Older adults’ experience of active computer gaming for falls prevention exercise: A mixed methods study

2.1 ACG system description

The ACG system was developed using Unity 3D software (Unity Technologies SF Inc., San Francisco, USA). It ran on an Alienware PC (Alienware Corps., Miami, USA.) connected to a Microsoft Kinect camera (Microsoft Corps. Redmond, USA) and displayed using a 32” LED screen. Four strength and balance exercises chosen from the Otago Exercise Programme [15] were developed for tracking using the Kinect camera (Supplemental file 1). Users’ body movements were tracked and displayed on the screen in real-time via an avatar. Additionally, each game provided positive or negative feedback indicating the outcome of each action, and a score board provided feedback of results (Supplemental file 2).

2.2 Overview of system development

Development of the ACG system was an iterative process by an interdisciplinary team of game developers and clinicians. It was informed by a systematic review of the literature [9]. Older adults were involved from early in the design phase (early development is reported [10]); and, findings from the first phase of user-testing were used to modify the system and inform study design (Fig. 1). Modifications to the system aimed to improve user experience; for example, modification of ACG system feedback to include audio and visual effects, and the addition of user profiling to record scores and ratings of satisfaction and difficulty after each session. The revised system was then tested in a new sample of older adults.

OPEN IN VIEWER

Fig. 1

User-centred testing. Findings from user-testing of the first prototype of the ACG system with n = 9 older adults (phase 1) were used to make changes to the system. User-testing of the second prototype was conducted with a new sample of n = 7 older adults.

The study design used mixed methods to explore older adults’ user experience with repeated use of the ACG system. Study visits were carried out at two urban day centres for older adults run by a third sector organisation in the UK. This study was approved by the Office for Research Ethics Committees Northern Ireland (IRAS ID: 187902).

2.4 Participants

Participants were recruited from the day centres following group information sessions about the study. Eligible participants were aged 65 years or over, able to walk with or without a walking aid, had stable physical health according to the Physical Activity Readiness Questionnaire [16], and were able to communicate (verbal and written) in English. Participants were excluded if they had a current acute or uncontrolled medical condition that would not tolerate physical activity (determined by their GP who was informed of study participation via letter and asked to indicate any concerns about participation), had significant cognitive impairment according to the Mini-Mental State Examination (MMSE < 21) [17], were unable to follow verbal or written instruction, or were unwilling or unable to consent to participate.

2.5 Procedure

Participants were invited to complete six 1-hour sessions with the ACG system. Sessions were delivered twice per week for three weeks in the centre. While independent use was encouraged, one researcher was present to provide supervision to ensure participant safety. The researcher demonstrated and provided instruction on the use of the system. Participants were encouraged to comment on their use of the technology, report problems and request additional instruction, as necessary. Participants were able to use two chairs placed at either side or their walking aid for hand support, with encouragement to use only the hand support they required. Supplemental file 3 includes photographs of a participant using the system with their walking stick.

2.6 Initial assessment

Demographic information and participant characteristics were collected before use of the system. Standardised tools validated for the study population included: physical function, using the Short Physical Performance Battery (SPPB) [18]; balance, using the Berg Balance Scale (BBS) [19, 20]; fear of falling, using the Falls Efficacy Scale-International (FES-I) [21]; and depression, using the 15-item Geriatric Depression Scale (GDS-15) [22].

2.7 User experience outcomes

Data collected to explore user experience, in terms of usability, acceptability and safety, is summarised in Fig. 2.

OPEN IN VIEWER

Fig. 2

Diagram summarising user experience data collection.

2.8 Data analysis

Statistical analysis was performed using SPSS software (version 23). The data were checked for normality, then appropriate descriptive analyses were used to summarise participant characteristics and outcomes. Associations between participant characteristics and their use of and feedback on the system were explored using Spearman’s correlation.

Interviews were digitally recorded and transcribed verbatim and verified by the researcher. The Technology Acceptance Model (TAM; Fig. 3) [28] was used to develop a coding framework. The TAM is widely used to understand the factors influencing technology adoption. The definitions used are summarised in Supplemental file 4. Data were coded based on participants’ perceptions of the usefulness and ease of use of the ACG system, and external variables that may influence perceptions and attitudes towards and actual behavioural intentions to use the system. Interpretation, synthesis and data reduction were undertaken by three members of the research team. To ensure reliability of the analysis, all anonymised transcripts were coded by two authors (100% SH, 50% IW, 50% KP). The codes were then compared and agreed by consensus through discussion.

OPEN IN VIEWER

Fig. 3

Technology acceptance model [28].

