Abstract
BACKGROUND: Although recommendations for breast cancer follow-up frequency exist, current follow-up guidelines are standardized, without consideration of individual patient characteristics. Some studies suggest oncologists are using these characteristics to tailor follow-up recommendations, but it is unclear how this is translating into practice. The objective of this study was to examine current patterns of oncologist breast cancer follow-up and determine the association between patient and tumor characteristics and follow-up frequency. METHODS: The Surveillance, Epidemiology, and End Results (SEER)-Medicare database was used to identify stage I-III breast cancer patients diagnosed 2000-2007 (n = 39,241). Oncologist follow-up visits were defined using Medicare specialty provider codes and the linked AMA Masterfile. Multinomial logistic regression determined the association between patient and tumor characteristics and oncologist follow-up visit frequency. RESULTS: Younger age (p < 0.001), positive nodes (p < 0.001), estrogen receptor/progesterone receptor positivity (p < 0.001), and increasing treatment intensity (p < 0.001) were most strongly associated with more frequent follow-up. However, after accounting for these characteristics, significant variation in follow-up frequency was observed. In addition to patient factors, the number and types of oncologists involved in follow-up were associated with follow-up frequency (p < 0.001). Types of oncologists providing follow-up varied, with medical oncologists the sole providers of follow-up for 19-51 % of breast cancer survivors. Overall, 58 % of patients received surgical oncology, and 51 % undergoing radiation received radiation oncology follow-up, usually in combination with medical oncology. CONCLUSIONS: Significant variation in breast cancer follow-up frequency exists. Developing follow-up guidelines tailored for patient, tumor, and treatment characteristics while also providing guidance on who should provide follow-up has the potential to increase clinical efficiency.