Assessing the Value for Money of AI-Assisted Technologies for Older Adults: Scoping Review of Economic Evaluations

Abstract Background As populations age globally, AI-enabled digital health interventions (DHIs) are increasingly being adopted to support integrated, person-centered care for older adults. However, despite rapid advances in AI technologies, their economic value in older care remains poorly understood. Objective This scoping review aimed to map the existing literature on the economic evaluations of AI technologies for older adults by (1) identifying the types, characteristics, economic outcomes, and methodological approaches reported; (2) mapping the evidence across the World Health Organization’s (WHO) Integrated Care for Older People (ICOPE) pathway; and (3) identifying evidence gaps and priorities for future research. Methods A scoping review was conducted following the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) guidelines. PubMed, Scopus, Embase, Web of Science, and EconLit were searched for studies published up to July 31, 2026. Eligible studies were economic evaluations of AI-based technologies in older health care. Conference abstracts, reviews, technical reports, protocols, letters, trial registrations, and non-English studies were excluded. Methodological quality and reporting quality were assessed using the Criteria for Health Economic Quality Evaluation (CHEQUE). Data were synthesized using descriptive statistics and narrative synthesis, with findings mapped to the 4-step ICOPE care pathway. Results Forty studies published between 2018 and 2026 met the inclusion criteria. Methodological and reporting quality were generally high (mean scores: 85.7/100 and 85.4/100, respectively), although equity considerations, subgroup heterogeneity, and model validation were frequently underreported. Most evaluations examined AI for screening and diagnosis (32/40), particularly in cancer and ophthalmology, and primarily used model-based approaches, including decision trees, Markov models, and discrete-event simulations. AI-related costs were frequently obtained from assumptions, manufacturer quotes, or expert opinion. Twenty-one of the forty evaluations reported AI interventions to be cost-saving, and another 17 reported AI interventions to be cost-effective. Confidence in these findings is constrained by methodological limitations, reliance on modeled assumptions, and incomplete reporting of AI-related costs. Key drivers of cost-effectiveness included AI performance, AI-related costs, population characteristics, disease burden, and health system context. Along the ICOPE pathway, evidence was concentrated in screening and diagnostic interventions, with little economic evaluation of personalized care planning or long-term monitoring. Conclusions This review summarizes the currently available economic evidence on AI-assisted technologies in older health care following the ICOPE care pathway. It extends beyond prior reviews that have assessed AI cost-effectiveness across general or disease-specific populations without addressing the distinct cost structures, care needs, and equity considerations of older adults. By mapping the included evidence against the ICOPE domains, this review identifies where economic evidence is concentrated and where it is critically lacking, providing structured guidance for policy design and future research across different components of the care pathway.