Keywords

1 Introduction

There is an increasing availability of welfare technology designed to assist elderly people. Elderly users of assistive technology constitute a heterogeneous group of users with varying degrees of capacity and function. The design of new technology does not always take into consideration this diverse user group’s needs and varying degree of function. Our research targets the needs of elderly users who struggle with using, or are unable to use, existing touch-based interfaces currently available in municipal care homes in Norway. We present a tangible prototype named GLiMT that aims to support existing competences and habits by porting selected functionality from a traditional screen-based solution to a physical solution. GLiMT attempts to make information more accessible to users struggling with touchscreen interfaces by allowing them to interact trough a physical interface. We included a total of 42 elderly participants (aged 67 to 92 years old). The prototype serves as a demonstrative artifact and allows us to compare current touchscreen-based interaction with a tangible alternative. We draw on the results of our interviews and usability tests to discuss the potential role of tangible interaction in the context of assistive technology. We use the case of GLiMT to discuss how tangible interfaces can help users experience assistive technology as available and familiar, and thereby help to lower the threshold for use and make information more accessible.

The paper is organized as follows. Section 2 provides a brief overview of related work on tangible user interfaces and frameworks in the context of elderly people. We then outline our empirical context and research methods of our study in Sects. 3 and 4, before we present the results in Sect. 5. In Sect. 6, we present a discussion of the implications of our study, and the potential of opportunities that arises when moving from screen-based to tangible interaction.

2 Related Work

One of the three dimensions of aging defined by Spreicer [1] is biological age. Reduction of bodily capacities, such as loss of fine motor movements, limits our ability to interact with certain interfaces. Prior research has demonstrated that screen-based interaction introduces new of challenges due to bodily changes [2, 3]. Spreicer [1] further argues that cognitive aging changes our cognition and information processing abilities, and thereby makes it difficult to adapt to new problems without extensive training.

There has been some relevant research on tangible interaction as a way to overcome the challenges of traditional screen-based interfaces. Gamberini et al. [4] introduces Eldergames, an interactive physical apparatus that provides social cognitive training. Cognitive rehabilitation has also been the goal of the work of de la Guía, Lozano and Penichet [5] who used a NFC-based solution to accompany touch-screens in a game-based setting. Iversen and Joshi [6] introduces various tangible and embodied prototypes for helping elderly deal with declining bodily and cognitive skills. The work of Häikiö et al. [7] presents a solution similar to ours as it relies on NFC-based interfaces in order to help elderly people solve daily tasks such as ordering meals. Findings from their paper suggest that decline in motor abilities does not necessarily prevent or complicate the use of the system.

Cho et al. [8] present a framework for tangible user interfaces tailored for elderly users. The framework introduces two dimensions; tangible interface properties and supportive interface properties. We draw on four of these categories in the discussion to understand the tangible properties of our prototype. More precisely, from tangible interface properties, we include manipulation and intuitiveness; from supportive interface properties, we have elicited accessibility and simplicity as the main implication relevant for our research. Simplicity is described as easy to understand in this framework, however designing for simplicity in the context of elderly people and assistive technology is not a simple matter, and our prior research has demonstrated how screen-based interaction, in particular, may become too challenging for the users [3]. Thus, our understanding of simplicity introduces a more tailored understanding of the term; simplicity is to be understood only through the experience of the user, rather than something context-detached, and it should build on mastery and context to support users.

3 Empirical Context

Our study is part of a larger research project focusing on assistive technology in cooperation with Oslo Municipality. Two local care homes served as our primary locations for understanding the problem area and gather data for user requirements. In addition to these two local care homes, we used a third activity center to recruit participants for our control group. The care home offers residency to elderly users with cognitive and physical impairments. As a part of the Municipality’s welfare program, each apartment comes preinstalled with a set of assistive technologies. One of these assistive technologies is a customized tablet installed in the apartment [3]. The tablet is portrayed in Fig. 1 (left) and served as the touch-based device which we used as a basis when porting functionality to our proposed tangible prototype.

