Digital mobile telephony potentially creates new opportunities to augment health care. Owing to their interactive features, large storage capacity, communication capabilities, and ability to access large knowledge databases, smartphones can present a novel means to deliver health care to individuals in the home. Consequently, smartphones are being used to deliver an increasingly wide range of personal health care solutions [1,2]. Although older adults have traditionally adopted new technology at lower rate than other age cohorts, the Pew Internet Research Center reports that the use of Internet technology by older adults is steadily increasing, with 2012 being the first year where more than half of people in the United States aged 65 years and older were using the Internet [3]. Recent studies have shown that older adults have a rich technology profile in terms of home appliances, TVs, PCs, and mobile phone apps and only differ from other technology using age groups in terms of Internet-based technology [4]. Although today’s older adults have better uptake of mobile technologies than previous older adult groups [5], contemporary mobile devices such as smartphones still present a substantial challenge for older adults [6]. These challenges can be a result of numerous factors such as unfamiliarity or fear of technology, lack of perceived usefulness (PU), lack of perceived ease of use (PEoU), diminished interactive capabilities, and poor usability characteristics of the devices in question [7,8].

Methodological steps can be taken during the design process for these devices to ensure that the demands of the device do not exceed the capabilities of the older adult user (ref). For example, steps can be taken to ensure that interface elements such as buttons and text are usable, the device navigation can be designed to ensure that basic tasks only require a small number of steps, and the supporting documentation can be presented in an intuitive and simple manner [9-11]. However, these design aspects may not mitigate a new user’s unfamiliarity with the device, and therefore, the potential for technology rejection may remain quite high [12]. This could have adverse effects when attempting to introduce smartphones to older adult users for the delivery of health care using an mHealth, telemedicine, or connected health infrastructure. In a previous study, a smartphone-based fall detection and prevention system was tested on a group of 39 older adult users over a 10-day period. Despite the system having undergone a full human-centered design process and the participants receiving adequate training on how to use the system, the system scored 70 of 100 on the System Usability Scale (SUS), indicating only average usability [13]. We suspected that unfamiliarity with smartphones and the specific demands of interaction with a smartphone, particularly the unique touch screen interactions required, may have led to poor usability outcomes. From this experience, we concluded that the outcomes of many trials and studies that involve the use of home health app design to run on smartphones could be compromised because of a lack of familiarity with the basic functioning of the device.

A period of pretrial introductory smartphone training in conjunction with concurrent recall-based learning tasks could present a potential solution to this problem. Effective training could provide a complete novice with a better chance of adopting the technology, thereby increasing the potential effectiveness of smartphone-based mHealth and connected health interventions for that person. In a study of a group of older adults who were being introduced to an mHealth pain management smartphone app, 61% of the participants cited “provide training on device use” as the main requirement if their potential use of the technology was to be enhanced and ultimately sustained, followed by 30% who cited “tailoring the device to the user’s functional needs (ie, usability)” as the secondary requirement [14]. Therefore, we can conclude that twice as many participants felt that adequate introductory training was more important than enhancements to the design of the app in terms of usability. With this important finding in mind, this paper will provide an enumerative and detailed methodology to achieve this introductory training as efficiently as possible and therefore potentially mitigate any potential usability problems the technology may have.

In this study, we trained 22 participants to use the same smartphone-based fall detection and prevention system, as described in the study by Stara et al [13], over a period of 5 days. This system is the Wireless Insole for Independent and Safe Elderly Living (WIISEL) system [15]. The system is designed to continuously assess fall risk by measuring various gait and balance parameters and is also designed to detect falls. The system is targeted at older adults who are at high risk of falling [16]. The system consists of a pair of instrumented insoles and a smartphone, which are worn by the user during daily activity. Data collected by embedded sensors in the insoles are sent to the smartphone and then uploaded via an Internet connection to a server in a clinic for processing and analysis. The data are presented in various ways to a specialist via a Web app and desktop-based gait analysis tool. The architecture of the system is illustrated in Figure 1.

The 22 participants were instructed to carry out a number of specific tasks with the WIISEL smartphone app (see the Methods section for details of these tasks). Half of the participants were provided with additional concurrent training in the general use of the smartphone, which began 2 days before the trial and continued. Our hypothesis was that the group that received the additional smartphone training would have a better user experience with the system.

Participants were recruited from the Galway city area. Twenty-two participants were recruited 74±5.5 years providing informed consent under ethical approval provided by University College Hospital Galway. Participants were split into 2 groups as outlined in Table 1. Whether a participant belongs to Group 1 (No extra training) or Group 2 (Extra training) was decided at random with 50% of participants belonging to each group.

