The socioeconomic gradient in all-cause mortality for women with breast cancer: findings from the 1998 to 2006 National Cancer Data Base with follow-up through 2011.

PURPOSE: To analyze the association between socioeconomic status (SES) and all-cause mortality among women diagnosed with breast cancer before and after controlling for insurance status, race and ethnicity, stage, treatment modalities, and other demographic and hospital characteristics.

METHODS

Data analyzed included follow-up through 2011 for 582,396 patients diagnosed between 1998 and 2006 with ductal carcinoma in situ or invasive (stage I-IV) breast cancer from the National Cancer Data Base. SES was measured by grouping patients into six income and education-level ZIP code categories. Hierarchical Cox regression models were used to analyze SES survival differences.

RESULTS

Five- and 10-year survival probabilities for the highest SES group were 87.8% and 71.5%, versus 79.5% and 61.5% for the lowest SES group. Controlling for all covariates reduced the highest-to-lowest SES hazard ratio from 1.69 (95% confidence interval: 1.64-1.74) to 1.27 (95% confidence interval: 1.24-1.31). Results were virtually identical in models that included comorbidity and invasive cancer patients only.

CONCLUSIONS

Differences in insurance status, race, and stage at diagnosis are important components of SES disparities and explain about two-thirds of the initial SES survival disparity. The residual SES effect likely mirrors underlying social determinants of health for all American women.