Quantitative and qualitative findings were synthesised concurrently to provide greater understanding of older adults’ experience with the system and the factors influencing their experience. A framework was developed to synthesise the relevant quantitative outcomes and qualitative codes related to the outcomes of interest. The framework was agreed after discussion by three members of the research team. The related qualitative and quantitative outcomes are summarised in Supplemental file 4. The framework was populated with a summary of the quantitative findings, including relevant relationships between quantitative outcomes, and the related qualitative excerpts collected within each code. The mixed methods results were presented based on each key quantitative finding alongside the supporting data collected within the related qualitative code.

3.1 Recruitment

Thirty-eight service users attended one of three information sessions held at the day centres to find out more about participation in the study. Fifteen service users declined screening for the study, due to health reasons, including chronic health conditions, musculoskeletal pain or mobility limitations (n = 8), or lack of interest (n = 7). Eligibility screening was completed for the remaining 23 participants; n = 7 individuals were excluded due to cognitive impairment (n = 5), registered blind (n = 1), or inadequate level of mobility (n = 1). Of the n = 16 eligible participants, n = 7 gave written informed consent and participated in the study. Thus, recruitment rates were 18% of the total number of older adults (7/38) who attended the information sessions, and 44% of those who were screened as eligible for the study (7/16). Seven participants (aged 73–88 years, 3 female/4 male) consented to participate and completed at least one session using the ACG system. Reasons provided for non-participation were health-related or related to scheduling, when individuals wished to attend other activities within the day centre.

3.2 Participants

Participant characteristics, baseline measures and the number of sessions completed are presented in Table 1. Three participants regularly used a walking aid. According to SPPB scores, n = 4 participants had a lower level of physical functioning (SPPB score < 10). BBS scores indicated that one participant was at high risk of falls. According to FES-I scores, 6/7 had moderate to high concern about falling (FES-I score 20–64); however, none of the participants reported having a fall in the last twelve months. According to the GDS-15, n = 2 participants scored on the cut-off for mild depression (GDS-15 score≥5).

The system was used a total of 22 times, representing 52% adherence to the 42 scheduled sessions, with n = 4 participants completing at least half of the scheduled sessions. No participants completed all six sessions. Reasons reported for non-completion of sessions were not related to user experience. Reasons reported included health-related reasons, such as head cold symptoms, a flare-up of osteoarthritic pain and stiffness (n = 12 sessions), or scheduling issues (n = 8), such as participants’ non-attendance at the day centre or preference to attend other activities scheduled at the day centre. Following all but one (21/22) uses of the system, participants agreed they would intend to use the system if given the opportunity (AFRIS item 6); while in qualitative feedback related to behavioural intentions, 5/7 participants did not intend to continue using the system, but two stated they would be interested in future.

Exploring participant characteristics alongside ACG session adherence, all the male participants (n = 4) completed at least half of the scheduled sessions, with n = 3 completing 5/6 sessions. All female participants (n = 3) completed fewer than half the scheduled sessions (1–2). Two of the females (Participants E and F) had scores representing the lowest physical function (SPPB) and balance (BBS), and the highest fear of falling (FES-I) and depression (GDS-15). Qualitative feedback supported this finding and provided greater understanding that the participants experienced changes in their physical health associated with ageing, suggesting this was an external variable influencing their engagement with the system:

Pt A: I could have worked anywhere, no balance trouble at all, but I wouldn’t even try that now because I know that I don’t have the balance.

Pt G: I always did have reasonable health and fitness but over this last, I suppose, this last four or five years things have just deteriorated.

Pt F: The first time I done it I was great, but I wasn’t well last week so that’s why I’m not doing it. I enjoyed doing it, so I did, and turning around and different things, but, not this week. I couldn’t do it this week.

3.4 Participant feedback following system use

SUS scores suggested high levels of usability. The median (interquartile range; IQR) SUS score across the completed sessions was 85 (75.6–92.5), which is considered excellent [24, 29]. AFRIS scores suggested high levels of acceptability and positive attitudes (median (IQR) AFRIS score 36 (35–39) out of 42). Participants most frequently rated moderate to high levels of satisfaction following their use of the system (21/22), with one negative rating following a session with technical difficulties that caused the participant some frustration. Participants reported experiencing low to moderate levels of difficulty with their ACG sessions (19/22). Of note, the highest rating of difficulty on the first session was “moderate” which was rated by the three participants with the lowest baseline of balance, physical function and balance confidence.

Perceived usefulness of the system influenced participants’ perceptions of the game. After every (22/22) system use, participants agreed that using the system would be good for them (AFRIS item 1). Qualitative feedback indicated that participants tended to believe that using the system would have a health benefit or that it would make exercise more enjoyable:

Pt B: I thought the game would actually help achieve a bit of fitness for me, bit of strength in my legs.

Pt F: Well, I thought I would get better walking . . . But exercise does do you good.

