Dementia carers’ perspectives on paper- and artificial intelligence (AI)-based assessment tools: a qualitative think-aloud study using the COM-B framework

Objectives

Carer observations provide valuable longitudinal and contextual information for dementia assessment, but traditional informant-based questionnaires may limit reporting of nuanced, fluctuating changes. Artificial intelligence (AI)-based tools may offer greater flexibility to report, but their successful uptake depends on carer capability and acceptability. Despite this, little is understood about carers' views on existing informant questionnaires and emerging AI-based tools, therefore, this study explored carers’ perspectives on barriers and facilitators influencing engagement with paper- and AI-based assessment tools.

Methods

Ten carers of people living with dementia completed think-aloud interviews, while reviewing the Cambridge Behavioural Inventory-Revised (CBI-R) and two AI-based tools, CognoSpeak™ and LUMEN (Language Usage for Monitoring Early Neurodegeneration). Using a hybrid inductive-deductive approach, themes related to barriers and facilitators to engagement were identified and subsequently mapped onto the Capability, Opportunity and Motivation-Behaviour (COM-B) framework.

Results

Carers valued the CBI-R’s simplicity and breadth but found its structure, complex language and limited free-text space restrictive. Verbal, narrative responses felt more natural and less burdensome, therefore, AI-based tools were seen as more engaging and better able to collect richer narratives, with the added advantage of remote access, however, digital literacy, sensory impairments, privacy concerns could limit uptake. Participants stressed that AI should support, not replace, human-led care.

Conclusion

Engagement depended on individual capabilities, available opportunities and motivation; factors fundamental to successful uptake. Future tools should be co-designed to ensure accessibility, cultural sensitivity and clinical relevance.