Long-Term Staff Acceptance of an AI-Supported Radar-Based Bed-Exit Surveillance System After Clinical Implementation

Background: Contactless bed-exit surveillance may support fall prevention, but little is known about healthcare professionals’ long-term acceptance after implementation. This study compared staff acceptance of two artificial intelligence-supported bed-exit surveillance technologies—an infrared camera-based system (Cogvis) and a three-dimensional radar-based system (QUMEA)—during technology selection and evaluated acceptance of the selected QUMEA system over three years of routine clinical use. Methods: Repeated cross-sectional surveys were conducted in 2021, 2022, and 2024 on a specialized acute geriatric ward for patients with delirium, yielding 52 questionnaires. Fourteen common acceptance items were rated on a four-point scale and dichotomized. Paired tests compared Cogvis and QUMEA in 2021, and mixed-effects logistic regression evaluated QUMEA acceptance over time. Results: QUMEA received higher positive ratings than Cogvis across nearly all items in 2021. The adjusted probability of a positive QUMEA response was 95.4% in 2021, 73.6% in 2022, and 86.4% in 2024 (overall p = 0.003). Acceptance remained lower in 2024 than in 2021 (p = 0.021). Conclusions: Staff acceptance remained high but was lower after technology selection and only partially reached former levels. Use over time revealed challenges related particularly to false alarms and perceived burden, emphasizing the importance of technical performance and workflow fit.