Artificial Intelligence–Supported Patient Navigation Addressing Social Drivers of Health among Vulnerable Adults in Family Medicine and Geriatrics Safety-Net Clinics

Abstract: Health-related social needs such as housing instability, food insecurity, and transportation barriers are associated with worse health outcomes and increased health care utilization, particularly among older adults. Although health systems increasingly screen for these needs, many lack the capacity to provide follow-up support at scale. This pilot evaluation described the implementation and early operational performance of an artificial intelligence-supported, on-demand patient navigation service integrated into an urban safety-net health system. From July 2024 to March 2025, providers placed 149 referrals from family medicine and geriatrics clinics. Outreach was attempted for all referred patients, and successful contact occurred for 78 patients. Navigators submitted 301 applications for public benefits and community resources; 145 were approved, while others remained pending or in process at analysis. Program costs during the observation period were approximately $5,000. These findings suggest feasibility while highlighting the need for further evaluation of patient experience and outcomes.