AI-enabled whole-of-home resident profiling and connection matching in residential aged care: a conceptual framework and simulation study
Loneliness is common in residential aged care, and staff across a home learn different things about each resident. This knowledge is rarely pooled to help residents connect. This paper proposes a whole-of-home framework enabled by artificial intelligence (AI). Observations from every staff role, including how residents get on with each other, are pooled into a living profile, and a transparent matching method suggests introductions that staff and residents decide on. The matching step was examined in 400 synthetic homes of 80 residents, built from published Australian statistics and stated assumptions. A hidden compatibility model, specified by the authors and never shown to the matching method, decided which pairings worked. No real resident, staff or facility data were used, and no AI model was trained or run on real data. In the simulation, whole-of-home matching placed an expected 48.6% of residents in a pairing that worked under the hidden model, against 31.7% for pairing residents on the same level (difference 16.9 points, 95% simulation interval 16.4 to 17.4). Widening the profile from care-team to whole-of-home observation was associated with a 5.7-point increase. Assigning the highest-scoring pairs first left 14.8% of residents with no suggestion, and residents preferring a language other than English gained least. The advantage fell to about 4 points when staff recorded a quarter of what was assumed. These figures describe behaviour under stated assumptions. They are not estimates of real-world effect and do not show that residents would feel less lonely. A pilot is the next step.
Authors
- Helen Dorathy Uba
- Ebere Josephine Uba
Institutions
- Western Sydney University (AU)
Publication Details
- Journal
- Research Square
- Published
- 2026-11-20
- DOI
- https://doi.org/10.21203/rs.3.rs-11440401/v1
- Primary Topic
- Geriatric Care and Nursing Homes
- Type
- preprint