The Usability of a Touchpad Active Video Game Controller for Individuals With Impaired Mobility: Observational Study

TPS Development

The TPS is an internally developed AVG controller designed to add physical activity to sedentary video games. This prototype translates physical contact into video game commands. Because the TPS is recognized as a USB controller, this system can provide commands to any video game title that features controller support. The current proof-of-concept has been adapted for use by people with impaired mobility.

We constructed 6 touchpads using a particle board base, conductive aluminum tape, electrical wire, and duct tape. The touchpads were wired to a MAKEY-MAKEY circuit board (MAKEY-MAKEY LLC, Santa Cruz, CA) that was connected to the video game computer as a controller. All but 1 touchpad measured 20 cm × 20 cm, and the final touchpad was larger (61 cm × 31 cm). Each touchpad was placed on an adjustable stand that enabled varying height, distance, and orientation (vertical to horizontal). The smaller touchpads were placed in front or to the side of the player. The larger touchpad was placed behind the player.

To ensure safety, adjustable parallel bars were located on both sides of the player during gameplay. The participants played video games approximately 1.8 m away from a flat-screen television. Touchpads were placed so they did not block the player’s view of the screen. All wires were secured away from the player to prevent any trip hazard during gameplay.

Design and Setting

This was a cross-sectional usability testing study. We assessed usability and embedded participant feedback to help explain our results. All written consent and data collection took place at the University of Alabama at Birmingham within the RERC RecTech Exercise Science and Technology Laboratory. Participants attended a single 60- to 90-minute session to test the usability of the system.

Ethics Approval

The procedures of this study were approved by the University of Alabama at Birmingham’s institutional review board (IRB 300003265).

Participant Recruitment

We recruited participants from the local community of Birmingham, AL, using flyers and word of mouth. Sample size estimates were based on identifying common usability barriers for the system, and issues specific to the 3 modes of play (standing, chair sitting, and wheelchair sitting). According to Cazañas et al [35], a sample of 17 individuals would reasonably identify 80% of common problems in the system, and groups of 4 to 9 are sufficient to identify problems specific to the mode of play. In total, 21 participants were recruited to account for modest attrition.

Interested individuals were included if they were an adult 18 to 75 years of age, had a self-reported mobility impairment, and possessed the ability to exercise with their upper extremities. Individuals were excluded if they were unable to converse in English, weighed greater than 181.4 kg (400 lbs), had significant visual impairment that prevented them from seeing a large flat-screen television, had cardiovascular disease within the previous 6 months, had severe pulmonary disease or renal failure, currently pregnant, ongoing exacerbation of a health condition, or any other condition that would interfere with testing procedures. Participants received a US $50 gift card at the end of their visit.

Measures

Participant usability of the TPS was assessed using the System Usability Scale (SUS) and the Health-Information Technology Usability Evaluation Scale (Health-ITUES). The SUS is comprised of 10 statements, and participants rate their agreement of each statement with a 5-point Likert scale from “strongly disagree” to “strongly agree.” Across 206 usability tests and 2324 responses, the SUS was found to be a robust and reliable (α=.91) tool to measure usability [36,37]. The SUS is robust to small sample sizes and applicable to many systems [36,38,39]. The SUS can produce a score from 0 to 100 and a score equal to or greater than 68 indicates above-average usability [37,39]. In addition to the single score, factor analyses suggest a 2-factor structure of learnability and usability [40]. The SUS scores were summed and converted from a 0 to 40 into a 0 to 100 scale, and a score of 68 was the threshold to indicate the TPS is “usable” [36].

Similar to the SUS, the Health-ITUES is made up of 20 statements that participants rate on a 5-point Likert scale from “strongly disagree” to “strongly agree” [41]. The subscales of the Health-ITUES demonstrate high internal consistency and reliability (α=.85-.92) [42], and the construct validity of the Health-ITUES has been established [43]. Health-ITUES scores can typically be difficult to generalize; however, a cut-point score of 4.32 has been suggested as a cutoff to represent a system is “usable” [44]. Health-ITUES was scored by calculating a mean score for each subscale, and a total score by calculating the mean of the subscale scores.

Participants reported their enjoyment with a visual analog scale (VAS), which is a 10-cm scale with anchor phrases at each end [45]. The anchors for our enjoyment VAS were from “not enjoyable at all” to “most enjoyable.” Using an electronic tablet computer, participants touched the line on the spot that best represented their enjoyment. The length of the line is used as a measure of their enjoyment and is reported to be the closest 0.1 cm. Enjoyment VAS scores were converted from centimeters to millimeters and reported from 0 to 100.

