Cancer Research - Montefiore Einstein Center for Cancer Care - New York City

Abstract

Introduction:

Racial differences in cancer survival have been well established based on national and independent cancer registries. Multiple extrinsic factors including stage, socioeconomic status (SES) and treatment have been postulated as causative, questioning whether race is a significant independent predictor of cancer survival.  The purpose of this study was to evaluate racial differences in overall cancer survival rates in a diverse NY population.

Methods:

Study population included 18,556 patients (28% White, 31% African American (AA), 27% Hispanic) diagnosed and/or treated for primary cancer from 1/1/2000 through 1/1/2007 at Montefiore Medical Center (MMC), identified through MMC’s Cancer Registry.  Demographics, clinical information and survival were obtained through the Cancer Registry.  SES was determined using census tract data. Kaplan-Meier survival curves and Cox proportional hazards regression were used to compare overall survival by race.

Results:

White patients were older (67.6 vs. 60.9 & 58.7) and of higher SES than AA and Hispanic patients, with similar rates of stage III/IV cancer and surgical treatment. Unadjusted analysis revealed that Whites had a significantly higher risk of death compared to all other patients (HR 1.29, 95%CI 1.23-1.35). Adjusting for age, sex, stage at diagnosis, surgical treatment and SES in all cancer types combined, revealed a significantly higher risk of death compared to all other patients (HR 1.14, 95%CI 1.06-1.21). However, race was no longer a significant predictor of cancer survival in multivariate models stratified by cancer type.

Conclusion:

Contrary to published literature, White patients had significantly lower survival rates than AA and Hispanic patients in unadjusted models.  Racial differences in cancer survival diminished in adjusted models, and were absent when stratifying by cancer type.  These results suggest that racial disparities in cancer survival may be better explained by variables such as age, stage at presentation, SES, treatment.