Since the use of ATs often requires or supports the contact and interaction with relatives, this study considered 2 participant groups: older adults and their relatives or a close friend (hereinafter referred to as relatives). The inclusion criteria for older adults were (1) aged ≥65 years; (2) residence in the area of Ulm (distance ≤50 km); (3) an independent or assistive living situation; (4) the ability to speak and read German; and (5) the ability to consent to the study participation (adults with dementia were excluded). In addition, older adults had to have intact vision and hearing (normal or corrected). The relatives had to (1) be a family member or someone close to the adult; (2) speak and read German; (3) give consent to study participation; and (4) have a smartphone, tablet computer, or PC with internet connection. Participants were recruited in Southern Germany using convenience (via analogous and digital recruitment methods such as flyers, social media, and clinic staff) and snowball sampling (asking participants for other potential participants). As this was an exploratory study, no sample size calculation was conducted. The aim was to recruit between 6 to 10 participants per selected device. This number was defined based on feasibility. For 4 different ATs with 5 available devices each, 2 rounds of the study with 40 participants can be performed within 4 months.

Ethical approval was obtained from Ulm University Ethical Committee (Nr. 230/21, 05.07.2021). Before entering the study, individuals received detailed information about the study and provided written informed consent to participate. Data analyses were performed on pseudonymized data. No financial remuneration was provided for participation.

Initially, 4 different AT were chosen for this interventional study. The ATs were selected from the 2020 product competition “Daheim Dank Digital” (at home thanks to digitalization) [25] aimed at startups and established companies in German-speaking countries (Germany, Austria, and Switzerland). Companies were asked to apply with ATs that could compensate age-related deficits and impairments described in 5 predefined use cases. The use cases were designed together with experts from different areas of expertise and focused on (1) nocturnal restlessness and fall risk; (2) loneliness, hearing and vision impairment, and forgetfulness; (3) inactivity and listlessness in daily routine; (4) urinary incontinence and reduced fluid and food intake; and (5) limited mobility, weakness, and loneliness.

A total of 9 companies presented their products, and a scientific jury evaluated the devices based on availability; readiness for use; and impact in at least 1 of the categories of communication, security, or autonomy. These categories were specified as target areas in the overall project. A total of 4 products were selected for this user study based on their availability. The identified devices focused on safety (an intelligent bed exit alarm from the company NevisQ and a smartwatch with an emergency button from the company CareIOT GmbH) and communication (the Media4Care tablet computer from Media4Care GmbH and the Eldertech app, which supports communication and coordination of care within a family, from Eldertech GmbH). At the time of the study, all products were commercially available and had a CE certificate.

We performed a prospective, real-world, pilot study of 4 different ATs with community-dwelling adults aged ≥65 years and their relatives. Together with the participants, the investigators decided which product was the most suitable for the older adult. Based on their individual needs, a decision was made for 1 of the targeted areas: communication or safety. Unfortunately, no candidates were identified for the evaluation of the bed exit alarm and the app; therefore, the selection was limited to the tablet computer and the emergency smartwatch. Both devices were used in their versions that were current as of August 2021. Details on the main functionalities of these 2 devices are listed in Table 1. Further technical information on the 2 products can be found in Multimedia Appendix 1.

In an introductory assessment session, all participants and their relatives were informed of the study procedure, goals, and possible risks and consented to participate. Baseline data were captured in a subsequent meeting at the participant’s home or in the study center (SM and BK). Participants were given the selected AT and a short introduction to the device. Participants then tested the AT in their daily life over a period of 8 weeks. In case problems with the use of the AT arose, participants were asked to first seek help in the manual, on the web, or via the company hotline before contacting the study center. This was done to reproduce a real-life setting as accurately as possible.

During the study period, the study team called the participants after 1, 2, 4, and 6 weeks to inquire about the use of the product; help solve possible issues; and remind participants to document errors, problems, and the frequency of use. After the test period of 8 weeks, the participants and their relatives were assessed a second time. This meeting took place at the participant’s home or in the study center. The relatives performed an independent assessment of the AT based on their user experience gathered during the testing period (assisting or interacting with older adults). Thus, the relatives’ ratings represent their own perception of the AT rather than the perception of the usefulness of the AT for the older adult.

