This study consists of two parts: one is to design and develop an exergame prototype (part 1) based on the research results [31] of the team’s previous analysis of the needs of older adults with MCI for exergames, and the other is to test the usability of the exergame prototype among older adults with MCI (part 2).

Regarding sample size, Macefield [37] suggests a reasonable baseline range for identifying usability issues is 5‐10 participants, while comparative studies aiming for statistically significant results should include 10‐12 participants. Additionally, usability study guidelines indicate that the optimal sample size is 10±2 participants [38]. Thus, to achieve adequate precision and representation, the sample size for the quantitative phase of this study was determined to be 12 participants. No power analysis was performed as no inferential hypothesis testing was planned. No stopping rules or interim analyses were used.

The primary outcome measure was usability. Immediately postsession, a self-developed questionnaire collected demographic data, while the System Usability Scale (SUS) [39] assessed the exergame’s usability. The SUS is a scientifically validated, reliable scale (eg, ω=0.91 [40]) that is easy to administer [39,41] and frequently used in game research [19,42-45]. It consists of 10 items covering effectiveness (learnability), efficiency (usability), and satisfaction [46], rated on a 5-point Likert scale (1=“strongly disagree,” 5=“strongly agree”) [39,41]. Scores are calculated using a specific algorithm: raw scores for odd-numbered items are subtracted by 1, and for even-numbered items are subtracted from 5. The sum is multiplied by 2.5 to yield a final score (0‐100). Scores ≥70 are considered “acceptable,” with 50‐70 as “marginally acceptable,” and <50 as “unacceptable” [47]. Secondary measures and covariates included demographic data (age, sex, and educational level). All measures collected in this study are included in this report.

Before this study, the first author (XC) briefed participants on the research objectives and prototype functions, provided a demonstration, and obtained consent for audio recording. Participants selected their preferred testing location (home, community center, or care facility) and performed a mandatory 10-minute safety warm-up. To ensure an authentic experience, participants used the prototype independently under XC’s safety supervision (Figure 2). The quantitative data collection occurred immediately after this session.

This quantitative component of this study used a cross-sectional, nonexperimental, single-group observational design.

For quantitative research, valid questionnaires were numbered, and research data were double-entered into Excel (Microsoft Corp) spreadsheets. Quantitative data were analyzed using SPSS (version 26; IBM) software. First, P-P plots were used for normality testing of quantitative data. When quantitative data followed a normal distribution, descriptive statistics were performed using mean and SD; when quantitative data did not follow a normal distribution, median, IQR, minimum, and maximum values were used for descriptive statistics. Frequencies and constituent ratios were used to describe categorical data.

Integration occurred at the interpretation level to identify convergence, divergence, and complementarity between the two datasets. Specifically, a joint display was used to systematically compare the quantitative SUS dimension scores with the corresponding qualitative themes and verbatim quotes. While the quantitative SUS scores established a baseline measurement of general usability, the qualitative findings were integrated to explicitly explain the underlying reasons (the “why”) behind these scores and the specific user pain points, thereby generating comprehensive meta-inferences that extend the understanding of the statistical patterns [52].

This study was approved by the ethics review committee of Nursing and Behavioral Medicine Research, School of Nursing, Central South University (E202397). All participants provided written informed consent before enrollment, acknowledging their right to withdraw at any time without explanation, and were advised that such withdrawal would not affect future services or opportunities. To ensure participant privacy, all collected data were systematically deidentified. Specifically, participants were assigned unique alphanumeric codes upon enrollment, which were used to label all demographic questionnaires, usability scales, and qualitative transcripts. During the verbatim transcription of the qualitative interviews, any potential identifying information mentioned by the participants was removed and replaced with generalized placeholders. Furthermore, all digital files were stored on a password-protected and encrypted storage system accessible exclusively to the core research team. No compensation was offered to the participants. Furthermore, no individual participants or users can be identified in any images included in the manuscript or supplementary material.

While our prior qualitative research [31] identified the specific exergame requirements of older adults with MCI, the current manuscript proposes an original, structured design framework based on those requirements. As a primary contribution of this paper, this original framework was constructed to guide the subsequent development process. See details in Table 1.

