiovascular diseases.

Smart home monitoring for ensuring safety of older adults with dementia or cognitive impairment had scarce evidence (Table 1 and Table S3 in Multimedia Appendix 5). Only 1/2 (50%) of studies found no reduction in admission to care homes for monitored patients [87]. In addition, 2/5 (40%) of studies noted progress in the ability to detect cognitive decline [90] and aggression [91]. Most studies found that technologies were not mature enough to detect activity changes or improve independence [88,89,92,105]. For monitoring older adults in general, % (3/N) of studies noted weak evidence for detecting changes in ADLs [82-84], and 2 studies showed improved QoL [85,86].

The fourth theme, rehabilitation adherence and encouraging an active lifestyle at home, included 11 studies (Table 1 and Table S4 in Multimedia Appendix 5). All studies showed positive outcomes for both disease-specific rehabilitation programs and for monitoring older adults in general. Monitoring improved adherence to cardiac rehabilitation, increased activity levels in patients with COPD, and improved motor control in patients after stroke [94-96]. Monitoring increased physical activity in older adults in general in 4 studies, and telerehabilitation with monitoring was found to be as effective as traditional rehabilitation [104]. Wearables helped to increase activity in patients with cancer, improve functioning in patients with arthritis, and improve QoL in postoperative patients [97-99].

As part of the scoping review, we included studies that focused on human factors such as acceptability of technology, ethical considerations, and costs. Although this was not the focus of question 1, we performed a basic analysis of the outcomes from these studies, recognizing the importance of human factors related to remote monitoring. Table 2 summarizes these 33 studies.

Of the 24 studies on acceptability, 8/24 (33%) showed good acceptability of monitoring technologies in general. For chronic diseases, 5/9 (56%) showed good acceptability in monitoring for heart failure, COPD, and management of amyotrophic lateral sclerosis, whereas 2/9 (22%) studies on heart failure and COPD were inconclusive. Monitoring acceptability was less clear for generalized home health monitoring and rehabilitation for older adult patients with fatigue. Costs, ethical considerations, and QoL were inconclusive or were descriptive studies.

Although there appear to be positive results from specific diseases, there is a need to continue studying whether remote health monitoring is acceptable for the older adult population. Human factor considerations have been prominent in several studies and deserve closer inspection. Applying appropriate frameworks to the design and development of home health monitoring technology that address human factors such as users’ ethical concerns, like privacy and usability, need to be explored to ensure that users feel at ease while using and installing devices that monitor their lives. Therefore, the second part of our review focused on the ethical concerns that older adults have regarding home health monitoring technology and frameworks that have been suggested to address such concerns.

The second research question in this study was “what frameworks have been proposed to address the ethical concerns that stakeholders, specifically older adults, have toward home health monitoring?” To learn which frameworks are available to address the ethical issues that arise with home health monitoring, the ethical issues themselves must be uncovered and discussed in more detail.

To explore the ethical issues, we conducted a literature review of ethical challenges. An initial search of 132 papers published since 2015 was conducted, focusing on the ethical challenges and concerns regarding smart home technology. After screening titles and abstracts, 19 papers discussing the ethical challenges and concerns regarding smart home technology were identified. The most discussed concerns are listed in Table 3. The papers were a mix of systematic reviews, literature reviews, qualitative research (including focus groups and interviews), and mixed methods research (qualitative and quantitative research, many using surveys to obtain results).

Health care tools are often designed without considering the intended users, their literacy levels, information goals, and preferences, which results in user dissatisfaction, leading to the abandonment of the technology and eventually the need to redesign it [150]. Despite constant advancements, web-based medical platforms have low adoptability or acceptance rates [1]. A reason for this is inadequate user involvement, especially in the early development phases [151-153]. We should aim to tailor the technology to the stakeholders’ needs and requirements to increase their usability and utility [154]. User-centered design (UCD) is a framework in which the requirements of stakeholders (including end users) are considered extensively at every stage of the product’s development and design [157]. Table 5 summarizes the various user-centered frameworks developed and adopted in various health care settings based on systematic reviews and individual articles. Moreover, the use of UCD-based evaluation instruments, such as the 11-item measure “UCD-11,” developed by Witteman et al [154] to quantitatively determine the user centeredness of the design and development of health care tools will ensure production of reliable and valid constructs.