The questionnaires were paper based and filled out by 1 of the investigators (SM or BK) during the interview sessions before and after the intervention. Older adults and their relative were interviewed in separate sessions to reduce the interview duration and to avoid interaction, except when older adults required support from the relative. Interviews were performed face-to-face with all older adults. Interviews with relatives were also performed by phone.

At baseline, sociodemographic data were collected including sex, age, living situation (alone, together with partner, or with someone else), education (<10 y or ≥10 y), level of care (administratively assigned level of care measuring a person’s care need and determining their claim for additional support), self-perceived health status (excellent, very good, good, fair, or poor), and self-perceived age. Additionally, information on interest in technology was collected (high, medium, or low).

To assess older adults’ ability to perform instrumental activities of daily living (IADL), the Lawton scale was used [26]. The corrected IADL score was calculated taking into account the possibility that the activities have never been performed [27]. To capture additional information about the participants’ social situation, the Lubben Social Network Scale-6 was used [28]. With the Clinical Frailty Scale (CFS), frailty and fitness of the older adults (scale from 1=“very fit” to 9=“terminally ill”) were assessed [29]. In addition, both older adults and relatives had to rate their life satisfaction (scale from 0 to 10, with higher values indicating higher life satisfaction).

For all outcome measures, descriptive statistics were calculated and included the mean (SD). The focus of the analysis was on the outcomes that assessed the usability aspects of the ATs, that is, the TUI and QUEST 2.0. In addition, exploratory statistical inference testing was performed to assess the effects of the ATs on the specified everyday issues of the older adults and their relatives (COPM). The assumption of a normal distribution was tested by the Kolmogorov-Smirnov test of normality. Subsequently, 2-tailed t tests were used to compare intervention groups using the Welch t tests (life satisfaction, TUI, and QUEST 2.0) and preintervention and postintervention data using paired t tests (COPM). Uncorrected P values are presented in the Results section. However, to control the false discovery rate, we used the correction of the P values via the Benjamini-Hochberg procedure [34]. If this correction changed the outcome of the statistical test, this is reported in the Results section. In addition, effect sizes were calculated using Cohen d. The statistical significance was set at P<.05 for all tests. Statistical analyses were performed using RStudio (RStudio Team).

The study was performed from August 2021 to April 2022. A total of 44 older adults were screened for inclusion in the study. From this sample, 18 older adults and their relatives met the inclusion criteria and were enrolled in this study. There was 1 dropout who stopped study participation due to dissatisfaction with the technology. A total of 17 older adults (tablet computer: n=8, 47% and smartwatch: n=9, 53%) and 16 relatives (tablet computer: n=7, 44% and smartwatch: n=9, 56%) completed the study and were included in the final data analysis. In the tablet computer group, there were 2 cases where the same relative belonged to 2 older adults, and there was 1 case where 2 relatives belonged to the same older adult.

The sociodemographic characteristics of the older adults and their relatives are shown in Tables 2 and 3, respectively. Older adults in both groups were aged >80 years on average but reported a younger self-perceived age. Even though participants in the smartwatch group were older on average, they felt younger than those in the tablet computer group. Data on self-perceived health status, level of care, corrected IADL, and frailty indicated that older adults in the smartwatch group were more dependent and slightly frailer than those in the tablet computer group. A total of 4 participants, all in the smartwatch group, perceived their own health to be poor or fair. However, older adults of the smartwatch group were socially more engaged than those of the tablet computer group. While most older adults (tablet computer: 7/8, 88% and smartwatch: 6/9, 67%) and all relatives indicated at least a medium level of technology interest, one-third (3/9, 33%) of participants in the smartwatch group reported a low level of technology interest. Relatives in the smartwatch group had a higher mean age and included more male persons than the relatives in the tablet computer group.

The frequency of use of the tablet computer and smartwatch in everyday life was heterogeneous across the study period. In the tablet computer group, 5 older adults reported varying use frequencies, which ranged between not using the technology at all (0× per week) and regularly using the technology (7× per week). Two participants of the tablet computer group stopped using the technology after the first week of the intervention. In the smartwatch group, 6 older adults used the technology consistently across all 8 intervention weeks (mostly 7× per week). In the smartwatch group, 2 participants stopped using the technology after the first 1 to 3 weeks. A total of 2 older adults (1 in each group) did not document their frequency of use.

