Chicago-based Northwestern Medicine radiologists used a generative AI system — designed in-house by Northwestern engineers — to boost documentation efficiency by 15.5%, according to a study published June 5 in JAMA Network Open.
The system was utilized across Northwestern Medicine’s 12-hospital network on about 23,960 radiographs analyzed between Nov. 15, 2023, and April 24, 2024.
Here are five things to know from the study:
- The radiograph interpretations were split into two equal groups, half having been developed with AI-assistance.
AI-assisted interpretations took an average of 159.8 seconds while interpretations without AI-assistance took an average of 189.2 seconds.
AI-assisted interpretations were 29.4 seconds faster on average, resulting in a 15.5% increase in documentation efficiency. - Peer review found no difference in clinical accuracy or textual quality between the two interpretation groups.
- The AI model flagged studies that showed unexpected pneumothorax with a 72.7% sensitivity and 99.9% specificity.
- “For me and my colleagues, it’s not an exaggeration to say that it doubled our efficiency. It’s such a tremendous advantage and force multiplier,” Samir Abboud, MD, chief of emergency radiology at Northwestern Medicine, said in a June 5 news release from the health system.
- “Our study shows that building custom AI models is well within reach of a typical health system, without reliance on expensive and opaque third-party tools like ChatGPT,” Mozziyar Etemadi, MD, PhD, assistant professor of anesthesiology at Northwestern University Feinberg School of Medicine, said in the release. “We believe that this democratization of access to AI is the key to drive adoption worldwide.”
Read the full study here.