Pt E: I thought that it would make doing exercises more interesting. . . I wasn’t doing anything I hadn’t done before. I found it quite good, you know. And it’s less boring than doing a thing and doing it on your own.

Pt B: The game is probably different, I think if it’s there you have a visual and you’re trying to do what that visual is doing, and I think it’s a better motivation to do the thing daily, on a daily basis even.

Overall, participants expressed positive perceptions of the ease of use of the system, which was influenced by the visual feedback provided by the system:

Pt B: If you see the object on the screen, what you need to do, then I think it helps you too. It’s better than actually doing it at home, say trying to walk along a table or that. You can see your result on the tv screen.

Pt E: It probably helps to do the things when you’re sort of watching yourself doing it. I think it’s beneficial that way.

Pt B: A score shows, a good score shows that you are capable of participation of the game and you can maybe try and increase it.

Perceived ease of use was reflected in responses to AFRIS item 4; after 21/22 system uses, participants agreed that if they wanted to, it would be easy for them to use the system. During the semi-structured interviews participants described the game they found most difficult, which was usually influenced by their physical limitations, and aspects of use that were challenging, such as requiring thinking speed. They also made suggestions to improve the ACG system based on their experience:

Pt B: I do think it would be advantageous if you had different levels.

Pt E: If they could slow it down a bit at first and have it gradual, and if it started the first couple of times at a slower pace and then speeded it up to what would be the correct speed, just so you would really get into the way of doing it properly and so on.

3.5 Use of hand support

A summary of the level of hand support required by participants during ACG system use is shown in Fig. 4. Games that involved standing on one leg (leg abduction and one leg stand) required the most hand support. While the level of hand support used by participants for each game varied, it remained consistent for that individual i.e. it did not tend to change with repeated use of the system. Observation of participants’ use of the system indicated that some participants tended to use the hand support available to them (two chairs placed at either side) even if it was more than they required.

OPEN IN VIEWER

Fig. 4

Summary of hand support used by participants at each ACG session.

Qualitative feedback suggested that having a chair available was an important external variable that influenced participants’ perceived ease of use when using the ACG system:

Pt B: If you have the chair. I do need the support, no doubt about it, and I would have to have the chair there all the time. But if I had the chair, I would feel secure enough to do the exercises, there’s no doubt about that.

Pt A: You see, with the least touch [of the chair] I can do it, but without it I can’t.

Pt F: You had the chairs set up to hold on to. I’m no good if I haven’t anything to hold onto.

The median (IQR) frequency of instruction related to set-up in a session was 2 (1–2). The highest frequency of instruction related to set-up was observed on the first use of the system, which consistently decreased on the second use of the system. However, for participants A, B and G, who completed the highest number of ACG sessions, the frequency of instruction related to set-up did not decrease with further uses of the ACG system (Fig. 5A).

OPEN IN VIEWER

Fig. 5

Summary of additional instruction during system uses.

The median (IQR) frequency of additional instruction related to play, for example to correct timing or technique of movement to improve success in the game, was 3 (2–8). The frequency of instruction related to play for participants A, B and G, who completed the greatest number of ACG sessions, decreased with use of the ACG system. For example, participant G required n = 23 additional instructions on the first visit, and on the last visit required n = 5 additional instructions (Fig. 5B).

Interestingly, lower in-system difficulty ratings across all completed sessions were associated with increased frequency of instruction (r = –0.525, p = 0.012). Feedback from participants suggested that the availability of someone there to provide additional instruction or support was an external variable influencing perceived ease of use and enjoyment:

Pt B: You had someone there to help and to try and tell you if you were doing something wrong, or if you should be trying it this way. I think that is a great help to people.

Pt B: I think someone there is a help. I couldn’t see myself enjoying it as much, possibly, on my own as I did when you were here when I was doing it. . . . You had someone.

Pt G: [I felt secure] under your instruction; I have to give you your bit as well, because that was totally new to me.

The importance of social interaction was a recurrent theme throughout the qualitative feedback. As described above, availability of the researcher for instruction and support was important to participants; however, many also valued peer support. Reasons for this varied from having company, to encouragement and an element of competition.

Pt B: That applies to anyone, I mean if you’ve anyone in the room at all. If you’ve two people doing it, one encouraging the other. Encouragement is a great thing.

Pt F: I watched her doing the things and she watched me. . . It’s company, isn’t it. Although you’re company. But somebody nearly the same as yourself, you know. . . I had never spoke to that woman, maybe said hello or something. But that’s like a friend I’ve made. I think that’s good.

Pt B: I think if you’re competing with someone else it is that extra bit of motivation. . . I’d look at myself and say that chap can do it, if I knuckle down here I can do it.

There were no adverse events related to use of the ACG system.