New musical interfaces for older adults in residential care: assessing a user-centred design approach

Abstract

Purpose:For older adults in aged-care, group music-making can bring numerous physical and psychological benefits, ultimately improving their quality of life. However, personalising music-making to optimise these benefits is often difficult given their diverse ages, experiences, abilities, cognitive and motor skills, and their experience with music technology.

Materialsandmethods:In this study, we conducted a 10-week group music-making intervention with twenty participants in an aged-care home, using a prototype digital musical instrument that we iteratively refined by following a user-centred design approach from direct resident feedback. The prototype instrument adopted a novel method for errorless learning in music-making settings, which we also refined, by increasing the difficulty level of the instrument's operation. We also assessed the residents' engagement with the sessions by obtaining feedback from caregivers and facilitators.

Results:Results show that residents' enjoyment decreased as the complexity (difficulty) of our errorless learning implementation increased. We also found that resident engagement increased when changes to the prototype digital musical instrument were provided, but not when residents were giving feedback. Results also found that participation over the course of the intervention, and the number of songs played during each session also enhanced engagement.

Conclusions: Overall, our results show the intervention was beneficial to residents, although we note some areas of enhancement for further interventions in designing prototype musical instruments for group music-making in aged-care settings.

 

IMPLICATIONS FOR REHABILITATION

• Older adults positively engage with novel music technology, and do so increasingly over subsequent sessions. Repeated sessions may have the potential to enhance longer-term adoption of technologies as well as any rehabilitative effects of the group music-making activity.

• There is significant potential for residents with different abilities to all make music together, although to maximise the sustainability of the devices, the sessions, and the subsequent rehabilitative benefits, residents must be given the right adaptation for individual interfaces that balances ambition and ability.

• Rapid DMI prototyping positively enhances engagement among older adults, suggesting that in the case of a custom DMI, an upgrade schedule should be aligned with key rehabilitative milestones. Similarly, in the case of pre-developed digital music systems, resident exposure to new features or functionality should be strategically introduced, so as to maximise engagement for key phases of resident rehabilitation.

Keywords:

• Music technology
• digital music instrument
• user-centred design
• residential aged-care
• older adults
• group music-making

Acknowledgements

The authors would like to acknowledge Matthew Breaden, Dave Anthony, Felix Dobrowohl, Alison Short, Ehab Ghobreyal, Khushali Surti and the lifestyle team at the residential aged-care facility in Western Sydney. This project was funded by the NSW Department of Communities and Justice under the Liveable Communities Grants Program (LC-0096).

Declaration of interest

The authors report no conflicts of interest.

Notes

1 The second period experienced a delay in completing the sessions due to an influenza lockdown at the facility.

2 Although illness and other reasons not to attend were not captured by this process.

3 See https://vbn.aau.dk/ws/portalfiles/portal/316447795/Music_in_Dementia_Assessment_Scales_MiDAS_.pdf for an example MiDAS rating scale (University of Aarhus, Denmark).

4 We had hoped for a positive predictive influence of DMI use on MiDAS ratings, as part of Hypothesis 2. Specifically, we wanted to evaluate the efficacy of the custom-made DMIs, as compared to off-the-shelf DMIs. Three different Bayesian models were fitted to assess the possible effect of DMI use. In each case, predictors for each DMI were added to the model in Table 4. For the three models these additional predictors were, respectively: (1) a binary code for whether each given DMI was played at least once per session per participant; (2) a count of the number of times each given DMI was played per session per participant; (3) same as (2) but also including a squared count term to allow for a nonlinear effect. All three models performed worse under cross-validation than the Table 4 model, which does not include DMI. This means that our data are not sufficiently informative for us to make useful inferences about whether the different DMIs have different effects on our participants’ MiDAS scores.