We conducted a 1-week usability test of the ENHANCE app prototype with older adults, including individuals from underserved backgrounds in the United Kingdom, using a mixed methods approach. Recruitment and data collection took place in London between November 2024 and February 2025. An overview of the study procedure is presented in Figure 1.

This project is part of the ENHANCE study, which will evaluate the feasibility, acceptability, and effectiveness of the ENHANCE intervention in reducing dementia risk among healthy older adults from underserved backgrounds.

The Institutional Review Board of University College London (ID 24235/001) gave ethical approval on June 19, 2023.

Before enrollment, all participants received a participant information sheet detailing the study purpose, investigator names, study procedures, and the expected time commitment (60-min introductory session, 7 days of independent app use, 60-min posttest interview, and a 5-min satisfaction survey). Participants were informed of their right to withdraw at any time without consequence and gave written informed consent before any data collection. All research participation was voluntary.

We pseudonymized collected data using alphanumeric codes. The code list and consent forms were stored separately from the anonymized data in locked cabinets (paper records) and on a password-protected, encrypted institutional drive (electronic records), accessible only to the core research team. Data are retained for 1 year following completion of the relevant analysis.

Participants were older adults residing in the United Kingdom. Eligibility criteria were as follows:

We predetermined a sample size of 10 participants, based on guidance suggesting that this number can identify approximately 95% of major usability issues [26]. We did not set eligibility criteria based on education, ethnicity, or area deprivation; instead, we used a recruitment strategy designed to reach underserved groups. Participants were drawn from a database of over 100 individuals (aged 60 years or older) recruited through community settings serving underserved populations in London, including food banks, neighborhood shopping malls, religious groups, and community organizations supporting low-income and ethnic minority communities. No centers serving people with dementia were included. Participants had previously taken part in ENHANCE coproduction activities [23] and had provided consent to be contacted for future research.

To inform purposive sampling, we extracted participants’ age, ethnicity, and Index of Multiple Deprivation (IMD) scores from self-reported data collected in the previous study [23]. The IMD is a UK-based measure of relative deprivation derived from participants’ postcodes and divided into deciles, with lower scores indicating higher levels of deprivation [27]. This information was used to purposefully recruit individuals from more deprived areas and non-White backgrounds.

TKCY, a PhD student with 6 years of dementia research experience, recruited potential participants by phone, email, or text. She informed them that no previous digital skills or home access to a tablet were required. Interested individuals completed a brief telephone screening for eligibility, covering age, dementia diagnosis status, as well as dementia health and lifestyle risk factors (Multimedia Appendix 1).

ENHANCE is a coach-supported, app-based intervention designed to help older adults aged from 60‐80 years, including those from the underserved background, reduce dementia risk by addressing 10 modifiable risk factors—hypertension, diabetes, physical inactivity, obesity, depression, smoking, excessive alcohol use, social isolation, hearing loss, and low education—through lifestyle behavior change and cognitive engagement.

The app’s features fall into 3 categories:

All participants, regardless of their educational attainment, were given access to the cognitive training games in the ENHANCE intervention as they are rewarding in themselves. For the remaining 9 risk factors, each participant selected 1 risk factor module to focus on during the 1-week testing period. Figure 2 illustrates screenshots from the ENHANCE app. Further details about the app design are reported elsewhere [23].

All quantitative data were analyzed using SPSS (version 25; IBM), and qualitative data were analyzed using NVivo 14 (Lumivero).

All audio recordings were transcribed using data regulation-compliant software (Happy Scribe [Happy Scribe Ltd]) and reviewed by TKCY for accuracy and then imported into NVivo 14. TKCY analyzed the data using 6-step reflexive thematic analysis by Braun and Clarke [30,31]. She also maintained a reflexive journal to consider how her background might have influenced the interviews and analysis. She used the COM-B (capability, opportunity, motivation—behavior) model as a guiding framework to identify facilitators and barriers influencing app engagement [32] to provide a comprehensive overview of internal and external factors influencing behavior. It has been widely applied in studies understanding app use and user engagement [32-35] and aligns well with qualitative interview-based data analysis.

She analyzed each qualitative data source separately, generating initial codes and developing them into candidate themes with theme refinement carried out through 5 biweekly meetings between TKCY, A Singh, GL, SGC, and HKP. As consistent patterns emerged, findings from each qualitative data source were integrated into a unified set of themes at the end.