Fig. 1.
figure 1

The current touch-based tablet (left) and the GLiMT prototype (right).

Full size image

3.1 GLiMT

The GLiMT prototype, illustrated in Fig. 1 (right), is a small wooden cube that receives a stream of four categories of online information through a wireless configuration. The information is identical to the information offered through the existing touchscreen-based tablet. The cube has four sides with corresponding symbols to represent the four categories of information. Incoming messages are signaled by having the corresponding side of the cube light up. Messages can be displayed on any bigger screen that the residents are comfortable with, e.g., tablet, cell phone or TV, by placing the lit side of the cube next to the back of the device or remote control (TV). The topside displays a small screen where a thumbnail of the sender of the message is shown inside a frame with the corresponding color. The bottom side has an inductive charger unit to charge the cube whenever it’s placed on its base. It is designed not to be confined to each apartment, but to be carried around in the whole local care home. A battery icon indicates charging, and this icon is repeated on the charging station. While charging, a battery icon symbolizes the charging status on the small screen on the top of the cube.

4 Research Methods

The research was conducted over three months during the autumn of 2015. The main goal of our mixed-method approach was to gain insight into challenges with current touchscreen-based interfaces and to compare the device to our tangible alternative. Table 1 presents an overview of all participants included in our study. We conducted six individual interviews, and two sessions of group interviews with a combined total of 10 participants. We carried out formative testing on the GLiMT prototype and held a summative test to compare real-task differences between the tangible prototype and the existing touchscreen-based interface.

Table 1. Overview of research methods and participants.
Full size table

The participants involved in this study were residents at the local care homes. We also recruited participants for the usability test from an independent activity center for a control group to reduce the likeliness of contextual bias. Most of the visitors at the activity center resided in private homes and came to the activity center during daytime to participate in scheduled activities such as dinner and board games.

The total number of participants was 42 people, with 25 in the experiment group and 17 in the control group. The age ranged from 67 to 92 years (M = 80). We recruited our participants through the method of convenience sampling by visiting both care facilities and joining activities in the common rooms. We aimed to achieve a balanced and representative distribution of age and gender within the target population, but the gender distribution was skewed towards women (61 %). While we did recruit several users with physical impairments, there were also instances where people declined to participate because of physical impairments.

4.1 Interviews and Group Interviews

We used semi-structured interviews and group interviews to capture the daily life in the common rooms of the care home to understand the problem space and establish an understanding of current challenges. The interviews were conducted in the common rooms of the care home over the course of three weeks. We spent the time visiting the care home on five separate occasions to take part in daily activities like dinner in the cafeteria, afternoon coffee, and reading group in the library of the care home. We asked questions about what technological devices they used regularly, as well as related questions regarding social aspects, daily activities, routines, and communication with relatives. We also complemented the interviews with observations of their daily life in these common rooms and discussed their current assistive technological devices, like digital pill organizers, safety alarms, and the customized tablet. Figure 2 portrays some of the interview and observations activities conducted.

Fig. 2.
figure 2

Examples of conversations, interviews and walkthroughs with participants.

Full size image

To more accurately present the results of the interviews and group interviews, as well as provide a more detailed analysis, we end this section by providing an overview of all participants who contributed to these activities in Table 2. Three participants (#5, #10, and #16) chose not to disclose their age.

Table 2. Overview of participants in interviews and group interviews.
Full size table

4.2 Usability Tests

We conducted two rounds of usability testing, one formative and one summative. All tests were conducted in the common rooms of the two institutions during times where there were organized activities such as dinner and afternoon coffee. The formative usability testing was conducted on an earlier version of the GLiMT prototype in order to get feedback on performance. More precisely, we registered the number of attempts and error-rate as participants, who struggled with the existing touchscreen tablet or were unable to use it, carried out a set of 10 daily tasks that encompassed all major aspects of the interaction. 19 users participated in this formative evaluation, 11 in the experiment group and 8 in the control group.

