The role of artificial intelligence in blood-borne virus opt-out testing in emergency departments

Blood-borne viruses (BBVs) such as HIV, hepatitis B, and hepatitis C continue to pose serious public health concerns, particularly within emergency departments (EDs), where patient volume and turnover are high. While opt-out testing strategies, where individuals are tested unless they specifically decline, have shown effectiveness in increasing diagnosis rates, their adoption in EDs is limited by challenges such as inefficient workflows, data fragmentation, and suboptimal patient engagement.

This narrative review aims to explore the application of Artificial Intelligence (AI) in enhancing BBV opt-out testing in EDs, focusing on how AI can address current operational and clinical challenges while supporting ethical and equitable implementation.

A structured narrative review approach was used following established guidelines. We searched PubMed, EMBASE, Web of Science, and grey literature from 2010 to 2024 using terms related to AI, blood-borne viruses, opt-out testing, and emergency departments. A total of 32 articles were included in the final synthesis.

AI demonstrates theoretical potential with limited BBV-specific empirical evidence in improving BBV testing outcomes through automated patient identification and risk stratification using electronic health records. Evidence from broader healthcare AI applications suggests workflow improvements may be possible through automated test ordering, real-time alerts, and adaptive scheduling systems. Data analysis tools have shown promise in other healthcare contexts for accurate test result interpretation and epidemiological trend identification. AI-driven patient communication tools such as chatbots and mobile apps show potential to enhance patient understanding and reduce opt-out rates. Follow-up and continuity of care could potentially be strengthened via automated notifications and predictive adherence models.

AI offers potential opportunities to improve the scalability, efficiency, and equity of BBV opt-out testing in EDs. However, successful integration depends on addressing ethical issues, algorithmic bias, and system interoperability, supported by interdisciplinary collaboration and continuous evaluation. Further research with BBV-specific evidence is urgently needed to validate these theoretical applications.