Participants rated their perceived exertion using the OMNI 10-point ratings of perceived exertion (RPE) scale [46]. The scale was shown and explained before data collection. Participants could point to or say a number from 0 (extremely easy) to 10 (extremely hard). A researcher confirmed the RPE number with the participant before recording the response. A score of 4-6 would be considered somewhat easy to somewhat hard.

Task self-efficacy was assessed using the video game play appraisal (Multimedia Appendix 1). This scale asks respondents to rate their certainty on 6 dimensions of video game play with the TPS from 0 (no certainty) to 10 (absolute certainty). The scale was created based on expert recommendations [26,47], and the video game play dimensions were chosen to represent the key steps in video game interaction [48]. Among a sample of 30 healthy adults, this scale exhibited high internal consistency (α=.95), and good test-retest reliability with an ICC_3,2 (intraclass correlation coefficient) 0.83 (95% CI 0.62-0.91) [34]. The participants’ understanding of the task itself is vital to the validity of a task self-efficacy scale [47]; therefore, prior to video game play, each participant watched an instructional video illustrating how the TPS is used while sitting in a fixed chair, standing, and seated in a wheelchair.

Participant Feedback and Researcher Observations

After the participant completed usability surveys, they provided open-ended feedback through surveys and semistructured interviews. The survey and interview questions were designed to explore the participant’s perspectives on accessibility, overall experience, and identify areas of improvement. The initial 8 participants were asked to answer open-ended questions on their own. To gain richer feedback from the responses, the remaining participants were interviewed by a member of the research staff using an interview guide (Multimedia Appendix 2). Interviews were audio recorded and later transcribed. The transcriptions were merged with responses to the open-ended questions where appropriate. Researchers documented written observations from gameplay sessions regarding modifications, adaptations, areas of improvement, and suggestions for future touchpad iteration.

Instruments

The TPS is an alternative video game controller that substitutes typical game controls with large movements. The system is designed to control a wide variety of video games available on PC and is easily adapted to users of varying abilities.

The TPS consists of 6 individual touchpads, circuit board, and laptop computer. Each touchpad is a square of particle board with a surface of conductive tape that was placed within a flexible stand that could be adjusted in any direction. All touchpads were wired into a small circuit board microcontroller (Makey-Makey LLC, Santa Cruz, CA) that converts electrical input into computer keys. Touchpads were connected to computer keys that corresponded with video game controls. The microcontroller functions as a video game controller and was connected to the PC via USB cable. The PC was used to run the 4 video game titles that participants played (Table 1). A 127-cm television flat screen was used to display the games 1.8 m in front of the participant.

To use the TPS, the player sat in a chair (Figure 1), sat in their wheelchair (Figure 2), or stood (Figure 3) within a set of parallel bars that they could hold or lean onto for added stability. Once the player was situated, the touchpads were moved around them at a distance that was within reach but made the player lean and reach to make physical contact. The hand or forearm activated touchpads in the front and to the sides of the player were 20 cm × 20 cm in size and placed at trunk height to the players. A larger 61 cm × 31 cm touchpad was created to go behind the player and was placed at scapula height for the seated player and hip height for standing players. To activate this larger touchpad, the participant wore a light harness with a conductive material either around their shoulders (seated) or hips (standing). The player leaned rearward to make contact and activate the larger touchpad.

Procedures

Each person was screened for eligibility by phone before visiting the laboratory. The visit began with reviewing test procedures and then obtaining written informed consent from the participant. Participants answered a baseline questionnaire that asked about their demographics, prior video game experience, and physical activity habits. Next, the participant’s heart rate and blood pressure were measured to ensure the participant was safe to engage in physical activity.

Before the TPS was used, the participant was shown the play area and watched an instructional video that demonstrated both the seated and standing use of the system. After watching the video, the participant rated their self-efficacy before playing. The TPS session consisted of playing 4 different video game titles for 5 minutes each with 5 minutes of rest in between each video game title. The TPS session began with playing a relaxing slow-moving game (Flower) to enable the participant to acclimate to the necessary movements. The sequence of the remaining 3 video game titles was randomized. Immediately at the end of playing each video game, participants were asked to provide an RPE and enjoyment score. At the end of the TPS session, the participant rated their self-efficacy and completed the SUS and Health-ITUES. The entire session was recorded with the participant’s consent.