The mean life satisfaction of older adults changed nonsignificantly from 7.9 (SD 3.3) to 7.3 (SD 3.1) in the tablet computer group (t₇=1.1, P=.31, d=0.29) and from 7.2 (SD 2.1) to 8.1 (SD 1.9) in the smartwatch group (t₈=1.1, P=.29, d=0.44). Similarly, the mean life satisfaction of relatives changed nonsignificantly from 8.9 (SD 0.9) to 8.0 (SD 2.1) in the tablet computer group (t₆=1.5, P=.17, d=0.53) and from 7.7 (SD 1.4) to 8.1 (SD 0.7) in the smartwatch group (t₆=0.9, P=.41, d=0.39). In the smartwatch group, 2 participants only provided the preintervention (n=1) or postintervention (n=1) value and were not considered in the statistical analysis.

Prior to the intervention, mean ratings for technology acceptance of older adults and their relatives ranged between 5.8 and 6.5 points for fearfulness and between 4.4 and 5.1 points for curiosity in the tablet computer and smartwatch group, respectively. For both items, the differences between the tablet computer and smartwatch groups were not significant (older adults: fearfulness P=.11 and curiosity P=.77; relatives: fearfulness P=.43 and curiosity P=.36). After the intervention, ratings from the older adults and their relatives of the tablet computer and smartwatch groups were similar for the items interest, accessibility, usability, and skepticism (older adults: interest P=.90, accessibility P=.64, usability P=.26, and skepticism P=.67; relatives: interest P=.65, accessibility P=.78, usability P=.64, and skepticism P=.39). This was different for the items usefulness and intention to use. TUI scores on the item usefulness were higher in the smartwatch group (older adults: mean 5.0, SD 1.0; relatives: mean 5.4, SD 1.4) than in the tablet computer group (older adults: mean 3.8, SD 1.5; relatives: mean 4.2, SD 1.6), although this did not reach statistical significance (older adults: t_12.3=1.9, P=.07, d=0.98; relatives: t_12.1=1.6, P=.14, d=0.83). TUI scores on the item intention to use were significantly higher in the smartwatch group (mean 6.1, SD 3.4) than in the tablet computer group (mean 2.7, SD 2.2; t_13.6=2.4, P=.03, d=1.14) in older adults (note that after correction for multiple comparisons, the P value exceeded .05). Although ratings on intention to use from the relatives were also higher in the smartwatch group (mean 7.8, SD 2.7) than in the tablet computer group (mean 5.9, SD 1.5), the differences were not significant (t_11.0=1.7, P=.11, d=0.86). Interestingly, in the tablet computer group, relatives’ ratings on the item intention to use (mean 5.9, SD 1.5) were on average twice as high as those of older adults (mean 2.6, SD 2.2; t_12.3=3.3, P=.006, d=1.66). In contrast, the relatives’ and older adults’ ratings on the item intention to use were similar in the smartwatch group (older adults: mean 6.1, SD 3.4; relatives: mean 7.8, SD 2.7; t_14.8=1.1, P=.28, d=0.54). Descriptive data on all items of the TUI can be found in Multimedia Appendix 2.

The overall satisfaction score of the QUEST 2.0 across all items was similar for the older adults and relatives of both groups (Figure 1). The mean satisfaction score of older adults was 3.2 (SD 0.6) in the tablet computer group and 3.6 (SD 0.4) in the smartwatch group (t_11.4=1.5, P=.16, d=0.75). Similarly, the mean satisfaction across all items rated by the relatives was 3.6 (SD 0.7) in the tablet computer group and 3.5 (SD 0.4) in the smartwatch group (t_8.6=0.3, P=.77, d=0.16).

The 3 most relevant rated items for both AT groups were safety and reliability (n=4), ease of use (n=7), and effectiveness (n=7). Thus, these items were analyzed separately (Table 4). This analysis suggested that older adults of the smartwatch groups rated all 3 items better than older adults of the tablet computer group. Likewise, the relatives of the smartwatch group rated the items safety and reliability and effectiveness better than those of the tablet computer group. This was different for the item ease of use. This item was rated better by the relatives of the tablet computer group than those of the smartwatch group. In addition, item ratings of relatives were better than those of older adults in the tablet computer group. However, none of the abovementioned comparisons yielded significant test results (older adults: safety and reliability P=.06, ease of use P=.57, and effectiveness P=.40; relatives: safety and reliability P=.21, ease of use P=.40, and effectiveness P=.21).