In the summative evaluation, we compared performance when using GLiMT vs. touchscreen-based devices. This usability test required participants to be capable of using any touchscreen-based device (e.g., provided a tablet or own mobile phone), and being familiar with common tasks and main interaction mechanisms. Thus, the number of participants was lower during this test. Seven users participated in the summative test, 3 in the experiment group and 4 in the control group. The participants were asked to perform 11 daily tasks during the evaluation.

The tasks used during the two tests provided insight into several aspect of the design such as (1) understandability of the concept of interaction, signals and messages; (2) perception of representation, typography and icons; (3) volume and quality of auditory feedback; and (4) difficulty with operational tasks (handling, manipulating, and charging). Participants were encouraged to think out loud while performing the tasks in order to capture any additional opinions. Examples of tasks included reading daily activities, identify the sender of a picture, and charging the device. The participants were given instructions and demonstrations of all tasks included before initiating the test, and the task order in both rounds was randomized to negate any learning effects. Figure 3 shows two of the elderly residents that participated in the summative evaluation, and how they used different touchscreen-based interfaces based on preference.

Fig. 3.
figure 3

Elderly users participating in the summative usability test.

Full size image

5 Results

5.1 Interviews and Group Interviews

We performed a thematic analysis of the gathered data from both the individual and group interviews. We categorized the data into four thematic areas by using an inductive approach [9] with two levels of clusters for each identified thematic area. Table 3 presents an overview of the four thematic areas that emerged from the data collected during the interviews and group interviews.

Table 3. Overview of elicited thematic areas.
Full size table

In our interviews, the customized tablet was repeatedly brought up as a topic of interests to the participants. 4 out of 6 participants from the individual interviews reported difficulties with the touch-based interaction. For certain residents, the touch-based interaction felt unfamiliar: “It is strange to use your hand like this [demonstrating with a sweeping motion with the hand], so I don’t really care much for it, haven’t really used at all” (Participant # 3). Connectivity problems led to the content of menu and activity calendars not being updated on a daily basis: “It has been [the same dish] on the menu for a week, one has to go down [to the common rooms] when the tablet does not work” (Participant #7). Two of the users reported that they preferred to go to the common room to read the poster on the wall because they found it easier than using the tablet. The manager of the care home reported that the tablet was of “limited use to the residents” because of these usability problems. The charging unit for the tablet was mounted on a wall in the kitchen of the apartments. This led to the tablet not always being available to the users who depended on assistance to get out of bed in the morning, or used wheelchairs, walkers or other aids.

Regarding the participants use of technology, in general, we found that all of the participants owned and used a TV regularly. The use of cell phones and tablets was common but to a varying degree. Some participants also had a personal computer that they used. Participation in social media and the internet varied considerably, from daily use to no use. Also, all residents had assistive technology such as automatic control of lighting and heating in their apartments. This increasing number of assistive technological devices was also a topic of interest for the residents. This was brought up by participants both in individual interviews and in the two group interviews; the participants expressed a feeling of being overwhelmed by the number of technology devices and concern regarding the implications for the quality of intrapersonal relationships between caretakers and the elderly people in need of assistance. 9 of the 16 participants in the individual and group interviews expressed a feeling of being overwhelmed by the increasing number of assistive technology devices in their daily lives, e.g., “It can be too much of a good thing” (Participant #6), and “Technology is fine, but many are not using it at all” (Participant #9).

Three participants reported that they thought it was difficult to follow the technological development in their personal use of technology: “I will attend a course next fall. It will be good to learn to use it [the tablet]. It’s easy for those who know it” (Participant #2). During both group interviews, there were participants who were uncomfortable with giving an opinion on the use of technology because they did not feel that they knew enough about it. Also, not all the participants who reported feeling overwhelmed by the amount of assistive technology had difficulties using the technology. Two of the participants actively used smartphones, computers and the tablet, but expressed a general interest in the development of the caretaking technologies. A third participant had a professional background where she had worked with computer systems since the early eighties and were interested in following new developments as well as their impact on the community.

