For hospitals facing immense pressure to report quality measures accurately while managing tight resources, implementation of large language models (LLMs) could prove incredibly useful.
Currently, quality reporting is both time-consuming and costly—some estimates show it can cost hospitals millions annually and require thousands of work hours. By automating elements of the process with LLMs, the authors of the NEJM AI Case Study believe that hospitals can drastically reduce the manual workload involved in reviewing patient charts, freeing up staff to focus more on patient care.
If properly implemented and monitored, LLMs can help streamline the complex, 63-step SEP-1 reporting process, making it easier for non-clinical staff to gather the required data quickly and accurately. For clinical staff, this means fewer disruptions in their workflow, less time spent on data entry, and more timely feedback on patient outcomes. This can lead to faster implementation of quality improvements, ultimately improving patient care.
Moreover, the authors envision that by using LLM-based systems human error and variability in reporting can be reduced, which would ensure consistency across different staff members and hospital sites. While these advancements can boost reliability and can provide hospital administrators with more accurate data for decision-making - it’s not just about efficiency - it’s about creating a more reliable and less stressful environment for hospital staff, enabling them to dedicate their energy to improving health outcomes.
Kudos to Aaron Boussina, Rishivardhan K, Kimberly Quintero for their insights and interests in leveraging technology to make health systems more efficient and ultimately better for patients.