The group discussion established the goal of developing a user-centered, immersive, and culturally relevant exergame. Table tennis was selected as the primary sport activity for three reasons: (1) its familiarity among Chinese older adults minimizes the learning curve and cognitive load; (2) its competitive nature effectively motivates physical engagement [53,54]; and (3) racket sports are the sports most conducive to health and longevity [55]. Furthermore, to accommodate the cognitive limitations of older adults with MCI, the design prioritizes simplified tasks over complex narratives to minimize frustration and ensure training continuity.

Finally, the group discussion selected the PICO 4 Pro (Figure 3) based on two critical metrics: clarity and comfort. Regarding clarity, the device offers 4K+ resolution and a high refresh rate, meeting the requirements for high-quality software streaming. Regarding comfort, it excels in three aspects: (1) supporting interpupillary distance adjustment and allowing users to wear prescription glasses; (2) combining a lightweight body (597 g) with a rear-mounted battery design to achieve front-rear balance; and (3) using a breathable face mask to prevent excessive perspiration during exercise.

Based on the above results, the “Preliminary Design Plan for Exergames Targeting Older Adults With MCI” has been formulated, as detailed in Table 2.

The research and development result is a VR exergame prototype titled Lehuo Pingpang, as shown in Figure 4.

Following the convergent mixed methods design, the quantitative SUS scores and qualitative themes were integrated using a joint display (Table 7) to generate comprehensive meta-inferences. Based on established usability evaluation frameworks [46], the 10 SUS items were structured into three core dimensions: learnability (items 4, 5, and 10), usability (items 2, 3, 7, and 8), and user satisfaction (items 1, 6, and 9). These quantitative dimensions and their raw item-level scores were mapped directly against the corresponding qualitative themes, subthemes, and illustrative participant quotes to explore convergence and complementarity.

This study successfully designed, developed, and evaluated an immersive VR exergame prototype tailored to older adults with MCI. Guided by the three-stage MIDE framework, we successfully developed “Lehuo Pingpang,” translating the target population’s requirements into an aging-appropriate and safe VR exergame prototype. Furthermore, our convergent mixed methods usability evaluation confirmed that the exergame prototype is safe and generally acceptable. By integrating our quantitative SUS scores and qualitative interview themes into three core dimensions—usability, learnability, and user satisfaction—we derived comprehensive insights regarding the prototype’s performance. Collectively, these integrated findings demonstrate that while the exergame prototype is highly viable and acceptable, the identified challenges provide essential, user-driven empirical guidance for subsequent iterative optimizations.

The average SUS score in this study was higher than the 70-point threshold required for an acceptable game [47], indicating that the usability of the exergame prototype among older adult individuals with MCI is acceptable. These results are consistent with the findings of Altorfer et al [57] and Shah et al [58], who reported even higher average SUS scores. Altorfer et al [57] explored the feasibility and effectiveness of a cognitive-motor dual-task training exergame in 39 older adult inpatients through a pilot randomized controlled trial, testing its usability with the SUS scale. The results showed an average SUS score of 83.60 (SD 13.72), indicating high usability among older adult inpatients. Shah et al [58] co-designed and developed a VR-based exergame with older adults living in or previously living in rehabilitation centers and physical therapists, using a mixed methods approach to evaluate its usability. Among 14 participants, the quantitative part yielded an SUS score of 83.75 (SD 13.3), confirming acceptable usability. The higher SUS scores in these two studies may be attributed to the fact that participants in this study experienced the exergame prototype fewer times and for a shorter duration, whereas the usability studies mentioned above measured SUS scores after multiple long-term exposures, allowing participants to become more familiar with and accepting of the exergame, thus achieving higher usability levels.

However, this study’s results are inconsistent with those of Thalmann et al [59], who conducted a usability study on their newly developed VITAAL exergame among 13 older adults with mobility impairments, reporting an average SUS score of 58.3 (SD 16.5), lower than this study’s findings. Possible reasons include two aspects: on the one hand, the participants in this study had a younger average age, and research by Bangor et al [60] showed a significant correlation between age and SUS scores, meaning older age may negatively impact SUS scores, which could explain why this study’s results were higher. On the other hand, the study by Thalmann et al [59] included older adults with mobility impairments, while the 12 participants in this study had no mobility limitations. Thus, compared to the sample from the study by Thalmann et al [59], this study’s participants may have been more suited to using exergames.

