11. Reversal of Trends in U.S. Breast Surgery Rates: An Analysis From 2005-2017 Using Three Nationwide Datasets

Abstract

Purpose: Despite equivalent oncologic survivorship, U.S. lumpectomy rates previously declined in favor of more aggressive surgical options such as mastectomy, often performed in conjunction with a contralateral prophylactic mastectomy (CPM) with or without reconstruction. Using three national datasets, this study evaluates longitudinal trends in lumpectomy/mastectomy, CPM, and breast reconstruction rates, determining characteristics most associated with current surgical practice. Methods: Trends in lumpectomy, mastectomy, and reconstruction rates were evaluated using the NSQIP, SEER, and NCDB databases from 2005-2017, further examining mastectomy with a focus on CPM. Longitudinal trends were analyzed with Cochran-Armitage Trend tests and Poisson regression. Upon determining a notable reversal in lumpectomy rates in 2013, we compared NCDB lumpectomy patients before (2011) and after (2017) this change. Multivariate logistic regression using NCDB identified predictors of lumpectomy, CPM, and reconstruction. Results: We analyzed 3,467,645 female surgical breast cancer patients. Surgical trends were found to be similar in all three databases. Lumpectomy rates reached a nadir between 2010-2013, with a significant increase thereafter (NSQIP: +1%/year; SEER +1.6%/year; NCDB: +1.6%/year, all p<0.001). Concurrently there was corresponding decrease in mastectomy rates. Both CPM and reconstruction rates increased significantly from 2005-2013 (p<0.001), but have since stabilized. Age distribution of lumpectomy patients from 2011 to 2017 demonstrated an increase in patients 60-79 years of age with a concurrent increase in the proportion of patients with Medicare (p<0.001). On multivariate logistic regression analysis, the strongest predictors of lumpectomy were older age, black race, treatment at a community center, and clinical N0 disease. The strongest predictors of CPM were younger age, white race, and residence in a zip code with a higher median income. The strongest predictors of reconstruction were younger age, white race, private insurance, residence in a zip code with a higher median income, and undergoing CPM. Conclusion: This is the first study to document a reversal of trend in lumpectomy rates since 2013 with an associated decline in mastectomies and stabilization in CPM and reconstruction. While the databases differ in size and population, the trends are consistent among all three databases. Longitudinal data demonstrate a reversal of prior trends which favored more aggressive surgical management of breast cancer. The etiology of the recent reversal in trends is likely multifactorial; however, an increase in age of the breast cancer population is likely related to this change in the trends. This is also the first evidence of level breast reconstruction rates since passage of the WHCRA. Further research is required to understand factors driving these recent practice changes and associated impact on patient reported outcomes.