All training was carried out in the participant’s home by the lead researcher who followed the same protocol for each participant. The researcher visited the participants’ home each day to teach them new tasks and to observe them carrying out previously learned tasks. A systematic cuing hierarchy approach and a “think aloud” protocol [17] were used by the lead researcher for the training procedure. These will be outlined in greater detail.

To indicate how participants would fare with the introduction of the new technology, their current mobile technology capability was assessed. Participants were split into subgroups based on their previous experience with mobile technology to allow for further analysis. When it comes to classifying users based on their expertise or previous experience with mobile technology, there is no set classification system. Most so-called “expert” users are simply users who have gained hands-on experience of using the technology and may not have used a user manual for their device [18]. However, using an understanding of users’ prior technical knowledge to predict their future adoption of technology is well supported [19]. We classified the participants into 3 separate categories based on their observed performance in carrying of a series of simple functional tasks with their mobile device. The functional tasks were chosen as representing real-world use requirements of the mobile device and were sufficiently challenging to highlight differences in skill level among user groups [18]. The 3 categories are outlined in Table 2.

In this context, “feature phone” is a retronym used to describe low-end mobile phones which are limited in capabilities in contrast to smartphones, that is, they are phones that have basic call and short message service functionality but do not have extensive media or Internet capabilities.

This procedure for practically dividing participants into technology experience groups is illustrated in the flowchart in Multimedia Appendix 1.

PEoU and PU are the 2 key usability indicators pertaining to technology design. The influence of PEoU and PU on behavioral intent, and hence technology adoption, has been supported for the use of technology by older adults and specifically for the use of communication technology [20]. To measure whether there was any correlation between either PU or PEoU and the eventual usability outcomes, we measured the participant’s PEoU and PU of smartphone technology before the trial started. We used a 7-point Likert scale using items from the technology acceptance model [21,22] to establish PU and PEoU.

Our training procedure was based on an approach known as the errorless fading of cues technique [23]. This technique involves reducing the cues on repetitive tasks until the user can complete the task without error. The overall training schedule for Group 1 (No extra training) and Group 2 (Extra training) are outlined in Table 3. Training blocks are broken up to ensure that the participants were not overburdened with new training. Overall, each participant in Group 1 was subject to 0.75 to 1 hour of training or testing time per day, whereas each participant in Group 2 was subject to 1.5 to 2 hours of training or testing time per day. This included regular breaks and the time taken for the researcher to record metrics after each task. The number of days of training was chosen to be long enough to give participants the best chance of achieving some sort of mastery [23] but short enough to allow for convenience in having participants and trainer available for consecutive days.

The WIISEL specific tasks which were to be carried out by both Group 1 (No extra training) and Group 2 (Extra Training) are listed in Table 4. These tasks were selected based on a use case analyses of the WIISEL system [24] and were split into 2 different lesson blocks to reduce the burden on the participant by implementing small measurable objectives [25].

Both groups were exposed to the WIISEL system over a course of 6 days (day 0 was training only, Days 1 to 5 were used to record performance, there was also some training on day 1, see step 2 in the following). This exposure consisted of the following steps:

An example of the user manual for the WIISEL app is shown in Multimedia Appendix 2 and was the product of a comprehensive human-centered design process which tested and informed design changes of the WIISEL smartphone interface and the layout and content of the manual. Both Group 1 and Group 2 used this manual. Group 2 was also provided with smartphone-specific training in parallel with their WIISEL training. The smartphone-specific tasks are listed in Table 5. The smartphone tasks were selected based on recent studies on most popular usage patterns for smartphones [29,30].

This training was started 2 days before the WIISEL exposure began (we will refer to these days as day −2 and day −1). The routine was completed as follows:

The LG-Nexus 5 User Manual outlined in a step-by-step format how to complete basic tasks such as making phone calls and sending text messages (Multimedia Appendix 3). This manual was used for the Group 2 smartphone training.

Data were interrogated using both qualitative and quantitative approaches. To compare how each group performed in terms of the WIISEL tasks (Table 4), t tests were used to seek statistical significance between groups for metrics such as errors made, cues required, completion time, and ASQ scores for each task. To reduce the effects of potential outliers for the task completion time metric (eg, a participant takes an unusually long time because of very slow typing or has to return to the beginning of the task because of a serious error), the logarithmic-based geometric mean was used [31]. Mean SUS scores were also compared for each group using t tests. Ethnographic observations were also made on the types of errors committed with the smartphone by reviewing notes and videos of the lessons.

The results are presented in a series of stages. First, we will present our findings on the technology profiles of the participants. Next, we will compare the individual metrics of task times, errors made, cues required, and ASQ score between the two groups for each WIISEL task. Next, we will compare the SUS score for the two groups from their use of the WIISEL system. Finally, we will compare the results with some other usability studies that have been carried out with the WIISEL system.