The most frequently reported everyday issues identified in the COPM were contact/communication with relatives (tablet computer group), entertainment/information on news (tablet computer group), perception of safety (smartwatch group), getting help in emergency situations (smartwatch group), and usability of the AT (both groups). The number of specified issues ranged between 1 and 4.

In the tablet computer group, the older adults and their relatives reported diverse scores for self-perceived performance and satisfaction, with both improved and reduced scores. Statistical analysis showed that self-perceived performance (older adults: t₇=0.09, P=.93, d=0.03; relatives: t₆=0.08, P=.94, d=0.03) and satisfaction (older adults: t₇=0.6, P=.54, d=0.23; relatives: t₆=0.5, P=.64, d=0.18) did not significantly change in the tablet computer group during the intervention (Figure 2A-D). This was different in the smartwatch group. All but 1 older adult and 1 relative reported improved performance and satisfaction in the defined issues. Statistical analysis showed that self-perceived performance (t₈=3.5, P=.008, d=1.17) and satisfaction (t₈=3.2, P=.01, d=1.06) significantly increased in the older adults of the smartwatch group (Figure 2E-F). In the relatives of the smartwatch group, self-perceived performance (t₆=2.5, P=.04, d=0.96) and satisfaction (t₆=2.3, P=.06, d=0.87) increased, although this result was not statistically significant (after multiple comparison correction; Figure 2G-H).

This prospective, real-world, pilot study analyzed the perceived benefit and satisfaction with 2 types of AT for older adults and their relatives addressing different user needs (communication related vs safety related). The devices were (1) a tablet computer with a video call function and entertainment content for older adults and (2) a smartwatch with an emergency button and GPS tracking. Both older adults and their relatives in the smartwatch group consistently reported improved outcomes in the domains of life satisfaction and self-perceived performance and positive ratings for technology acceptance and satisfaction. Several participants emphasized feeling more secure when going out while wearing the device. In contrast, experiences with the tablet computer were more diverse, with several participants reporting poorer technology acceptance, satisfaction, and self-perceived performance.

Wearables with sensor-based risk assessment or fall detection and GPS locating can contribute to reducing older adults’ fear of going outside and thus preserve participation in the community, autonomy, and mobility [35-37]. However, issues with device aesthetics, reliability, and ease of use can negatively impact device acceptance [35,38,39]. Most available studies that focus on the accuracy of device measurements are conducted within a laboratory setting. Contrary to our study, the main sensor location in existing studies is the waist or lower back [36,37]. However, it has been shown that a sensor location at the wrist (eg, watch or wristband) is an important feature for better acceptance of the device [40]. Another study found the highest satisfaction for a pendant worn on a key chain or around the neck [35]. Both forms have the advantage of being familiar or comfortable—aspects that are essential for technology adoption as stated by Fischer et al [41]. Although many emergency smartwatches for older adults have entered the market in recent years, only a few intervention studies with devices similar to the smartwatch tested here exist. Thus, the benefit and results for older adults and their caregivers remain inconclusive [38,42]. Future interventional studies in a real-world setting that take into account the wearers’ wishes and focus on reliability, familiarity, and ease of use are needed.

Tablet computers for older adults are designed to improve participation, provide social support, and reduce loneliness or anxiety. Based on the existing research, the effectiveness of these systems remains inconclusive [5,13,43]. In our study, older adults in the tablet computer group reported on average reduced life satisfaction after the 8-week test period and worse scores for safety and reliability when compared to the smartwatch group. Several participants reported technical failures. Some of the issues mentioned included video calls that did not work, difficulties in receiving pictures from relatives, and defective charging. Considering the high value put on device reliability, these problems are part of the explanation for the negative results of the tablet computer. Additionally, the device is operated via a touchscreen, a technology older adults are mostly unfamiliar with. Consistently, the item ease of use was rated higher by the relatives in the tablet computer group, who belong to a generation already familiar with the use of a touchscreen [43].

