
Variations among women in breast cancer screening are due to provider and health system factors rather than patient characteristics, according to a study published online August 3 in the Journal of General Internal Medicine.
Because breast cancer screening uptake can be influenced by factors associated with patients, primary care providers, practices, and health systems, a team led by Tracy Onega, PhD, of Dartmouth College sought to clarify the effects of these factors on breast cancer screening.
The study consisted of a web-based survey distributed at 15 primary care practices that represented 306 providers and served 46,944 women. The survey covered patient visits between January 2011 and September 2014.
Overall, 73.1% of women underwent breast cancer screening. Patient ethnicity and the number of primary care visits had strong associations with screening rates, the group found.
However, after adjusting for patient-level characteristics, Onega and colleagues discovered that 24% of the overall breast cancer screening variation among primary care providers was due to practice factors, 35% to health system factors, and 41% to variations among providers within practices.
"After accounting for patient characteristics, the variation in breast cancer screening metrics was largely due to provider and health system variation rather than specific structural characteristics and process measures," the researchers concluded.
![A normal mammogram confirmed by three-year radiologic follow-up illustrates reader-marked regions of interest (ROIs) during (A) unaided (round 1) and (B) artificial intelligence (AI)–assisted (round 2) reading. Each colored dot represents an ROI for recall by a human reader. Readers could mark more than one ROI per case, represented by multiple dots of the same color. During AI-assisted reading, the AI system displayed three visible prompts: two with suspicion of malignancy scores of 35% (left mediolateral oblique [L MLO] and craniocaudal [L CC]) and one with a suspicion of malignancy score of 10% (right craniocaudal [R CC]), shown as polygonal overlays. Without AI, six of 10 readers (60%) marked a false-positive ROI. With AI assistance, this fell to two of 10 (20%). R MLO = right mediolateral oblique.](https://img.auntminnie.com/mindful/smg/workspaces/default/uploads/2026/07/2026-07-14-radiology-mammogram-ai-auto-bias.H0bYO8QlWs.jpg?auto=format%2Ccompress&fit=crop&h=100&q=70&w=100)







![A normal mammogram confirmed by three-year radiologic follow-up illustrates reader-marked regions of interest (ROIs) during (A) unaided (round 1) and (B) artificial intelligence (AI)–assisted (round 2) reading. Each colored dot represents an ROI for recall by a human reader. Readers could mark more than one ROI per case, represented by multiple dots of the same color. During AI-assisted reading, the AI system displayed three visible prompts: two with suspicion of malignancy scores of 35% (left mediolateral oblique [L MLO] and craniocaudal [L CC]) and one with a suspicion of malignancy score of 10% (right craniocaudal [R CC]), shown as polygonal overlays. Without AI, six of 10 readers (60%) marked a false-positive ROI. With AI assistance, this fell to two of 10 (20%). R MLO = right mediolateral oblique.](https://img.auntminnie.com/mindful/smg/workspaces/default/uploads/2026/07/2026-07-14-radiology-mammogram-ai-auto-bias.H0bYO8QlWs.jpg?auto=format%2Ccompress&fit=crop&h=112&q=70&w=112)










