Women’s Attitudes Towards AI in Breast Screening

I have to confess to having struggled to write this blog.  Why?  Because it’s difficult for me to understand why someone wouldn’t want a piece of consistently performing software code to remove the variation, call it luck, from the screening, diagnosis, and treatment process for breast cancer.  Especially if that artificial intelligence (AI) based code had demonstrated its effectiveness and safety in clinical trials and was being utilized by a licensed provider.  However, I had to stop and realize that I work in the world of healthcare AI everyday which provides a unique perspective into its advantages and opportunities.  

Fortunately, several studies have confirmed that patients are receptive to AI in the breast screening process.  The following highlights a few of the studies and their findings:  

• The Evaluation of EMRAD AI in Breast Screening Project: Final Report measured women’s attitudes towards the use of AI in the breast screening program resulted in the key finding that adult women of and under screening age were generally positive about the introduction of AI into the breast screening service. 
• A study conducted this year by BMJ Health & Care Informatics entitled, Women’s attitudes to the use of AI image readers: A case study from a national breast screening programme found that women of screening age are “ready to accept the use of AI in breast screening.” 
• In The Future of Breast Cancer Screening: What do participants in a breast cancer screening program think about artificial intelligence? researchers validated “a high level of trust in computerized decision-making.”  

While research has validated that women are open to their providers utilizing AI within the screening process, more work is needed to ensure that women understand the screening process and how to effectively engage with their providers around the screening results, diagnosis, and treatment planning.   

Additional opportunities include:  

• Creating greater transparency around the generalizability of AI algorithm performance.  This includes the accuracy of the model in detecting breast cancer within specific ethnicities or breast densities and/or variation in model performance on specific equipment types.
• Extending the power of AI beyond detection and into treatment planning and delivery.  

The use of AI within the breast screening, diagnosis, and treatment process holds great promise.  In the near term, we must provide easy to understand information to individuals in the screening process that help them understand the role of AI and how its information is being utilized to improve their health and, ultimately, outcomes.