United StatesHealthcare and life sciences2024 – ongoingNorthwestern Medicine
In-house generative AI for radiology report drafts
Northwestern Medicine reported a 15.5% average efficiency gain from AI-assisted radiology reporting
Problem
Rising imaging volumes and a shortage of radiologists created reporting backlogs. Preparing reports took time that radiologists needed for interpreting scans.
Approach
Northwestern Medicine physicians and engineers built a generative AI system using clinical data from their network. It analyzed images and drafted reports for radiologists to review and finalize. An evaluation covered nearly 24,000 reports across an 11-hospital network over 5 months.
Results
Takeaway
Radiologists retained responsibility for the final interpretation. The 15.5% average gain was smaller than the gains individual clinicians described.
Categories
- Size
- Large enterprise
- Industry
- Healthcare and life sciences
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Generative AIComputer vision
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
Northwestern described an internally developed AI system that drafted radiology reports during clinical work. It reported productivity gains of up to 40% and said radiologists remained responsible for checking interpretations.
Northwestern Medicine reported a 15.5% average efficiency improvement across nearly 24,000 reports. It described radiologists reviewing AI drafts before finalizing reports.
Full details are in the original sources.