Rhys Mantell (r.mantell@unsw.edu.au, website)
School of Population Health, University of New South Wales, 2025
Abstract
This thesis takes aim at the cognitive health crisis facing older people “ageing at the margins”—those experiencing homelessness, incarceration, or persistent poverty. Despite extraordinarily high rates of age-related cognitive decline in these populations, cognitive conditions (i.e., dementia) remain underdetected due to limited service access, institutional mistrust, and inappropriate assessment tools. Game-based cognitive assessments (GBCAs) offer promising alternatives to traditional screening by providing engaging, scalable, cost-effective, and ecologically valid approaches. However, realising their potential requires addressing two interconnected challenges: (1) designing GBCAs that appropriately engage underserved cohorts whilst demonstrating adequate psychometric properties, and (2) understanding the structural, social, and psychological conditions that enable or constrain implementation in challenging contexts like prison. Anchored in critical realist philosophy, this work investigates both challenges through co-design focus groups with older people in prison, systematic review of existing GBCA literature, stakeholder analysis of implementation barriers, and mixed-methods evaluation of ASCAPE-C—a novel GBCA battery designed specifically through this work for people ageing at the margins. Prison serves as a critical test case: if GBCAs can work in this restrictive, under-resourced environment, they have genuine transferability to other underserved contexts. The thesis makes several contributions. Theoretically, it identifies the Failure State Conundrum—cognitive assessment requires users to encounter failure states to measure their limits, yet such failures risk triggering frustration and disengagement—and specifies conditions under which this tension can be resolved. The Miyazaki Principle (challenge with comprehension tends to create satisfaction; challenge without comprehension creates frustration) explains why some design choices enable engagement whilst others trigger disengagement. Methodologically, critical realist-inspired causal analysis identifies six mechanisms affecting user engagement and three mechanisms constraining organisational implementation. Additional retrodictive analysis reveals how these mechanisms interact across laminated structural, institutional, and individual levels through downward causation, compensatory processes, and failure cascades. This demonstrates how critical realism can support the design and adoption of health interventions in complex health settings by explicating generative mechanisms and their emergent interactions—insights that remain obscured in conventional analytical and evaluative approaches. Practically, co-design with older people in prison generated several design principles that informed ASCAPE-C’s development. User evaluation demonstrated that when GBCAs are designed with rather than for people ageing at the margins, they can engage users whilst collecting reliable cognitive data, demonstrating the viability of this approach to address the cognitive crisis in underserved populations.
Methods
Mixed