Extent to Which County of Residence Matters in Predicting Breast Cancer Outcomes

In the US, the burden of breast cancer is not borne equitably among every woman. There are disparities in late-stage breast cancer risk, not receiving the recommended treatment protocol, and breast cancer survival that are related to race, ethnicity, urban/rural residence, and socioeconomic status (SES). The objective of this study is to assess the relative significance of individual-level factors and the latent county of residence effect on late-stage breast cancer risk, receiving radiation and surgery, and breast cancer survival. We also aim to evaluate the relative significance of latent urban/rural county designation and county-level SES effects on breast cancer outcomes. A causal-comparative study that used data from the Surveillance, Epidemiology, and End Results (SEER) Program Georgia (Atlanta and rural GA) registries between the years 1992 and 2007 for individual-level factors, breast cancer outcomes, and county of residence and the US Census for urban/rural designation and county-level SES. To estimate late-stage breast cancer risk and receiving radiation and surgery, we use the random-intercept logistic regression. For the survival analysis, we use the accelerated-time form parametric function with gamma shared frailty. We first model the latent county effect. In secondary analysis, we model the latent urban/rural assignment and county-level SES effects. Our results suggest no significant latent county effect on late-stage breast cancer risk; however, there were significant latent county effects on the likelihood of receiving radiation, surgery, and breast cancer survival. Patients belonging to counties with similar SES-levels were more highly correlated in breast cancer outcomes when compared to patients belonging to counties with similar urban/rural designation. The race of the individual was also significant for almost all outcomes with Blacks having worst outcomes as compared to Whites, followed in some cases by Asians. Ethnicity however, was not significantly associated with any of the outcomes. Younger women were found to have a higher risk for late-stage breast cancer while older women received less radiation, surgery, and had worse breast cancer survival. Factors affecting each level of the breast cancer continuum are distinct and should be examined separately. Our results show strong evidence for the importance of county of residence in predicting receipt of radiation, surgery, and breast cancer survival. Some of these effects are due to the county's SES-level, urban/rural assignment, or some other county-level characteristic. For these three breast cancer outcomes, the SES-level of the county had greater explanatory power in predicting the outcomes when compared to urban/rural assignment.