Data Analyses

Participants’ characteristics are reported as mean (SD) and range. The interitem reliability of SUS, Health-ITUES, and self-efficacy scores was assessed using Cronbach α. A 2-way mixed effects ICC_3,2 was used to assess test-retest reliability of self-efficacy scores. Because self-efficacy and baseline questionnaire responses were not normally distributed, nonparametric statistics were used. A Wilcoxon signed rank test was used to determine if self-efficacy differed from pre- to postgame play. Spearman ρ correlations were calculated to explore relationships among physical activity minutes, video game minutes, usability, enjoyment, perceived exertion, and self-efficacy. For consistency, nonnormally distributed data are reposted as mean (SD). Interview responses and open-ended questions were reviewed by a member of the research team to extract common themes and feedback. Data were analyzed using SPSS (version 27; IBM Corp).

The audio recordings of the participant feedback and written researcher observations were transcribed and combined into the same database. Participant feedback and researcher observations were examined and organized by 2 members of the research team separately. The same 2 researchers classified these data into 3 main categories: accessibility, overall experience, and areas of improvement.

Participant Characteristics

Usability testing was completed by 21 participants 48.8 (13.8) years of age. Our sample consisted of 15 males and 6 females and reported their race as either Black (n=8), White (n=12), or Asian (n=1). Participants reported impaired mobility due to stroke (n=9), spinal cord injury (n=3), amputation (n=3), cerebral palsy (n=2), spina bifida (n=2), or other (n=2). The primary mode of mobility included walking without assistive device (n=7), cane (n=6), prosthetic leg (n=1), rollator walker (n=1), and manual wheelchair (n=6). Participants used the TPS to play either standing (n=8), seated in a 4-legged chair (n=7), or seated in their own manual wheelchair (n=6). All participants were able to complete data collection. Some participants required slight modifications to play such as altering the height of the touchpad, moving the touchpad closer to the player, and adjusting the tilt of the touchpad.

Measures

Participant responses to the baseline questionnaire can be found in Table 2. When asked to rate enjoyment of certain activities from 1 “strongly disagree” to 5 “strongly agree,” participants reported high agreement with both leisure time physical activity 4.6 (0.6) and video games 4.1 (1.1). However, there was high variability in our sample with reported weekly physical activity of 375 (257) minutes and video game play of 398 (643) minutes. There were no sex differences found for any baseline or outcome measures. The participants preferred playing video games on either video game consoles or cell phones.

Usability and subscale scores are reported with summary group scores in Table 3. The SUS demonstrated good reliability (α=.89). Participants reported above-average usability with an average SUS scores of 80.1 (SD 18.5). The Health-ITUES demonstrated excellent reliability (α=.92). However, the mean Health-ITUES score of 4.23 (SD 0.67) did not meet the suggested cutoff of 4.32.

Overall enjoyment and perceived exertion scores including scores by video game title can be found in Table 4. Across all video games, participants moderately enjoyed gameplay with an overall mean VAS score of 70 (SD 22) mm. Perceived exertion of the participants was approximately “somewhat easy” with a mean RPE of 4.3 (SD 2.0).

Self-efficacy scores by dimension and mode of play are included in Table 5. The 6 dimensions of the self-efficacy scale exhibited excellent internal consistency for both the preplay (α=.95) and postplay measures (α=.97). Similarly, the self-efficacy scale demonstrated good test-retest reliability with average measures (ICC_3,2=0.88, 95% CI 0.71-0.95; P<.001). Participants were highly certain in their abilities to use the TPS (mean 7.6, SD 2.2). Self-efficacy did not differ from pre- to postplay.

We found no relationships between weekly physical activity or video game play minutes and outcome variables. Both usability measures were moderately correlated with each other (r_s=0.50). Enjoyment and perceived exertion did not exhibit a relationship with other outcome variables. Self-efficacy after using the TPS was moderately correlated with Health-ITUES scores (r_s=0.43).

Participant Feedback and Researcher Observations

Principal Findings

We assessed the usability of the TPS and explored the enjoyment, perceived exertion, task self-efficacy, participant feedback, and researcher observations. In total, 21 individuals with impaired mobility played several video game titles (Table 1) while sitting in a chair, standing, or sitting in their own wheelchairs. A promising result was that every participant was able to use the TPS to play all video game titles for at least 5 minutes; however, numerous modifications were needed to foster the experience for many of the participants.

