Breast cancer care delivery is known to be uneven in the United States [1]. Clinicians and researchers have long observed differences in the process and outcomes of care for women diagnosed with breast cancer that can be attributed to observable individual and societal characteristics [2–6]. Researchers have reported noteworthy differences in screening, diagnosis, treatment, and mortality by age, race, ethnicity, geographic location, education, and income. It is important to recognize there are contradictory findings in the literature, specifically regarding the relationship between race/ethnicity and quality of chemotherapy care. The seminal Institute of Medicine report Unequal Treatment recognized these differences and the broader impact of disparities in health care delivery on the nation’s health [7]. The report’s authors noted the complexity in both understanding and remedying the current state. A persistent challenge to correcting disparities in breast cancer care delivery is the limited understanding of the reasons for the observed disparities [8]. Currently, we have a problem of uncertain size, scope, and etiology. Research studies that include triangulated data across sources, such as cancer registries, clinical registries, and surveys of patients and providers, will help disentangle current confusion and stimulate evidence-based quality improvement and policy efforts. In this context, the paper by Lipscomb et al. in this issue of Breast Cancer Research and Treatment is a necessary piece to solve this complex puzzle. A key contribution from this paper is that the findings stem from a unique combination of cancer registry data and clinical records. The authors used the Georgia Comprehensive Cancer registry to identify incident cases of breast cancer and abstracted records on 868 women treated in 27 oncology settings located in relatively rural Southwest Georgia. These linked data enable us to understand, at least in part, how several personal characteristics may influence variation in high-quality breast cancer care. The authors should be commended for their exhaustive attention to potentially confounding variables through sensitivity analyses and examination of interaction effects. Their careful attention to these issues increases the confidence in the principal study findings. In multivariable models, they found no differences in chemotherapy initiation by race. However, women who resided in impoverished areas were significantly less likely to initiate chemotherapy. This is an important finding that merits further attention and discussion. Of all the women in the study who initiated chemotherapy, 10% did not complete their originally intended plan, primarily due to toxicity. In a multivariable model that included interaction terms between race and marital status and race and comorbidity, unmarried black women had higher likelihood of chemotherapy completion compared with unmarried white women. Unmarried women with comorbid conditions were less likely to complete chemotherapy than unmarried women without comorbid conditions. Differences in chemotherapy initiation and completion were negligible for married women in the sample. One way to interpret these interesting findings is through the lens of social support. Marital status may serve as a proxy for the larger concept of social support, which may influence treatment decisions and outcomes. While social support was not This is an invited commentary to article doi: 10.1007/s10549-011-1916-1
Read full abstract
Dec 20, 2012
Mary Larkin 
Open Access
