Housing insecurity and cancer in South Australia: artificial intelligence-enabled insights from a statewide cancer support service

OBJECTIVE: This study demonstrates the use of a locally hosted, offline artificial intelligence approach to identify, quantify, and describe housing-related needs documented in routine case notes from a statewide cancer support service. METHODS: In this retrospective, exploratory service evaluation, a locally hosted, offline large language model screened 25,031 staff-generated case notes recorded between December 2020 and December 2025 for housing-related interactions and grouped confirmed interactions into categories based on a predefined framework. Flagged records and categorisation were independently verified by at least two reviewers, and confirmed interactions were summarised descriptively. RESULTS: Of 25,031 interactions, 308 (1.23%) were confirmed as housing-related interactions. The most prominent categorisations were lack of support from health and social services (44.16%), financial barriers to housing stability (37.66%), and homelessness or reliance on emergency or crisis housing (36.04%). In more than half of confirmed interactions, housing was not the initial reason for contact and emerged only during discussion. CONCLUSIONS: Findings highlight the utility of a privacy-preserving artificial intelligence approach to identify, quantify, and describe community needs at the intersection of housing and cancer care within large-scale, routinely collected service data. Although drawn from a single service, the findings align with existing evidence supporting coordinated pathways linking oncology, social, and housing services, and greater policy recognition of housing within cancer care.