5.2 Usability Tests

During the formative test, we registered error rates during task-solving for the 12 participants with the GLiMT prototype. Every participant in both the experiment group and control group was able to grip fully and manipulate the object, and there were no apparent challenges with handling the cube for any participant. 6 out of 12 participants struggled with at least one of the screen-specific tasks (e.g., reading a message from the screen). The reasons varied from small letters and weak contrasts to navigation and interaction challenges.

As Fig. 4 demonstrates, the average number of errors per task peaked at 1.428 (for task #1 and #10), yet remained below 1.0 as average errors per task for the 12 participants. Out of the 120 tasks solved by the 12 participants, 19 errors were registered. One participant in particular (male, 81 years) made 5 of the total 19 errors. Thus, most participants made it through all tasks without making a single error. Only one task was solved by everyone without making errors, but we did not register any significant reasoning for this one error-free task.

Fig. 4.
figure 4

Average number of errors

Full size image

During the summative test, we ran comparative tests on the GLiMT prototype vs. the touch-screen device. Figure 5 presents an overview of the time spent during task solving for the 7 participants. The red-colored line illustrates the number of seconds spent per user with their currently preferred touch-screen device, which in this case was their mobile phones; the blue-colored line illustrates the similar performance time with GLiMT. An independent-sample t-test was conducted to compare the performance time, and the results from the t-test are reported in Table 4. While only three of the participants were above the average for the touch-screen device, we still registered a significant difference in time spent when using the touch-screen device vs. GLiMT (p = .028). The average time spent per user was significantly lower for users of GLiMT (M = 8.428, SD = 4.198) compared to the touch-screen devices (M = 23.14, SD = 16.608), and no single participant performed faster with their existing device than with the GLiMT prototype despite this being a novel and previously unexperienced interface for most participants.

Fig. 5.
figure 5

Time spent on task solving for GLiMT and touch-screen device

Full size image
Table 4. t-test comparing performance on GLiMT vs. touch-screen device
Full size table

6 Discussion

GLiMT has served as an interesting prototype in the exploration of how tangible interfaces can open up new interaction opportunities in the context of assistive technology. For elderly people who are unable to use touchscreen-based interfaces for various reasons, our demonstrative prototype has provided empirical evidence of how porting functionality over to new interfaces can reconnect users with important assistive technologies. Furthermore, it demonstrated how using familiar movements and actions as input rather than screen-based commands enabled participants to operate new interfaces with a very low error rate (as illustrated in Fig. 4). Moving beyond the particular functionality of our prototype, this study has presented some new opportunities for interaction through tangible interfaces. In this section, we discuss aspects of interaction with assistive technology supported by our empirical results.

Considering the category of accessibility from the framework of Cho [8], the interviews revealed a strong general opinion of touch-screen devices being inaccessible due to high thresholds of use; the interface was too difficult to use and required too much effort to understand for most people to get started. Increased complexity raises the threshold for interaction, something that may scare away elderly users [10]. Participants also stated that the touch-based interface felt unfamiliar to them, contributing to their feeling of being overwhelmed. These results confirmed similar trends in our previous research [3], where a similar interface was perceived as complex. One of the significant effects of porting functionality over to a tangible interface in this study was the effect on how the users perceived technology and digital information. The tangible interface was perceived as more familiar and thereby more accessible to most users. Both the interviews (Table 3) and results (Fig. 5) demonstrated fewer errors and faster completion time for daily tasks when using GLiMT. By porting selected functionality to a tangible user interface, we aimed to simplify the interaction and build on the users’ existing experiences of bodily interaction and manipulation of physical objects. With our extended understanding of simplicity [3], mastery is an essential success factor for achieving interaction with low thresholds and in turn intuitiveness [8]. Several participants claimed that they only needed to deal with touchscreen-based interfaces after moving into the local care home, and they did not have any prior experience or particular desire to learn new interfaces. The threshold for mastery required new incorporation of interaction mechanics (such as swipe, drag, and pinch), and relied on cognitive capacities.

