OBJECTIVES: Clinical guidelines now are commonly used in many settings, though the impact of these guidelines on real world practice remains less clear. The USPSTF recently dropped recommendations for breast cancer screening among women aged 75+ years. Prior studies using survey data suggest that the guideline revision had no effect on screening behavior. Using registry and claims data, we assessed changes in screening patterns, and stage at diagnosis, among female Medicare beneficiaries aged 75+years compared with 65-74 years. METHODS: We examined 2003-11 biennial mammography screening rates using national Medicare fee-for-service claims (20% random sample), and linked SEER registry-Medicare claims databases. Using linear regression models with a patient-level fixed effect, we estimated the change in proportion of subjects receiving screening over time (as well as the rate of incident cases of earlier vs. later stage cancer), among subjects aged 75+yo compared with younger subjects, i.e., a difference-in-difference design. RESULTS: We find a decrease in screening rate among women aged 75+ years after the revision (2011 vs. 2003-2008, and the difference in changes compared with younger women was 13.2%  (95%CI:13.1-13.3%). Similarly, we find a relative decrease in the population rate of incident cases of earlier-stage cancer (0.12%, 95%CI: 0.07-0.17), increases in later-stage cancer. CONCLUSIONS: With the guideline revision, there were relative decreases in screening rates among women aged 75+ years, compared with younger women. There also was a breast cancer stage shift, with more later-stage cancers and fewer earlier-stage cancers detected. Future evaluations on survival are needed. Moreover, as health insurance plans increasing link benefit designs with guidelines, the likelihood that guidelines will impact practice patterns could grow.