Shedding light on the future of dementia diagnosis

Abstract

As the global burden of dementia continues to rise, there is a growing need for accessible, scalable, and precise approaches to cognitive and functional assessment in dementia diagnosis and care. Digital cognitive and functional assessments hold promise in this regard, particularly as disease-modifying treatments increase the demand for timely and accurate diagnosis. However, digital tools are often introduced before their psychometric quality, clinical relevance, and fit with end users have been sufficiently established. This thesis therefore aimed to strengthen the evidence base for digital cognitive and functional assessment across the dementia diagnostic pathway, from population-based and primary care settings to the memory clinic. The key findings show that cognition and everyday functioning, although related, are distinct constructs that should be assessed separately, as cognitive performance explained only a modest proportion of variance in daily functioning (Chapter 2). A scoping review of smartphone- and tablet-based cognitive assessments revealed rapid tool development but sparse and inconsistent psychometric evidence, underscoring the need for more rigorous validation before clinical implementation (Chapter 3). End users, including older adults across the Alzheimer's disease continuum, their caregivers, and primary care physicians in the Netherlands and Norway, were generally positive about digital tools, while also identifying concrete requirements for successful use, such as flexibility to accommodate individual preferences, transparency about validity, ease of use, and integration with existing clinical workflows (Chapters 4, 5). Building on this, the web-based cognitive test cCOG was shortened without loss of diagnostic accuracy, refined based on end user feedback, and shown to have good test-retest reliability when completed remotely and unsupervised; population-based norms further enabled meaningful interpretation of individual scores (Chapters 6, 7). The digital Amsterdam IADL Questionnaire (A-IADL-Q) was, for the first time, validated in primary care, demonstrating strong diagnostic accuracy across clinical stages and minimal influence from demographic factors (Chapters 8, 9). Together, this thesis provides the first evidence that both cCOG and the A-IADL-Q can be reliably and meaningfully used across population-based, primary care, and memory clinic settings. The research in this thesis highlights the persistent gap between the promise of digital innovation and its translation into daily clinical practice, and shows that closing this gap requires not only robust psychometric evidence, but also active engagement of end users and attention to the structural preconditions for implementation. Finally, this thesis provides key recommendations, grounded in Knoster's model of complex change, to guide the road towards a digitized memory clinic.