Development of a Serious Game App (Digimenz) for Patients With Dementia: Prospective Pilot Study for Usability Testing in Inpatient Treatment and Long-Term Care.

Background: In the face of an increasing treatment need among people with dementia, effective and efficient interventions with a focus on quality of life need to be established. In this context, serious games have received increasing attention. However, there is a lack of apps specifically designed for people with dementia. Objective: In this prospective pilot study, we examined the usability of a newly developed serious game app ("Digimenz"). Methods: A total of 43 people with cognitive impairment and mild to severe dementia completed the repeated-measures study procedure. Participants were recruited from an inpatient geriatric psychiatric ward and a long-term care facility. Participants were asked to complete 4 conditions in randomized order, including playing 3 different serious games (experimental conditions) and reading a newspaper (control condition). Each condition was completed once, and the total duration was 60 to 90 minutes per participant. Data on app usability were collected through self-ratings and observation after each condition. We tested for differences in usability among the conditions and the recruitment sites, and analyzed the relation of usability to cognitive capacity. Results: The serious games were accepted in both settings (long-term care: 30/30, 100% interested; psychiatric ward: 31/41, 76% interested), although study completion was lower in the psychiatric subsample (15/41, 37%) than in the long-term care subsample (28/30, 93%). Global usability was rated good (System Usability Scale global mean score: 79). More severely impaired patients had more pronounced difficulties in learning how to play the games (ρMMSE, Learnability=-0.61, 95% CI -0.78 to -0.36; P<.001) and playing them alone (ρMMSE, Support=-0.49, 95% CI -0.69 to -0.19; P<.001). Nevertheless, playing the games was associated with a more positive mood (likelihood ratio χ23=25.09; P<.001), independent of the level of cognitive functioning (likelihood ratio χ21=0.64; P=.42). All games were played with a moderate error rate (0.19-0.49). Conclusions: Our results indicated a positive association between serious game usage and well-being in patients with dementia, given adequate support. This is a valuable addition to the understanding of serious game usage in dementia care. Although challenging, user-centered development of serious games with people who are severely impaired by dementia is an important research target. Limitations like low data quality and a simplified design are inherent in this study population. Nevertheless, we demonstrated how usability testing in this target group is possible through careful definition and operationalization. The inclusion of different data sources, different recruitment sites, and different levels of cognitive impairment increased the generalizability of the findings. To accommodate severely impaired patients, future developments should incorporate a broader range of difficulties and adaptations to group settings.