Race/ethnicity and disparities in mastectomy practice in the Breast Cancer Care in Chicago study.

PURPOSE: To examine racial/ethnic disparities in mastectomy practice and explore mediating factors to explain the disparity.
METHODS: Participants included 989 females aged 30-79 years, from a population-based study of newly diagnosed (primary in situ/invasive) breast cancer patients, in Chicago, Illinois, from 2005 to 2008, who completed an interview. Medical records were also abstracted for tumor, diagnostic, and treatment information. Multivariable logistic regression models with model-based standardization were used to estimate risk differences. Differences in rescaled coefficients were used to estimate the proportion of the disparity that could be mediated by patient and tumor characteristics.
RESULTS: Mastectomy prevalence overall was 40 %. Factors significantly associated with increased rates of mastectomy (p < 0.05) included the following: non-Hispanic (nH) black and Hispanic race/ethnicity; younger age at diagnosis; lower socioeconomic status (SES); lack of recency of and adherence to screening mammography; and higher tumor pathologic stage and grade. In adjusted models (age, body mass index, comorbidity), compared to nH white patients, mastectomy was increased by 10 % points in both nH black (95 % confidence interval [CI] 0.03, 0.18; p = 0.007) and Hispanic (95 % CI 0.01, 0.19; p = 0.028) patients. After accounting for the proportion of disparity mediated by tumor stage, the disparity was reduced by about a third in nH black (risk difference = 0.07, 95 % CI -0.01, 0.14) and half in Hispanic patients (risk difference = 0.04, 95 % CI -0.05, 0.13). Additional control for mediation by SES and other tumor-related factors almost completely eliminated the nH black:nH white disparity.
CONCLUSIONS: The best approach to reducing the racial/ethnic disparity in mastectomy rates would be to intervene on factors that could affect stage at diagnosis.