In the usability dimension, the low perceived complexity and high ease of use demonstrated by the quantitative analysis strongly converged with the “simple and convenient operation” reported by participants in the qualitative evaluation. Participants generally appreciated the system’s intuitive physical layout and convenient operational experience. This success is primarily attributed to the strict adherence to age-appropriate interface design principles during system development. To accommodate age-related degenerative changes, the prototype incorporated appropriate error tolerance rather than overemphasizing precise manipulation [61]; concurrently, interface elements were arranged in a regular and simplified manner to reduce the attentional burden on older adults [61]; and sufficient reaction time was provided to bolster their confidence [62]. Furthermore, to assist older adults with MCI in processing information rapidly, the interface used large fonts and icons while avoiding complex technical jargon [61]; the layout adhered to visual movement rules of left-to-right and top-to-bottom, presenting crucial prompt information in the center to guide the user’s line of sight [62]; meanwhile, the system integrated multichannel feedback modes, such as audio-visual synchronization and dynamic graphics, to satisfy the information readability needs of older adults [63]. This series of design practices aligns with the design principles of Bogza et al [64] for a web-based decision aid tailored to older adults with MCI, which similarly emphasized simplifying design, using clear and easy-to-understand phrasing, and using pictograms as visual reminders to aid user memory. Moreover, qualitative interviews played a critical complementary role to the aforementioned positive ratings. Under the subtheme of “deficiencies in game design,” some participants pointed out that the “confirm” and “exit” buttons were too small. This complementary finding reveals minor shortcomings in the actual execution of centering and sizing guidelines within the interface layout, thereby providing a precise, user-experience-driven direction for subsequent iterative UI designs.

In the learnability dimension, the prototype received moderate ratings, indicating that some participants faced a distinct learning curve. On one hand, the quantitative high functional integration of the system strongly converged with the qualitative appreciation for the practical safety boundary function, with participants stating that the customizable safety zone provided a profound sense of security. This is scientifically supported by the system’s built-in safety boundary setup: the VR headset first displays the real-world scene through pass-through video, and only presents the virtual world scene after the user has autonomously designated a safe range; when the user exits the safety zone, the device issues a prompt and automatically switches back to the real-world scene, ensuring safety during exercise. Furthermore, to prevent overexertion and injury, sufficient time was allocated between high-intensity exercises for older adults to relax and recover [65]. On the other hand, the higher learning costs and moderate need for technical support shown in the quantitative analysis were critically complemented by qualitative interviews—participants generally mentioned their lack of prior VR experience and expressed confusion regarding the complex controller button layout. The study by Tuena et al [32] pointed out that handheld controllers can be complex for some older adults, particularly those with dementia or MCI; this is consistent with the findings of Malinowski et al [66], who revealed that older adults with MCI face greater difficulties in using technology than those without any cognitive impairments. Additionally, Gerling et al [67] emphasized that providing a technology-based training system capable of building technological confidence is of paramount importance, enabling a better gaming experience and more successful training. Multiple prior studies have similarly confirmed the importance of age-appropriate design and robust technical functionality to the usability of exergames [44,68,69]. Regarding learning costs, Thalmann et al [59] also pointed out that a brief period of familiarization and learning is required during the first training session of an exergame. Collectively, these findings highlight the necessity in future clinical applications of providing structured preuse training, allowing sufficient adaptation time, and designing simplified, tactile controllers customized specifically for older populations.