Participants received no training prior to using the ATs in their daily lives. This procedure was chosen to create a situation that is as realistic as possible. In most cases, older adults purchasing ATs need to rely on the assistance of their relatives or the company support hotline [35]. In both groups, there were 2 participants who stopped using the ATs after the first few weeks of the intervention period, potentially due to a lack of adequate preparation or lack of support. Indeed, previous research indicates that older adults wish to receive additional training of the technology items they use in the home [44]. Thus, it might be necessary to incorporate training, particularly for more innovative ATs with reduced familiarity and potentially challenging features [38].

This study included 3 older adults characterized as vulnerable or prefrail (CFS score=4) and 4 with some level of frailty (CFS score>4). Physical frailty is a “medical syndrome with multiple causes and contributors that is characterized by diminished strength, endurance, and reduced physiologic function that increases an individual’s vulnerability for developing increased dependency and/or death.” (p. 4) [45]. Frail older adults face specific challenges when it comes to using and deriving benefits from ATs [46]. However, this subgroup is frequently underrepresented in intervention studies, thereby impeding a comprehensive understanding of their requirements [47]. In this study, all vulnerable or frail older adults reported a high likelihood for intention to use the ATs, indicating that this population is overall open to using ATs in their daily lives. Additionally, vulnerable or frail older adults had an average technology satisfaction score of 3.5 (SD 0.5), that is, their satisfaction was between “more or less satisfied” and “quite satisfied.”

As clearly deduced from the introductory WHO definition [1], the term AT is a very broad umbrella term, and it covers an extremely heterogeneous group of products. Each product targets users with specific needs and specific goals. Measuring such diversity is challenging within scientific research. The COPM is unique in the sense that it allows the analysis and quantification of individual user needs and specific AT aims. Thus, it can be adapted to different settings, targets, domains, and participants—irrespective of their age, sex, or other attributes. The tool has been used in other studies investigating the effectiveness of ATs [48,49]. The TUI and QUEST 2.0, however, include predefined items or questions that are not developed specifically for older adults. A recent study suggests that independence, affordability, ease of use, and ethics are the most important AT evaluation criteria for older adults [23]. It is possible that these domains are not adequately represented by the TUI and QUEST 2.0. Assessments focusing on older adults or specifically adapted to the needs and wishes of this group might be better suited in future studies.

The 2 ATs studied target different domains (communication vs safety). Thus, the comparability and generalizability of results are limited. However, the used of the COPM allowed an item-specific evaluation according to its properties to help resolve the identified issues. Although the overall interest in the study was high, only 17 older adults and 16 relatives were included in this study. Thus, the statistical power of the study was relatively low, and further studies with larger samples are required to draw robust conclusions. Having the participation of a relative as a prerequisite for enrollment was one of the biggest obstacles for recruitment as they often declined to participate due to time restrictions. However, as many ATs require or support the interaction with relatives, we consider it important to investigate both perspectives, that of the older adults and that of the relatives. A large systematic review on ATs for older adults with dementia analyzed 571 studies and found that most investigations of clinical effectiveness were conducted with small sample sizes of <20 participants [50]. Conducting real-life intervention studies on the effectiveness of ATs requires a significant amount of effort and time. Combined with difficult recruitment among the population aged ≥65 years, this explains the low sample sizes. Unfortunately, we did not find any older adults–caregiver dyads who were willing to test the smart bed rail or the care planning app and able to participate in the study assessments. Both products are possibly of interest for frail older adults with a higher risk for fall or need for care. These individuals might be too sick to participate in a study such as this, live in some form of assisted living facilities or nursing homes, and be more difficult to reach. Older adults with no interest in technology did not participate in the study, causing a certain sample bias. The study was conducted with the device versions (including the software) that were current as of August 2021. The reproduction of the study using newer device versions could potentially change the results.

This prospective, real-world, pilot study confirmed the potential of ATs to support older adults and their relatives, especially for safety-related issues, but also highlighted the remaining obstacles for widespread use. Frail older adults and their relatives, who would potentially benefit the most from ATs, are especially difficult to reach. Thus, future research and technical developments of ATs should take into account the preferences, problems, and goals of older adults. In addition, this study highlights that individualized measures such as the COPM are necessary to identify the needs and assess the user benefits of the ATs in real-world applications.