Briefing: Examining the Impact of Artificial Intelligence on Healthcare Safety (R18)

STATUS: EXPIRED (March 17, 2026). Limited case-by-case acceptance may apply under NIH late/continuous submission policies; contact eRA Service Desk.

Research Focus

This Agency for Healthcare Research and Quality (AHRQ) program funds research on how artificial intelligence (AI) systems affect patient safety in clinical practice. The funder defines AI as software applications capable of learning, decision-making, automated analysis, model inference, and content generation analogous to human cognition. Research must address two priority areas: (1) whether and how breakthrough AI uses affect patient safety, and (2) how AI systems can be safely implemented and used in healthcare delivery settings.

Breakthrough AI is defined as a significant advance removing hazards or barriers in healthcare processes—including large language models (LLMs) and generative AI for clinical documentation and clinician inbox queries. The program seeks evidence on whether these systems actually reduce clinician burden and burnout, improve documentation quality, decrease "pajama time" (work completed outside normal hours), and whether they introduce new patient safety risks. Research may also explore AI applications in underserved areas (e.g., rural clinician shortages) and their safety implications.

The second focus examines implementation science: workflow integration, usability, human-computer teaming, governance for post-implementation monitoring, and user interaction with AI systems that have already demonstrated efficacy in healthcare (e.g., medication selection, diagnostic accuracy, patient monitoring, treatment prediction). This is not a development program; it funds assessment of actual AI deployments in healthcare systems.

At a Glance

• Who can apply: Not stated (see AHRQ eligibility guidance and SF424 R&R Application Guide)
• Funding & project length: Not stated
• Award mechanism: R18 (Research Demonstration and Dissemination Project)
• Key dates: First application due September 25, 2024 (expired); peer review ~4 months after receipt; earliest start ~4 months after peer review
• Best fit for: Health services researchers, implementation scientists, and clinical informaticists studying AI safety, clinician burden, workflow integration, and patient outcomes in hospital and ambulatory care settings