Consistent with other AVG controllers for people with impaired mobility, the participants found the TPS usable (Table 3) [13] However, the Health-ITUES scores were slightly below the suggested benchmark [44]. Only the participants who used the TPS standing reported a minimum score above the usability threshold of 68. The participants who played sitting in a chair reported a mean score barely above the threshold. This suggests the TPS can be improved to be more usable to seated players. Additionally, our overall Health-ITUES mean did not meet the suggested cutoff score of 4.32. Interestingly, our chair-sitting group reported the lower usability scores with higher variability than those who played sitting in their own wheelchairs, which is not consistent with the usability results of another AVG controller we were testing in our laboratory [13]. The chair we provided gameplay was consistent but every wheelchair player used their own device. Therefore, participants’ own personalized devices likely provide a more comfortable gameplay environment, which may have affected their usability scores.

Similar to previous studies using AVGs among individuals with impaired mobility, participants moderately enjoyed using the TPS (Table 4) [20] Similar to their peers without impaired mobility, participants had moderate to high self-efficacy using the TPS before and after gameplay [34,49]. Our lowest self-efficacy scores were reported by participants who played sitting in a chair, who also reported the lowest usability, enjoyment, and highest perceived exertion. It is possible that the participants sitting in a chair to use the TPS had a less positive overall experience than their peers who played standing or from their own wheelchair. The only 2 participants who reported below moderate enjoyment were individuals’ poststroke who reported not playing video games (0 minutes per week). Both individuals also reported low self-efficacy postplay and low usability scores. While AVGs can enhance self-efficacy [29,32], lower enjoyment and lack of experience playing AVGs may reduce self-efficacy.

These data suggest that our participants perceived their exertion as somewhat easy to somewhat hard during video game play using the TPS (Table 4), which is consistent among individuals with neuromuscular conditions [50]. It is unclear why the 3 participants with the highest mean perceived exertion were men >50 years. To better understand the influence of the warmup game (Flower) on perceived exertion, we calculated an exercise RPE by removing all the warmup game RPE values. This did not alter exertion. Our observed perceived exertion scores are consistent with RPE observed from previous TPS testing [34]. Our RPE findings are also comparable to scores reported by adults with impaired mobility, playing video games using an adapted Nintendo Wii balance board and an adapted gaming mat from a sitting and standing position [18,20].

The participants regarded the touchpads as novel, fun, and entertaining but they did encounter some accessibility barriers. Modifications such as repositioning the touchpads, adding conductive tape to prosthetics, and providing moisturizer for dry skin were not anticipated because they were not encountered in previous testing of the TPS [34] or with a similar controller in our laboratory [13]. While successful efforts were made during the study to overcome accessibility barriers and enable players who had difficulty using the palm of their hand to contact the touchpads, the need to adjust the position of the touchpads multiple times in a single session may have detracted from the participant’s experience. The subgroup that encountered the most frequent difficulty contacting the touchpads using their palm played sitting in a chair, which may account for the lower usability among this subgroup.

For future iterations, the stands used to mount the touchpads need to resist movement when a player exerts considerable force to activate. Due to the large body movements and quick reactions required to play video games using the TPS, the touchpads need to stay fixed in position during gameplay. Even though each touchpad was made large to mitigate the need for movement precision, touchpad movement during gameplay may cause the player to lose screen focus and thus introduce frustration. The TPS needs to remain robust to a varying degree of force from multiple directions to better accommodate individuals with impaired mobility. The TPS is currently being refined based on usability, participant feedback, and researcher observations. This system can be refined to work with adaptive video game equipment such as the Microsoft Adaptive Controller. Because the TPS is not limited to a single user, future research should examine the use of this system for multiple players simultaneously.

Limitations

First, the participants in this sample represented various mobility impairments; therefore, the number of players within each mode of play is few, which makes mode comparisons difficult. Second, our participants reported a wide range of weekly physical activity minutes (Table 2), limiting our ability to generalize these results to sedentary individuals with impaired mobility, and it is possible that physically active individuals may find the TPS more usable. Third, video game selection was limited to specific titles we felt could be demonstrated and learned quickly. Therefore, we chose video game titles that required simple commands and may not be indicative of more complex games. Additionally, the featured video games may not have appeal to some participants. Finally, we did not standardize touchpad placement relative to the player due to varied levels of dexterity. We used a consistent touchpad layout, and modifying the touchpad layout may limit the generalizability of our results; however, we feel this decision engendered rich feedback regarding the accessibility of the TPS.

Conclusions

The TPS enabled people with impaired mobility to participate in AVG play. The TPS allows typical sedentary video games to be played by adults with impaired mobility while sitting, standing, and with their own mobility aids. Participants found the TPS to be usable, experienced moderate enjoyment, and achieved moderate intensity physical activity through gameplay. Key areas of improvement to the system were identified based on our measures, participant feedback, and observation.

Data Availability

The data sets generated and analyzed during this study are available from the corresponding author upon reasonable request.