AI Speeds Up Cancer Care Safety Reviews by 29 Times

A new artificial intelligence tool is helping doctors catch safety problems before they harm patients, and it works nearly 30 times faster than traditional methods.

Scientists at Memorial Sloan Kettering Cancer Center created an AI system that analyzes medical safety incidents with remarkable speed and accuracy. The tool reviews reports of errors and near-misses in just minutes, a process that previously took hours of careful human analysis.

Dr. Abbas Jinia and his team trained the AI on over 1,500 incident reports, teaching it to identify patterns and classify problems using methods borrowed from aviation safety. When tested on 350 real cases, the system matched expert reviewers' conclusions 88% of the time.

What makes this AI different is its transparency. Unlike many artificial intelligence tools that work like mysterious black boxes, this system shows its reasoning to human reviewers. Doctors and safety officers can ask questions and understand why the AI reached certain conclusions, building trust in the technology.

The system uses a framework called the Human Factors Analysis Classification System, originally developed to investigate airplane accidents. MSK adapted this approach to healthcare, recognizing that medical environments share similar complexity and high-stakes decision-making with aviation.

Medical errors remain a persistent challenge despite rigorous safety protocols. Learning from incidents quickly is essential to preventing future problems, but traditional review processes are labor-intensive and can take weeks. This bottleneck delays the implementation of safety improvements.

Why This Inspires

This breakthrough means hospitals can spend less time on paperwork and more time designing safer patient care systems. Healthcare teams who once spent hours categorizing incident reports can now focus on actually solving the problems those reports reveal.

The faster learning cycle could prevent countless medical errors across hospitals that adopt the technology. When safety teams can review incidents in minutes instead of hours, they can spot troubling patterns earlier and intervene before patients are harmed.

The research, published in npj Digital Medicine, represents a thoughtful integration of AI into healthcare that prioritizes human oversight and understanding. Rather than replacing human judgment, the tool amplifies it, allowing safety experts to work at unprecedented speed without sacrificing accuracy.

Medical institutions worldwide struggle with the same safety review challenges, making this innovation potentially transformative at a global scale.