Format

Send to

Choose Destination
Cancer Causes Control. 2008 Dec;19(10):1305-18. doi: 10.1007/s10552-008-9202-1. Epub 2008 Aug 14.

Race/ethnicity and breast cancer estrogen receptor status: impact of class, missing data, and modeling assumptions.

Author information

1
Department of Society, Human Development and Health, Harvard School of Public Health, Boston, MA 02115, USA. nkrieger@hsph.harvard.edu

Abstract

OBJECTIVE:

To test whether reported associations between race/ethnicity and breast cancer estrogen receptor (ER) status are inflated due to missing ER data, lack of socioeconomic data, and use of the odds ratio (OR) rather than the prevalence ratio (PR).

METHODS:

We geocoded and added census tract socioeconomic data to all cases of primary invasive breast cancer (n = 42,420) among women diagnosed between 1998 and 2002 in two California cancer registries (San Francisco Bay Area; Los Angeles County) and analyzed the data using log binomial regression.

RESULTS:

Adjusting for socioeconomic position and tumor characteristics, in models using the imputed data, reduced the PR for the black versus white excess risk of being ER--from 1.76 (95% CI: 1.66, 1.86; adjusted for age and catchment area) to 1.47 (95% CI: 1.38, 1.56). The latter parameter estimate was 16% greater (i.e., 1.56) in models excluding women with missing ER data, and was 43% greater when estimated using the OR (i.e., 1.82).

CONCLUSION(S):

Studies on race/ethnicity and ER status that fail to account for missing data and socioeconomic data and report the OR are likely to yield inflated estimates of racial/ethnic disparities in ER status.

PMID:
18704721
PMCID:
PMC2635927
DOI:
10.1007/s10552-008-9202-1
[Indexed for MEDLINE]
Free PMC Article

Supplemental Content

Full text links

Icon for Springer Icon for PubMed Central
Loading ...
Support Center