Our tangible prototype aimed at relying less on heavy cognitive abilities by reducing the number of steps in the interaction to a minimum (as demonstrated in Table 4), as well as building on familiar gestures and movements for interactions that had been a part of their prior experience. Past experiences, along with cognitive capabilities, have been demonstrated to have an influence on task performance [11]. Reduction in memory load is a significant advantage of the proposed design in [5]. Building new technologies on familiar metaphors and experiences was also the strategy of [10, 12] when designing for elderly people. Another aspect of mastery is allowing elderly people to engage in interactions that respect their bodily capacities. With bodily changes, a lot of touch-based interfaces became inaccessible due to physical or psychomotor disabilities or reduction in capacity. This point was also made by [2]. One particular participant stated that physical capacities complicated the interaction: « I’m not very good at that, I can’t make it work, it is difficult to start [the interaction], to press » (Participant #2). GLiMT was designed to offer more flexibility with regards to the bodily configuration in order to acknowledge the importance of manipulation [5, 8] also argues for more flexibility. By designing the interface as a light-weight cube, we opened up new interactions mechanisms that were unavailable to the users when interacting with the tablet. For instance, people with only one good hand were in most cases unable to use the tablet in any other configuration that on a table, while they were fully able to operate GLiMT. The four leveled sides provided good stability, and its non-fragile components did not break like glass if the user were to drop it.

One challenge with the design of our tangible prototype was the consistency which affected the intuitiveness. Intuitiveness has been an important part of the design of similar studies, e.g., [4]. Several participants in the usability tests expected the device to charge by putting the charging icon on the screen down towards the charging icon on the base station. This indicated a perceived break in intuitiveness; they did not notice the battery icon on the bottom of the device at first. Similarly, we found that not all elderly users were familiar with commonly used iconography representing functions in technological devices, such as the battery icon. Some participants did not recognize the icon and did not understand the interface metaphors: “It reminds me of a remote control or something, maybe a temperature regulator. Maybe if the tip was more articulated… it wouldn’t take much” [elaborating on what the icon represented to him and how it could be altered to remind him more of a battery]. Again, this demonstrates how building on familiar metaphors and gestures is an important implication as there might be breaches in the cultural knowledge of representations [2], thereby violating the intuitiveness. Furthermore, by involving the users in the design process and explore what different representations means to these users, one can tailor representations that build on their existing knowledge [13]. Not designing the cube of our tangible prototype after the concept of tailoring became an evident limitation of the design. While our prototype outperformed the existing touchscreen-based interface, we do not claim this prototype to be a definite piece of technology; instead, it serves as a demonstrative prototype that illustrates the potential of utilizing the key concept of tangible interaction, for instance, tailored representation.

We also registered how the tangible interface incorporated better with the contexts of use, most notably how the elderly users perceived the technology as a part of the home. Familiarizing the technology helped us re-establish interaction, yet we still saw how it also needed to align with the surroundings of the homes of each participant. One participant said that the design of the existing technology should be “black or light so it could go with anything. Then it would be more appropriate in my home”. However, most elderly people in our empirical context had a different aesthetic in their homes and did not feel that the black, chrome, edgy and flashy screen-based devices blended well with the context. The wall-mounted tablet stood out in most wooden-furnished homes as an exterior addition to the household. Thus, we aimed at developing a wooden cube that would embed well in the households and reduce the chance of disrupting daily activities. This idea of embedding technology into the everyday environment is also advocated by [7]. This was also the reason for extending the functionality of GLiMT to televisions which were a piece of technology all participants in our study had and used regularly. Building on top of existing technology already incorporated into daily routines further increased the accessibility and did not introduce any major changes to their spatial setup. This also increased the freedom to refrain from adapting to new undesired technology and instead continue relying on familiar technologies such as TV.

7 Conclusion

This paper has used GLiMT as a demonstration of how tangible interfaces building on habits and prior experiences can help elderly users of assistive technology reconnect with lost interactions. The aesthetic nature of the tangible interface allowed the design to immerse better with the contexts of use, as well as to signal a more appealing and accessible interface. We have applied a thematic analysis as well as traditional usability tests to compare the performance of the tangible alternative with existing solutions, and demonstrated the potential of porting screen-based functionality to tangible interfaces.