Reversals of association for Pap, colorectal, and prostate cancer testing among Hispanic and non-Hispanic black women and men.

Background: Several studies have found that Hispanics and non-Hispanic blacks have statistically significantly higher adjusted OR for cancer screening tests compared to non-Hispanic whites, even though their crude percentages were lower than, or about equal to, those for the non-Hispanic whites. Most documentation is for mammography. This article investigates the prevalence of such unadjusted-to-adjusted "reversed associations" (RA) for Pap, colorectal, and prostate testing. We also investigate large percent changes (LPC) to the unadjusted ORs.

Methods: Data were from the 2004/2006/2008 Behavioral Risk Factor Surveillance System (BRFSS) and the 2000/2003/2005/2008 National Health Interview Survey (NHIS). Analyses used a consistent set of covariates.

Results: RAs were more common for non-Hispanic blacks than Hispanics, but Hispanics had a greater number of LPCs. RAs and LPCs occurred more often for Pap testing than colorectal and prostate testing. However, results from the BRFSS and NHIS were often not consistent.

Conclusions: Attention should be given to the National Breast and Cervical Cancer Early Detection Program, as well as public programs addressing other cancers, as possible contributors to RAs and LPCs. Hispanics may show more RAs in analyses of future data. Discrepancies between the BRFSS and the NHIS also must be recognized and explained.

Impact: This research highlights the need for vigilance regarding the results of analyses to identify race/ethnicity as a correlate of cancer screening. Results also direct attention to aspects of the results of multivariable analysis other than ORs and confidence intervals.