Feasibility of a Co-Designed Personalized Cognitive Reablement Program Delivered via Digital Voice Assistants for Older Adults Living With Early-Stage Dementia or Mild Cognitive Impairment: Pilot Randomized Controlled Trial.

BACKGROUND

Home-based reablement programs have improved everyday functioning for people living with early-stage dementia or mild cognitive impairment (MCI), but the substantial time commitment required for delivery by health professionals has created a barrier to widespread implementation. Digital technology could facilitate delivery, but traditional devices such as smartphones and tablets present usability challenges for this population. We have co-designed a cognitive reablement program, VOICE-COG, that can deliver personalized interventions broadcast via voice-controlled digital voice assistants (DVAs).

OBJECTIVE

This pilot nonblinded randomized controlled trial examined the feasibility and potential effectiveness of VOICE-COG for older adults living with dementia or MCI to guide the design of a full-scale trial.

METHODS

Older adults (aged 60-85 years) with a self-reported diagnosis of dementia or MCI (N=30) were recruited via Facebook or the StepUp-for Dementia Research database and randomized to either a DVA intervention (n=15) or control (n=15) group for 12 weeks. Each participant randomized to the intervention group and their care partner met with a psychologist via Zoom to select 2 or 3 personalized goals from the VOICE-COG program. The personalized program was input into the software (BuddyLink) by researchers for broadcast via a DVA (Amazon Alexa Show 8). Participants in the control group received enhanced usual care, consisting of treatment as usual, plus weekly informational emails. Primary feasibility outcomes were recruitment, retention, device-recorded adherence, and the System Usability Scale (SUS). Secondary outcomes were self-reported by the participant and care partner via web-based Qualtrics forms and included the Goal Attainment Scale-Light (GAS), changes in participant mood, cognition, and everyday functioning, and carer burden.

RESULTS

The recruitment rate was 86% (30/35) and the retention rate was 97% (29/30), with 1 participant in the intervention group lost to follow-up. Adherence to the VOICE-COG program ranged from 29% to 100% (mean 70%, SD 17%). The VOICE-COG program was rated as marginally acceptable (mean SUS score 56.3, SD 18.1). After 12 weeks, 60% (9/15) of the intervention group reported having met or exceeded their goals (GAS-Light t-score ≥50; mean 57.6, SD 15.4). The intervention group had a significantly higher increase in Multifactorial Memory Questionnaire-Strategy subscale score compared with control, mean difference 5.0 (95% CI 0.03-10.1, P=.049), representing a medium effect size (d=0.55). There were no other significant differences between groups.

CONCLUSIONS

This co-designed VOICE-COG program, delivered via DVAs, was feasible for older adults living with early-stage dementia or MCI, resulting in most participants meeting their reablement goals and increases in the use of memory strategies. System usability issues should be addressed in future adequately powered trials.