OUH study: AI could halve mammography reading workload without losing cancers
By Mohammad T. Elhakim and colleagues (Odense Universitetshospital, Region Syddanmark)Built and shared by Mohammad T. Elhakim and colleagues (Odense Universitetshospital, Region Syddanmark), not by me. Summarised from the public sources below. Check them before you rely on a number.
- Who
- Mohammad T. Elhakim and colleagues (Odense Universitetshospital, Region Syddanmark)
- Category
- Healthcare
- Source
- Web
- Dates
- Published 7 Mar 2025 · Found 17 Oct 2026
Researchers at Odense University Hospital tested AI on more than 249,000 screening mammograms from the Region of Southern Denmark. Replacing one of the two human readers with AI, or using AI for triage, cut the reading workload by about half while cancer detection held steady.
What they built
A retrospective study simulated three ways of using one commercial AI product in the Danish double-reading workflow: AI replacing the first reader, AI replacing the second reader, and AI as a triage tool that decides which exams need human reading. The results were compared with the actual radiologist readings.
The result
Replacing the first reader cut reading volume by 48.8% with the same cancer detection. Replacing the second reader cut workload by 48.7% and recalls by 2.2%, with a small drop in sensitivity (-1.5%). Triage gave the largest workload reduction, 49.7%, with slightly better detection.
Sources
Odense doctors test an AI scribe that writes the patient note during the visit
By Region Syddanmark (Odense Universitetshospital and Lokalpsykiatri Odense)
Region of Southern Denmark is testing an AI tool that listens to the conversation between doctor, patient and staff and drafts the journal note. A psychiatrist in Odense says it saves him three to four hours a week, while an emergency physician says the notes are not yet structured well enough to save him time.
Copenhagen-area emergency departments roll out AI that reads X-rays for fractures
By Region Hovedstaden / Region Østdanmark
The Capital Region and Region Zealand are rolling out an AI algorithm that analyses X-rays for suspected fractures in emergency departments and urgent care clinics. It lets nurse practitioners and junior doctors decide more often on their own whether a patient can go home.
TALT: Danish municipalities test AI that writes care documentation from speech
By Københavns, Aarhus og Aalborg Kommune with KL and Systematic
After two years of testing in municipal health and elderly care, the TALT project shows that AI can be a valuable aid when nurses and care staff write documentation. Copenhagen, Aarhus and Aalborg led the project with Systematic, but the solution needs a legal basis before it can be used in normal operations.
Want this in
your business?
Tell me which case caught your eye. I will tell you straight whether it fits, and what it would take.
Start a conversation →