TKCY is a female PhD student in her 30s from Hong Kong with a background in psychology and research experience in aging and dementia research. She is a core member of the team that co-designed the ENHANCE app. Her role in the project may have influenced how she conducted interviews and interpreted data. Her familiarity with app usability literature may also have shaped how she identified themes during coding. As a young, educated researcher proficient in technology, she acknowledged a risk that she may underestimate usability challenges and overlook barriers faced by older adults. To mitigate this, she remained open to participants’ perspectives, actively probed for any concerns raised by participants during interviews, and regularly discussed her reflections with supervisors throughout the process.

TKCY analyzed quantitative and qualitative data separately, then integrated them through collaborative discussions with the research team of A Singh, GL, SGC, and HKP. These discussions involved comparing and contrasting findings across data sources to identify areas of convergence and divergence.

All 10 participants completed the introductory session, the 1-week app testing period, telephone follow-up, posttest interview, and the satisfaction survey.

Participants used the app in testing for a mean duration of 9 (SD 3, range 7‐16) days. The participant with the longest testing period (16 d) had a delay in follow-up due to illness, which postponed the follow-up interview.

Participants’ mean age was 68 (SD 6) years; 7 were female; 4 were White and 6 were non-White (South Asian, East Asian, or Mixed); and 7 had English as a first language. Participants lived in areas spanning a wide range of neighborhood deprivation (mean IMD decile 4, SD 2, range 1‐8). Mean years of full-time education was 14 (SD 3). Refer to Table 1 for full demographic details.

Most had previous experience with technology (smartphone: n=9, computer: n=8, and tablet: n=6); and 5 stated that they were comfortable or highly comfortable using it.

Common dementia risk factors included hypertension and hearing impairment (n=6 each), physical inactivity and depression (n=4 each), and diabetes (n=3). Moreover, 6 participants had 1‐2 risk factors, and 4 participants had 4 or more.

Among participants enrolled in the 1-week testing—5 chose hypertension, 2 hearing impairment, 2 diabetes, and 1 depression.

Some participants with physical or sensory limitations found that certain app features in ENHANCE did not sufficiently support their physical capabilities, creating barriers to engagement. For example, a swipe-based game (Night Sorter) was difficult for a participant with reduced movement due to an arm injury. The speed required to play the game interfered with their ability to participate:

Concerns were also raised regarding low-contrast visuals and time constraints (Multimedia Appendix 5).

Some participants raised concerns about potential effects from certain games, particularly related to visual discomfort. These experiences undermined motivation and, in some cases, led to disengagement from the games. Some participants reported visual discomfort from specific game elements. For example, Stump Shuffle—which involves rapidly moving visuals—was visually demanding:

Similar concerns were raised about Hive Finder due to small, cluttered screen icons (Multimedia Appendix 5).

Participants consistently described initial face-to-face coaching as essential, providing real-time support and reassurance.

Hands-on use during coaching sessions enhanced opportunity to learn by reinforcing understanding. Several participants suggested encouraging users to perform tasks themselves during coaching sessions.

Participants also appreciated ongoing support through phone calls and in-app messaging after the initial session. This helped maintain engagement by providing a sense of personal care and reassurance.

Most participants appreciated the ENHANCE app’s educational videos, which combined expert-delivered information with relatable personal stories. This blend of trusted information and real-life experiences encouraged reflection on health behaviors and motivating change.

Practical, actionable advice from professionals also motivated participants to apply recommendations in their daily lives. One participant using the diabetes module appreciated the clear dietary guidance:

Personal stories created emotional resonance that deepened engagement and relevance. One participant expressed a clear preference for human stories over diagrams:

Furthermore, a participant living with long-term depression found comfort in the simplicity of a story from the “Better Mood” module (Multimedia Appendix 5)

Table 4 summarizes the themes mapped to the COM-B model and their associated design implications, discussed further in the Discussion section.

Our 1-week usability testing suggests that the ENHANCE intervention prototype is acceptable and usable for older adults. All participants met minimum-use targets, with many engaging further, making extensive use of core features (games, videos, and questionnaires). Most also used the in-app messaging feature. All reported high satisfaction with the app and coaching sessions on the postuse survey.