The breakdown of participants into the different technology categories and age categories are presented in Table 6. The only significant correlation that was observed indicated that users with greater experience with mobile technology had a higher level of PU of smartphones. There was a minor correlation observed between increased technology experience and PEoU of smartphones. No significant correlation was found between age and PEoU (R=.2, P=.21) or between age and PU (R=−.04, P=.86). Weak positive correlation was found between technology experience and PEoU (R=.39, P=.15); Strong positive correlation was found between increased technology experience and PU (R=.6, P=.005).

Table 6 shows the results of the PE and PEoU questionnaire applied to each participant before the study began. The scale runs from 1 to 7, with 7 indicating a higher level of PU or PEoU. It provides a summary of correlation analysis between these questionnaires and age and technology categories of the participants.

In terms of mobile technology ownership, only 3 of the 22 participants (13%) did not own any sort of phone, whereas 6 (27%) owned a smartphone. The rest of the participants (60%) owned feature mobile phones, with the Nokia feature mobile phone (various models) proving the most popular. Five participants owned tablets although only 2 could be said to be proficient and frequent users. Of the 5 tablet users, none owned smartphones (all owned feature mobile phones). Therefore, the number of participants with some sort of touch screen experience was measured at 11 (50%).

SUS scores showed significant differences (alpha=.05) between groups (Figure 10). In addition to measuring the total SUS, the scale was also split into its subscales to show learnability and usability scores. Group 1 versus Group 2 averages showed scores of 73 versus 87.25 (P=.007), 36.25 versus 63.75 (P=.04) and 82 versus 93 (P=.02) for SUS total, learnability, and usability, respectively.

When comparing the SUS outcomes to previous studies of the WIISEL system, we can see how dramatic the effect of the supplementary smartphone training on the Group 2 participants in this study really was. In Figure 11, we can see that Group 1 exhibited similar SUS results to the participants who took part in a controlled usability test [24] and an open trial [32]. In both of these studies, the participants received no supplementary smartphone training. The Group 2 participants scored significantly higher than the other three groups.

This study aimed to understand how supplementary smartphone training could enhance an older adult’s user experience of a smartphone-based connected health system (WIISEL). The results show that overall Group 2 (Extra Training) had a more positive user experience with the WIISEL system according to the statistically significant difference observed in the SUS (and its subscales) scores between the two groups at the end of the 5-day use period. We observed that in the first 3 days of WIISEL use, Group 2 outperformed Group 1 (No extra training) in key usability categories such as errors committed, cues required, and ASQ scores. No significant difference was observed in task times between each group. It was observed that by the fourth day of use, both groups were recording similar performance metrics, implying that there was a ceiling effect, above which no extra smartphone training could have any significant influence on WIISEL use performance. On the basis of the evidence presented in the Results section, we can conclude that providing the users with extra systematic training on the device had a highly positive effect on user experience.

We observed a number of recurring problems encountered by the older adult users in each group. These problems occurred with more frequency in Group 1, which resulted in poorer usability metric outcomes. We grouped these problems into 3 categories—touch sensitivity, touch quality or accuracy, and user interface feedback. Within the first category, the users encountered problems when they held buttons for too long which would either deactivate the button press or initiate a secondary undesired function of the button. For example, holding a keypad letter for too long would input a number or symbols instead of letters. A related problem was observed when users were scrolling through a menu, such as in the settings page. Heavy touches while scrolling resulted in the user unintentionally entering an option (Multimedia Appendix 4). The user would then quickly become lost as they would not recognize the screen they were now presented with.

Users unfamiliar with touch screens had a tendency to leave their touch finger hovering near the screen when they were not interacting with the screen, causing unintentional and sometimes unnoticed screen presses. Unintentional touches also occurred if the user gripped the phone incorrectly. Problems within the second category, touch quality or accuracy, included inaccurate button striking and poor quality striking. For example, there was tendency to aim too low when attempting to strike a button, usually owing to the angle at which the user held the screen (Multimedia Appendix 5). This, at times, led to excessive tapping, where the user would rapidly tap the screen in the hope of hitting the button correctly, leading them to unintentionally press a button in close proximity or to inadvertently press a button on the next screen. Finally, inadequate or unrecognized feedback on screen caused problems for some users. For example, many touch screen elements do not look like traditional buttons with clearly marked borders. This caused problems with striking accuracy. Sometimes users did not recognize the subtle changes in color or form that indicated a button had been successfully pressed. Other buttons had strange shapes which the user did not recognize (Multimedia Appendix 5), such as a triangle or an arrow for a send button, which led to hesitation and confusion.

