Date of Award

Spring 2019

Document Type

Open Access Dissertation


Epidemiology and Biostatistics

First Advisor

Swann A. Adams


Introduction: Despite a lower incidence of breast cancer (BrCA) among Black women in the U.S. compared to White women, Black women experience consistently higher mortality rates. The aims of this study were 1) to assess the relationship between race and diagnosis-to-treatment times 2) to assess racial disparity in mortality among Black and White BrCA patients in SC and 3) assessment of the validity of the Mortality-to- incidence ration (MIR) as a proxy for survival and geospatial investigation of racial disparity among breast cancer (BrCA) patients.

Methods: Breast cancer cases diagnosed between 2002-2010 were obtained retrospectively from the SC Central Cancer Registry, linked with administrative data from a private payor source and Medicaid Plan. The main exposure variable for all analyses was patient’s race (White vs Black women). For aim 1, outcome variables were diagnosis-to-treatment time for BrCA-related surgery, radiation, adjuvant hormone treatment (AHT) and chemotherapy; Chi-square tests, logistic regression and generalized linear model analyses were conducted to compare patients’ diagnosis-to-treatment times among Blacks and Whites. For aim 2, the main outcome variable was mortality characterized by vital status and total survival time; Cox proportional hazard analyses were conducted to compare hazard ratios among Blacks and Whites to assess disparities in mortality. For aim 3, MIRs were computed from cancer incidence and mortality data which were obtained from the SC Community Access Network (SCAN). ArcGIS 10.2 was utilized to map BrCA MIRs by race (46 counties and 4 regions). MIR were categorized into seven levels using the national BrCA MIR for White women as reference in county maps; in all other maps, categorizations were based on natural breaks in ArcGIS. Survival percentage, Cox Proportional hazard ratios and survival-MIR correlation analyses were computed for all BrCA cases in each county/region utilizing SAS software and data on BrCA cases which were obtained retrospectively from the SC Central Cancer Registry from 2002 to 2010.

Results: A total of 2155 BrCA patients with 1557 White women and 598 Black women were reported in the study period (2002-2010). For aim 1, multivariable linear model regression showed that there was statistically significant increase in adjusted least square means in receipt of AHT by 54 days, 36 days, 63 days and 46 days among unmarried, not being on best chance network and late surgery; multivariable logistic regression showed that the odds of late receipt of surgery was 1.96 (95% CI: 1.38-2.79) among unmarried Black women compared with unmarried White women; 1.89 (95% CI: 1.32-2.71) among Black women who lived

Conclusions: Mortality was higher among Blacks who lived in the Low Country region of the state and among Blacks who lived in urban areas. Health region ranking utilizing the MIR correlated with 12-year survival time in the overall population, Whites and Black-White difference. To improve overall timely receipt of AHT, efforts need to be directed at Black BrCA patients that are not married, not on BCN, and those that received late surgery. To improve overall timely receipt of surgery, efforts need to be directed at Black BrCA patients that are not married, lived in urban areas and lived


© 2019, Oluwole Adeyemi Babatunde

Included in

Epidemiology Commons