In the user satisfaction dimension, quantitative indicators exhibited high subjective confidence and extremely low perceived system inconsistency, which strongly converged with the qualitative approval of the prototype’s “beautiful and realistic design” and “professional and interesting content.” A plausible explanation is that table tennis is a highly familiar sport and leisure element in the daily lives of most Chinese older adults, which is crucial for enhancing game acceptance [70] and consequently boosting their confidence to participate. The validation of the prototype’s “beautiful and realistic design” stemmed from the aesthetic sensory stimulation provided by its realistic visual graphics and audio effects [71]; the endorsement of its “professional and interesting content” is consistent with the findings of Shah et al [58], which confirmed through semistructured interviews that older adults find VR-based exergames highly engaging and interesting. The “professionalism” of the system is chiefly reflected in its multidisciplinary, collaborative design philosophy: the development process deeply integrated opinions from various experts across hospitals, universities, research institutes, and technology enterprises, covering four key domains—geriatric nursing, clinical medicine, rehabilitation medicine, and AI; moreover, the entire design scheme was fully evidence-based, strictly adhering to the guiding principles in the “Guidelines for Designing Exergames to Enhance Physical and Social Activity of Older Adults” [61].

However, a prominent methodological discrepancy and contradiction emerged between the exceptionally high subjective confidence and the relatively moderate intention to use the system frequently. Qualitative interviews expanded on this paradox, revealing that participants’ long-term willingness to use the game was likely constrained by immediate physical discomforts such as VR motion sickness (dizziness), eye fatigue, facial pressure, and physical fatigue. These feedback reports of physiological discomfort were validated by the systematic review of clinical and research applications of VR in older adults by Tuena et al [32], which also identified usability issues including discomfort and motion sickness. In addition, research by Stanney et al [72] and Sharples et al [73] pointed out that eye strain is one of the most frequently reported symptoms during the use of VR technology. This concern is further supported by similar exergame research; for instance, in a study by Song et al [74] evaluating the usability and feasibility of VR and mixed reality Tai Chi exergames for community-dwelling older adults, although the physiological comfort of the screened sample was acceptable and severity levels were low, they initially excluded 18.6% of participants due to VR discomfort. Prior studies have emphasized that VR cybersickness is a common but frequently underestimated barrier that can lead to user abandonment [75,76], which explains the underlying reason why participants in our study held reservations regarding long-term, frequent use.

From a deeper perspective, this paradox of high scale scores coexisting with physical discomfort can be explained by the “novelty effect” commonly observed in usability testing with older populations. The enjoyment of the experience stemmed largely from the “novelty” brought by the game [70]. The highly immersive nature of VR was an unprecedented experience for most first-time participants with MCI, and the intense freshness and excitement of the initial trial temporarily masked or diluted their physical discomfort when filling out the quantitative assessments. This key contradiction underscores the indispensable value of the convergent mixed methods design used in this study: the quantitative survey established a positive baseline of general acceptability, while the qualitative interviews successfully unmasked the true physical barriers and technical friction, thereby ensuring a balanced and objective evaluation.

In response to the usability issues, an iterative design scheme for the exergame prototype is proposed. First, in terms of game design, the “confirm” and “exit” buttons should be enlarged, and the distance between them shortened; text in the game scene should be prevented from being obscured; and the distance between the table tennis table and the gymnasium walls in the virtual scene should be adjusted to a reasonable range.

Studies have shown that compared with nonimmersive devices such as desktop computers and projectors, the use of head-mounted displays is more likely to cause head pressure, eye fatigue, physical fatigue, and cybersickness symptoms [77]. For standalone devices such as the PICO 4 Pro, these symptoms are often influenced by software optimization. The technical specifications of our prototype’s Unity (Unity Technologies) build target a 90 Hz screen refresh rate, a rendering resolution of 1920×1920 per eye, and an end-to-end latency of 12 ms. To address issues such as strong light inside the glasses, eye strain from prolonged use, discomfort from the VR headset feeling tight on the face, and dizziness, future software optimizations will focus on several critical avenues. First, we will implement Application SpaceWarp to significantly reduce rendering latency and improve performance, allowing the application to render at half the actual screen refresh rate without affecting display quality [78]. Second, we will enable Adaptive Resolution to automatically scale the viewport resolution based on the current GPU workload—decreasing it during high GPU loads and increasing it when resources are sufficient—maximizing GPU resources to enhance image quality without dropping frame rates [79]. Third, adopting the Universal Render Pipeline will optimize PICO’s GPU capability and improve rendering performance specifically for XR [80]. While software optimization remains our primary pathway, for users who still experience persistent discomfort, we will consider using a surface television instead of a head-mounted display in future studies to reduce the aforementioned discomfort while maintaining a sense of immersion.

