Changes in Use, Cost, and Value of Breast Cancer Screening Among Older Women in the US

The clinical and economic impact of breast cancer screening varies based on the modality, frequency, and age of the screened population.

To characterize changes in use and cost of breast cancer screening for older women.

Serial cross-sectional study using data from SEER-Medicare, 2009-2019.

Women 67 and older enrolled in Medicare fee-for-service.

Screening use and cost by age, frequency, and modality. We further categorized screening as cost-effective or cost-ineffective based on published economic analyses rather than guidelines. Cost-effective screening included biennial mammography among women < age 80, while cost-ineffective screening included annual mammography, addition of digital breast tomosynthesis (DBT), screening ultrasound, and any screening among women 80 and older. We estimated total annual spending on screening in Medicare fee-for-service, inflated to 2019 dollars.

Our sample included a mean of 229,683 (range 222,400- 244,793) Medicare beneficiaries annually. Biennial screening was stable among women 65-79, at 11.2% (95% CI 11.0-11.4) in 2009 and 11.9% (95% CI 11.7-12.0) in 2019. Annual screening was also stable at 32.5% (95% CI 32.3-32.7) in 2009 and 30.0% (95% CI 29.8-30.2) in 2019. Among women 80 and older, screening (annual or biennial) declined from 19% (95% CI 18.8-19.3) to 12.9% (95% CI 12.7-13.2). Between 2009-2019, use of DBT rose from 0% to 70.3% of screened women. Total spending on cost-effective screening rose from $569 million per year to $735 million per year, a 29% increase. Spending on cost-ineffective screening rose from $548 million to $1.025 billion, an 87% increase. By 2019, spending on cost-ineffective screening accounted for 58% of total spending.

Screening costs for older women have risen, driven by expenditures on technologies that may not be cost-effective. Reducing use of low value screening could result in savings that could be reallocated toward high value screening and follow up testing.