BACKGROUND
Characterising older adult engagement is important to determine the effectiveness of interventions. Engagement is the occupying of oneself in external stimuli and is observable across multiple dimensions of behaviour. Engagement of older adults is commonly investigated on a single behavioural dimension.
OBJECTIVE
In this article, we present a multidisciplinary approach for measuring and characterising engagement of older adults using techniques appropriate for people with varying degrees of dementia.
METHODS
Contexts for engagement included a dyadic reminiscence therapy interview and a 12-week technology driven group reminiscence therapy. Participants were older adults (8 female, 1 male, mean age: 79) who attended a day respite facility. Audio-visual recordings of the sessions were processed to analyse facial movement, lexical use, and prosodic patterns of speech. Facial movement was processed using OpenFace to measure the presence and intensity of facial movement. Lexical use was processed using the Linguistic Enquiry and Word Count to measure personal pronoun use, affective word use, and emotional tone of words in speech. Prosodic patterns of speech were processed using custom scripts written in Praat and Python, to measure mean duration of utterances, mean words per utterance, articulation rate and variability of F0. Mixed-effects modelling was used to assess effects of treatment conditions on dependent variable outcomes.
RESULTS
Results indicate measuring engagement through a multidimensional approach can sensitively capture older adults’ engagement.
CONCLUSIONS
Application of this method can enhance a researcher’s ability to measure older adult engagement, provide means to compare across interventions and contextual environments, and further develop the science of psychosocial intervention research.