To tackle the problem of limited game types, we will continue to enrich the variety of games in future research to provide more options for older adults. In addition, older adults with MCI require sufficient time to learn, understand, and adapt to the exergame, as well as adequate preuse training, and sufficient support during use. Finally, regarding hardware-related issues of the game, such as a slightly excessive number of controller buttons, susceptibility to network influences during use, and the lack of reminders or guidance for older adults to adjust the relative height between the controller and the ground, continuous improvement and optimization will be made in subsequent processes.

Based on our design and development experience, several key recommendations are proposed for future exergame designs tailored to older adults with MCI. First, ensuring safety during exercise is paramount, which requires adopting a UCD approach throughout the design and development process in close collaboration with stakeholders such as older users and health care professionals. To maximize physical accessibility, the interface must be simple and easy to understand, comfortably accommodating both standing and sitting postures alongside single- or two-handed gameplay. Second, cognitive load should be minimized by incorporating widely familiar elements, providing high error tolerance, and offering sufficient reaction times during play to prevent user frustration. Furthermore, exergames should prioritize health benefits and focus on individual differences, dynamically balancing user abilities and game difficulty through adaptive difficulty adjustments. Finally, to foster long-term engagement, designers should offer diverse task scenarios, use positive incentive strategies while avoiding negative reinforcement, integrate social support functions to enhance social connections, and deliver multichannel feedback with real-time prompts and responses.

Some limitations of this study need to be acknowledged. First, regarding the design and development phase, this study only designed and developed one game scenario and one exergame. Second, there was a lack of direct end user (patients with MCI) participatory co-design during the intermediate preliminary design validation (phase 3). While this tailored UCD approach was considered because the cognitive decline and potential difficulties in abstract thinking in patients with MCI may make it challenging for them to evaluate noninteractive, low-fidelity UI mock-ups, it deviated from a strict UCD cycle. Relying solely on multidisciplinary experts at this stage meant that certain UI issues, such as inappropriate button sizes and spacing, were only identified after the high-fidelity prototype was fully developed. Nevertheless, these identified UI issues served as empirical, user-driven data that directly informed the concrete, multidimensional iterative design scheme proposed in the Iterative Optimization and Future Design Considerations section for subsequent optimization. Future iterations will explore adapted co-design methods with end users earlier in the process, using simplified but interactive mock-ups to prevent such UI failures.

Regarding the usability evaluation phase, several limitations must be noted. First, the sample size of participants was relatively small, and all participants were from the same region, which may lead to selection bias. Additionally, this study only explored the usability of the exergame prototype from the perspective of older adults with MCI, without investigating this issue from the perspectives of other stakeholders (such as caregivers, medical staff, etc). Notably, although the sample size of 12 participants aligns with usability testing guidelines for identifying usability issues, the high SD in the SUS scores (12.89) indicates significant individual differences among older adults with MCI. Future research should involve a larger and more diverse sample size to better account for this variance and to further validate the system’s usability across a broader user group. Finally, the duration of the usability study is relatively short, which may not be able to fully explore the existing usability issues.

This study successfully designed, developed, and evaluated “Lehuo Pingpang,” a tailored VR table tennis exergame prototype for older adults with MCI. The innovation of this work lies in the formulation of an original, evidence-based design framework that systematically translates the complex cognitive-motor requirements of MCI users into safe, immersive physical-digital interactions. Furthermore, this study contributes to the field of gerontechnology a replicable, multistage development workflow and critical mixed methods methodological insights, unmasking the paradox where high quantitative usability scores can temporarily mask qualitative physical discomfort (eg, VR dizziness and facial pressure) due to novelty effects. In the real world, this prototype provides a highly viable, safe, and motivating dual-task rehabilitation tool. It can be easily deployed across community, institutional, and home-based care settings, ultimately enhancing cognitive-motor capacity, lowering barriers to digital interventions, and promoting active, healthy aging among cognitively vulnerable older populations.