Artificial intelligence promises to revolutionise hospital quality reporting, study finds
A pilot study led by researchers at the University of California San Diego School of Medicine suggests that advanced artificial intelligence (AI) could revolutionise hospital quality reporting, making it faster, simpler, and more efficient while maintaining high levels of accuracy. This transformation, researchers argue, could pave the way for enhanced health care delivery by easing the reporting process and optimising resource use.
The study’s findings, published on 21 October 2024 in the New England Journal of Medicine (NEJM) AI, reveal that a large language model (LLM)-driven AI system can accurately process hospital quality metrics, reaching a 90 per cent agreement rate with traditional manual reporting. The potential impact, the authors say, could be profound, as the system may replace laborious manual methods with a faster, more reliable approach.
In partnership with the Joan and Irwin Jacobs Centre for Health Innovation (JCHI) at UC San Diego Health, the study’s researchers demonstrated that LLMs could accurately abstract complex quality measures, including those required for the challenging CMS SEP-1 measure, which evaluates severe sepsis and septic shock. This capability is a key milestone, as the SEP-1 measure alone demands a rigorous 63-step review of patient records, a process which often takes weeks and requires multiple reviewers.
Aaron Boussina, the lead author and postdoctoral scholar at UC San Diego School of Medicine, expressed optimism about the transformative impact of AI on health care systems. “The integration of LLMs into hospital workflows holds the promise of transforming health care delivery by making the process more real-time, which can enhance personalised care and improve patient access to quality data,” he said. “As we advance this research, we envision a future where quality reporting is not just efficient but also improves the overall patient experience.”
Traditional SEP-1 reporting involves a painstaking review of detailed patient records, a task that currently requires weeks of staff time across multiple reviewers. By using LLMs, however, the study found that hospitals could vastly reduce both the time and resources needed for this process. AI systems could rapidly scan patient charts and generate critical insights within seconds, automating significant portions of the abstraction process while maintaining accuracy.
This AI-driven approach not only simplifies reporting but also allows health care staff to redirect their focus from manual administrative tasks to patient care. “We remain diligent on our path to leverage technologies to help reduce the administrative burden of health care and, in turn, enable our quality improvement specialists to spend more time supporting the exceptional care our medical teams provide,” said Chad VanDenBerg, co-author of the study and chief quality and patient safety officer at UC San Diego Health.
Beyond simplifying processes, the study highlights several other potential benefits of using AI for quality reporting:
- Increased Accuracy and Error Reduction: LLMs can detect and correct errors in real time, reducing discrepancies in reporting.
- Reduced Administrative Costs: Automation of routine tasks allows for significant cost savings.
- Near Real-Time Quality Assessments: LLMs can provide more immediate quality evaluations, enabling quicker responses to patient care needs.
- Scalability Across Various Health Settings: The adaptability of LLMs means they could be applied in different health care environments with ease, potentially standardising quality reporting.
The study’s next steps involve validating these findings through larger trials and working towards implementing the system in real-world health care settings. The researchers are hopeful that, in time, this innovation will support reliable, efficient data and reporting methods that are essential for quality care provision.
The study authors include Shamim Nemati, Rishivardhan Krishnamoorthy, Kimberly Quintero, Shreyansh Joshi, Gabriel Wardi, Hayden Pour, Nicholas Hilbert, Atul Malhotra, Michael Hogarth, Amy Sitapati, Karandeep Singh, and Christopher Longhurst, all affiliated with UC San Diego.
The research was partially funded by the National Institute of Allergy and Infectious Diseases (1R42AI177108-1), the National Library of Medicine (2T15LM011271-11 and R01LM013998), the National Institute of General Medical Sciences (R35GM143121 and K23GM146092), and JCHI.