The literature on home monitoring is limited. First, the study quality was limited to secondary prevention in heart failure and COPD from 7 high- or medium-quality studies and encouraging rehabilitation and physical activity in 3 studies. Most systematic reviews were of low or critically low quality. We have presented the results of a subgroup analysis of only high- and medium-quality studies to mitigate our own risk of bias. Higher-quality systematic reviews are critical for chronic disease management and generalized monitoring.

Second, the older adult population is not always the focus of reviews. We found only 1 fall detection study in older adults using monitoring devices. A technology survey of fall detection systems found that only 4 out of 57 studies even included older adults in their study [170]. In the screening process, >120 reviews were excluded, as they did not focus on adults aged >65 years. It is important to differentiate these populations, as the acceptability of technology and challenges of multiple morbidities may change the effectiveness of monitoring.

Third, the evidence was sparse for monitoring older adults in general or patients, with most studies suggesting that the technology is not yet mature enough to detect ADLs or cognitive decline [89,92,93]. Most acceptability studies have focused on the domain of general home health monitoring, with the majority being descriptive or showing mixed acceptability. The question of whether it is worthwhile to monitor older adults in general or those with dementia remains unanswered. In addition, telemonitoring and telemedicine are often not differentiated, making it difficult to differentiate the benefits of clinician support from devices and technological support.

From an evidence-focused perspective, monitoring is effective for specific diseases, but challenges remain in researching generalized monitoring for older adults. Clinical outcomes are difficult to measure because objective physiological markers of aging and health utilization need to be measured across years to determine effectiveness. Second, technical obstacles remain in acquiring data from multiple sensors, synchronizing outputs in real time, storing the data, and performing data analytics to detect anomalies or track wellness, all of which can be time consuming and expensive. Finally, the drivers of long-term monitoring are older adults or families focused on the personal benefits of aging in place. Outcomes focused on improving function and independence and older adults’ sense of security and identity may not draw as much broad attention or funding compared with long-term clinical effectiveness and cost-effectiveness studies. Considering the promising results of remote monitoring in specific apps, it is worth researching broader generalized monitoring to improve clinical outcomes.

The main limitation of our review was that we did not have a 2-person validation for screenings and extractions. This may have introduced a bias in the included studies and the information extracted. To mitigate the validation of screenings and extractions, we used common inclusion and exclusion criteria and communicated with team members regularly on questionable papers. In addition, our search was limited to clinical databases to focus on clinical effectiveness rather than technological developments; this may have biased our results related to human factors and monitoring. Moreover, it is possible that multiple studies were included in multiple reviews, which may have biased the results. However, this cannot be mitigated. Finally, we identified gaps in the data but were unable to quantify the effect sizes.

Regarding the ethics- and user-centered framework portion of the review, we did not follow a scoping review methodology, as we found that the qualitative nature of the findings could be summarized with a narrative review. However, we may have missed some trade-offs regarding the acceptability of monitoring technologies.

An interesting area for future research is the extraction of what acceptability tools were used in each study. For the second half of this review, we did not provide an exhaustive list of the ethical concerns that end users have regarding home health monitoring technology. We also limited our study to older adults and did not consider family members; health care professionals; and other stakeholders such as engineers, computer scientists, and designers. More value tensions will arise and need to be addressed with the involvement of more stakeholders.

This scoping review provided a summary of the clinical evidence for monitoring older adults in their homes, ethical implications, and user-oriented frameworks found in the literature. Overall, there is promising evidence for monitoring specific diseases and for rehabilitation support, but generalized monitoring for older adults, including cognitive and physical decline, has not been well researched. More clinical research is required for the long-term monitoring of aging in place to provide evidence for its use. To conduct these future studies, we performed a review of important ethical and user considerations and existing user-centered frameworks that must be considered when conducting these monitoring studies. This study demonstrated the need to develop technology with stakeholders rather than for stakeholders to build the evidence for home health monitoring. User-centered frameworks allow stakeholders’ ethical concerns to be addressed and open iterative design opportunities to improve adoption. In developing a system that achieves ethical and UCD, researchers can collect long-term, meaningful data to demonstrate the efficacy of home health monitoring systems for aging in place.