The regression analysis performed on the usability metrics and how they affected the overall usability outcome from the SUS show that increased errors made and cues required were related to lower overall SUS scores. Increased ASQ scores (indicating greater task satisfaction) were related to higher SUS scores. Task completion times showed a weak to moderate relationship with SUS scores. This could be explained by the fact that many of the required tasks were not time intensive and had a very linear path toward the desired goal. Participants could not really “get lost” within a task or deviate too much from the optimum task path. Most participants who took longer to complete tasks did so because they were simply progressing at that pace. Therefore, task completion time may be a good indicator of overall usability and may depend on the context of the interaction and tasks involved.

With regard to the relationship between technology adoption predictors such as age, technology experience, and PU/PEoU on usability outcomes, we found no strong link for age and technology experience category although there was a moderate link for PU and PEoU.

Our methodology meant that the smartphone-specific training was maintained in parallel with the first 3 days of WIISEL training for Group 2 (see Table 3). There is a concern that this concurrent exposure, rather than implementing a cut-off where the smartphone-specific training ceased before Group 2 began using the WIISEL app, affected the outcome of the WIISEL usability data for Group 2 owing to the group having more total smartphone exposure time during the WIISEL app exposure. However, this approach was chosen such that the participant could achieve the benefit of the training they received on days −2 and −1, by recalling what they had learned from day 0 onward and receiving the appropriate guidance to optimize the training [33]. It is in this context that the participant achieves the benefit of the extra training, thereby improving their user experience and increasing their PEoU with the WIISEL system. We understand that for this training approach to be further operationalized, it may need to be streamlined or indeed undergo some structural changes. However, given the context of the study, we felt that our approach worked for the groups in question and allowed us to properly observe and measure their progress with the smartphone and the WIISEL app.

Regarding the specific problems we observed, we can make some instructional recommendations based on our training experience. These are presented in Textbox 1 and are presented within the categories of problems we identified in section Observed Problems.

Further to the specific recommendations for training older adults outlined in Textbox 1, we can make some more general recommendations. First, when approaching the older adult population, it is important to consider any cultural resistance and concerns that could act as hindrance for the uptake of technology-enabled health care. In particular, security, intrusiveness, lack of control, confidentiality, and usability issues can lead to a lack of trust in such technologies. We can assume that some cases of technology rejection are because of a lack of proper, scaffolded (scaffolded learning of technology is the use of progressive steps to allow the learner to gain independence in their use of the technology), and systematic introduction to the new technology [34,33]. This planned exposure, which should come in the form of appropriate training, must achieve the desired positive impact within a short window in order for technology acceptance rather than technology rejection to occur. The training must not only overcome apprehensions about how difficult the technology is to use but must also overcome any misconceptions about how useful the technology is. Older adults, as with many other user groups, overwhelmingly reject technology when there is unclear evidence of personal benefit or improvement in quality of life [35-37]. A key strategy of the training is to build and increase the user’s perception of and trust in the technology. Although instructional activities can take the form of written materials, computer-based programs, or face-to-face communications, our experience in this study shows the benefit of short, yet intensive, periods of task-based learning with direct corrective feedback. This approach may have applications in other domains such as mobile learning in education [38].

In this paper, we have discussed significant findings from a usability design study where a training intervention was developed and tested to introduce elderly users with limited or novice technology experience to a smartphone-based connected health care system. Our findings show the importance of properly introducing—in a scaffolded and systematic fashion—older adults to technology to improve their technology acceptance and enhance their user experience. Through our research in this specific context, we feel that it is possible to build or increase trust in connected, mobile, and wearable health technology through the design and deployment of meaningful instructional activities. Although there currently does not exist a structured training methodology for mHealth with older users, the methods and lessons learned in this paper can be used to conceptualize, design, and implement appropriate, bespoke training strategies. We understand that effort and time will not always be available to carry out training on a prolonged basis; however, scaffolded and structured intervention is necessary to ensure successful adoption of useful mHealth technology by elderly users. In addition to informing our future work, we hope the development of the WIISEL system and the guidelines and geragogical (geragogy is a theory which argues that older adults are sufficiently different that they warrant a separate educational theory) activities enumerated here will be widely used for those designing and developing connected health devices and infrastructure for older adults. More importantly, the involvement of end users is a key strategy for recognizing and removing barriers and mitigating design limitations, but our research has shown that this must be carefully planned to influence, drive, and refine systematically the iterative development of connected, mobile, and wearable health care technologies. We think this paper provides a useful enumerative approach to plan and conduct usability evaluations of smartphone apps and to gather user experience validation data, particularly in the domain of education and learning.