Qualitative interviews identified several facilitators of engagement, such as (1) face-to-face coaching support, (2) appropriately challenging and adaptive games, (3) in-app virtual rewards, (4) visual progress tracking, (5) familiar interface elements, and (6) health information combining expert advice with personal stories. Barriers included unclear interfaces or instructions, design that did not adequately account for users’ physical limitations, and visual discomfort linked to specific aspects of game design.

Overall, our positive usability and acceptability findings are consistent with the Prevention of Dementia Using Mobile Phone Applications (PRODEMOS) study—the only other UK-based, app-based dementia prevention program targeting underserved populations. The PRODEMOS reported that among its UK participants (n=600), 69% found the app acceptable, 71% considered it feasible, and 77% viewed it as appropriate at 3-month follow-up [36].

While both our study and PRODEMOS included features such as SMART goal setting, educational content, cognitive games, and health tracking, ENHANCE offered more intensive human interaction—including face-to-face onboarding, weekly telephone follow-ups, and in-app messaging with a coach. In contrast, PRODEMOS delivered initial sessions remotely due to COVID-19 restrictions, with follow-up support limited to asynchronous in-app chat. Participants in PRODEMOS reported difficulties staying engaged, partly attributing this to delays in coach responses [36,37]—a challenge similarly noted in Healthy Ageing Through Internet Counselling in the Elderly (HATICE) internet-based dementia prevention trial [38,39]. .

Our qualitative findings highlighted the value of in-person coaching. Participants described face-to-face onboarding as “necessary” for real-time support and reassurance—echoing findings from the HATICE study [38]. For ongoing support, telephone check-ins were more acceptable than in-app chat: all participants completed phone calls, while only 7 out of 10 used the chat feature. Nonusers tended to be older, less educated, and less familiar with tablets, and were unaware of the chat function despite onboarding demonstrations. This aligns with evidence that human-led approaches—in-person or voice-based coaching—are preferred by older adults and may foster deeper engagement than text-based or AI-driven approaches, such as chatbots [40,41]. In-app chat also posed greater challenges for those with lower digital literacy consistent with evidence that text-based interfaces can present direct barriers when reading or digital skills are limited [42].

While PRODEMOS assessed usability at 3 months, ENHANCE was evaluated after just 1 week, leaving its long-term engagement and feasibility untested. PRODEMOS reported a decline in task completion from 91% in the first 2 weeks to 41% by month 3 [36], reflecting a common pattern of reduced app engagement over time due to loss of novelty and motivation [43,44].

Using the COM-B framework [32], our qualitative findings can reflect how capability, opportunity, and motivation shaped participants’ engagement with the ENHANCE app.

This study has several strengths. Recruitment via community settings serving underrepresented groups yielded a sample spanning ethnic minority backgrounds and a wide range of neighborhood deprivation, addressing a key gap in usability evidence for inclusive digital prevention [11]. We used multiple data sources—questionnaires, interviews, and system logs—consistent with the Medical Research Council’s guidelines on mixed methods [64] and evidence that combined approaches provide a more comprehensive usability picture than single methods alone [65]. The 1-week testing period captured how user experience evolved across repeated use rather than relying on a single interaction [66], and participants used the app in their natural settings, supporting ecological validity.

Limitations include the 1-week timeframe, which leaves longer-term usability untested, and the small, London-based sample, which limits generalizability to rural areas and other countries. Social desirability bias may have affected interviews, although backend usage data showing use beyond minimum requirements suggests genuine engagement. Despite targeting underserved settings, the sample had relatively high educational attainment; usability barriers may be greater among those with lower literacy. High engagement occurred under supported conditions—device provision, face-to-face onboarding, coach support, and study compensation—which may not reflect real-world implementation. Finally, eligibility relied on not having self-reported dementia without formal clinical assessment. Future studies should incorporate more rigorous eligibility measures.

This study explored how community-recruited older adults, including those underrepresented in dementia prevention trials, engaged with the ENHANCE app over 1 week. The intervention was well-received, with high engagement, positive feedback, and strong usability and acceptability across participants. Key facilitators included face-to-face coaching, familiar interface design, appropriately challenging games, visual progress tracking, and health content that combined expert advice with peer stories. These findings offer practical design insights for developing inclusive digital dementia prevention tools that better accommodate diverse older adults, while future research is needed to evaluate maintained long-term engagement.