Presenter: Jeffrey H. Silber Affiliation: The University of Pennsylvania
Since the 1970s, breast cancer mortality rates in the U.S. have been higher among African-American women than among Caucasian women.
The factors contributing to this disparity are multi-factorial and not well understood.
Reducing racial disparities in breast cancer survival has been a federal priority since the early 1990's, and programs such as the Center for Disease Control's National Breast and Cervical Cancer Early Detection Program have been established to provide access to breast and cervical cancer screening services to underserved women in the U.S.
The goal of this study was to investigate the change in survival disparities over time using a novel statistical method called sequential multivariate matching.
They authors also investigated differences in demographics, presentation, and treatment as potential etiologies for the observed survival disparity.
The study included women over 65 years of age in the SEER-Medicare database diagnosed with breast cancer between 1991 and 2005. Twelve SEER sites selected because data from these sites were available for the entire study period.
The pre-Taxane era was defined as 1991-1998, and the post-Taxane area was 1999-2005.
There were 5,251 black patients and 72,695 white patients, with the majority of patients in both groups presenting with early-stage breast cancer.
The cohort of all black cases represented the focal group, or "control" group, for all subsequent matches.
White cases were matched to black cases in three distinct groups:
Matched on age and year of diagnosis (Demographics, 2 variables)
Matched on age, year of diagnosis, stage, and other presentation variables such as co-morbidities, estrogen receptor status, etc. (Presentation, 82 variables)
Matched on age, year of diagnosis, presentation variables, and treatment variables (Treatment, 126 variables)
5-year survival was compared between black controls, and white cases matched on demographics, presentation, and treatment, as well as between all cohorts in the pre- and post-Taxane eras.
When whites were matched to blacks on demographics, 5-year Kaplan-Meier survival was 69.2% vs. 56.7%, P < 0.0001.
Matching additionally on presentation (stage, estrogen receptor status, etc), 5-year Kaplan-Meier survival was 64.1% vs. 56.7%, P < 0.0001.
Matching further treatment data, the disparity was reduced to 61.6% vs. 56.7%, P < 0.0001.
The majority of the disparity was explained by differences in presentation, not treatment.
Comparing pre- and post-Taxane era 5-year survival, differences between whites and blacks matched for demographics were 12.4% in the pre-Taxane era vs. 12.5% in the post-Taxane era, which was not a significant difference (p = 0.389).
Of note, survival in both white and blacks increased in the post-Taxane era but disparities in survival between whites and blacks remained.
Although improvements in overall survival across races are observed in the post-Taxane era, racial disparities in breast cancer survival have not changed despite important policy initiatives.
Although differences in both presentation and treatment contribute to the observed disparity, differences in presentation account for a large proportion of the disparity.
Interventions should be aimed at improving risk factors at presentation.
This analysis emphasizes that in spite of advances in treatment of breast cancer, racial disparities in outcomes are ongoing.
Strengths of this study include the large number of variables included in the model, such as tumor specific characteristics and over 30 co-morbidities.
Limitations include the advanced age of patients included in the SEER-Medicare database, substantial missing staging data, and lack of information on molecular and biologic tumor characteristics.
Recent data suggest that black women with breast cancer may be more likely than white women to have a biologic variant of the disease characterized by earlier onset (peak age of 50) and more aggressive phenotype (Demicheli, Cancer 2007). Using Medicare data may exclude a substantial proportion of black women with this early-onset variant diagnosed before age 65.
Staging data were missing for 17% of black cases and 14% for white cases. Although "missingness" of data was matched for, there may be important differences contributing to missing data among whites vs. blacks. Furthermore, the difference in missing data between the two groups was statistically significant.
Estrogen receptor status was the only tumor-specific characteristic provided. As molecular characteristics become more widely used for treatment and prognostic stratification, these variables will become critical for modeling outcomes. These data are not currently available in the SEER-Medicare database.
In spite of these limitations, the findings of this study demonstrate that in spite of efforts to improve access to breast cancer screening, racial disparities in survival remain.
Further investigation into racial disparities is needed to better understand how disparities can be minimized.
Multilevel interventions targeted at patients, families, and provider teams, as well as state and national health policy measures are needed to reduce disparities in the future (Gorin, JNCI 2012).
Sep 2, 2014 - Several abstracts involving potential biomarkers of prognosis in cancer treatment were presented at a press briefing Nov. 18 at the American Association for Cancer Research -- National Cancer Institute -- European Organisation for Research and Treatment of Cancer International Conference, "Molecular Targets and Cancer Therapeutics," held from Nov. 15 to 19 in Boston.