Disparities in Preoperative Imaging for Breast Cancer: Not So Black and White

6061 Background: Controversy exists about racial/ethnic differences in the care of breast cancer patients. Imaging plays a critical role in evaluation, but little national data exists on breast imaging by race. This study was performed to determine if imaging disparities exist in Medicare patients, when adjusting for socioeconomic (SE) factors. Methods: Surveillance Epidemiology and End Results (SEER) data linked to Medicare claims were reviewed for patients diagnosed with invasive nonmetastatic breast cancer between 2000 and 2005, undergoing surgery as their 1st treatment modality. Diagnostic imaging was reviewed during the interval between the 1st physician visit and surgery (“preoperative interval”[POI]) and imaging use was defined as ≥1 imaging-specific claim. SE factors included rural/urban setting, and education, poverty, and median income per census tract. Results: Among 39,790 patients, there were 34,774 White (87%), 2,341 Black (5.9%), 1,459 Hispanic (3.7%), and 1,216 Asian/Pacific Islander (3.1%) patients. Univariate racial differences were noted for all imaging types (p’s <0.0001 to 0.007). During the POI, when adjusting for demographics, tumor characteristics, and SE factors in multivariable analysis, use of mammogram, ultrasound (US), breast MRI, CT, and bone scan were not different between Blacks and Whites. Education level was a predictor for US (p=0.007), breast MRI (p<0.0001), CT (p=0.006) and bone scan use (p=0.002), but not mammography (p=0.90). There were varying correlations between Hispanics or Asians vs Whites for all imaging types. Stage, histology, U.S. region, and comorbidity index were most consistently predictive of imaging use of all types. When evaluating 6 mos prior to the 1st physician visit along with the POI, 96.4% of all patients had ≥1 mammogram, with 94.1% of Blacks having mammograms vs 96.5% of Whites (OR 0.7, 95% CI 0.6-0.9; p=0.001). Conclusions: There is little racial difference in imaging performed during the POI for Medicare patients having breast cancer, and efforts to determine causes of survival disparities in such patients should be focused elsewhere. Whether these findings apply to the privately insured or the uninsured requires additional exploration.