🙏 Amazing team:
#AmberSalisbury (primary researcher)
@rich-norman.bsky.social
@juliasteinberg.bsky.social
#AmeliaSmit
#AnneCust
#CynthiaLow (consumer expert)
#KarenCanfell
#LouisaCollins
#MichaelCaruana

Thanks! Some of the main drivers of the cost effectiveness in this study were the higher rates of both recall and arbitration. We didn’t include any treatment costs or other impacts of overdiagnosis etc, but would be interesting next time!

Was a great session! Thanks for sharing some excellent insights.

Thanks! We also have an accompanying piece in @aunz.theconversation.com where you can read more...
bsky.app/profile/aunz...

Why AI might actually hurt breast cancer screening - you might be interested: @lisawilkinson1.bsky.social, @ariannahuff.bsky.social, @drkerrynphelps.bsky.social, @melissasweetdr.bsky.social, @oncodaily.bsky.social, @natashamay.bsky.social, @hollypayne.bsky.social, @brittanytrang.bsky.social

Huge thanks to my fantastic collaborators:
🌟Stacy Carter, U Wollongong
🌟Helen Frazer, St Vincent’s Melb
🌟Nehmat Houssami, @daffodilcentre.org
🌟Mary Macheras-Magias, BCNA
🌟Genevieve Webb, Health Consumers NSW
🌟Luke Marinovich, @daffodilcentre.org

And funders: Sydney Cancer Institute Seed Grant

📊We created an infographic summarising the results: www.alisonpearce.net/wp-content/u...

🔚 Conclusion: AI has promise, but we need to proceed cautiously. Women need reliable info on AI’s risks and benefits. Addressing public concerns is crucial for thoughtful & equitable AI integration in healthcare.

🗣️ My Take: Our research aligns with global trends – women see the benefits of AI but want it to support, not replace, clinicians. Current AI systems for breast cancer screening don’t have the evidence needed for trust. Rushing implementation could undermine screening programs.

🌟Key finding 3: What women want…

Overall, if AI is going to be used in breast cancer screening, they want it to be:
🎯 more accurate than radiologists,
⚡ faster to get results,
🦘 an Australian-owned algorithm,
🌏 representative of Australian women.

🌟Key finding 2: Impact on screening uptake 👩🏽‍⚕️

Most concerning: the way AI is implemented could change participation. Up to 22% might ⬇️their breast screening if they don't like how AI is being used.

This could lead to cancers being missed or diagnosed later meaning they are harder to treat.

🌟Key finding 1: Mixed feelings⚖️

Australian women have mixed feelings about using AI in breast cancer screening:
✅40% are open to AI if it’s more accurate than radiologists,
❌42% are opposed,
❓18% have concerns that would need to be addressed.