Background: Self-report survey data show little difference in crude mammography rates by race. In contrast, Medicare claims and HEDIS mammography data show large disparities. We hypothesize that African Americans over-estimate their use of preventive care more than whites based on stereotype threat. Stereotype threat refers to apprehension by respondents that their behavior will reinforce stereotypes regarding their social group.
Aims : 1) To determine whether there are discrepancies between racial/ethnic disparities in preventive care derived from self-report measures and those derived from Medicare claims. 2) To examine potential explanations for possible over-reporting of preventive care by racial and ethnic minorities. 3) To examine potential explanations for possible under-representation of preventive care to minorities in Medicare claims. Methods: We use Medicare Current Beneficiary Survey to examine these aims. We compare estimates of racial disparity in mammography, sigmoidoscopy, and PSA testing based on self-report with estimates from corresponding Medicare claims. We systematically examine other explanations for discrepancy in disparity age, education, income and source of care by sequentially adjusting for these factors. Significance: The National Health Care Disparities Report will rely heavily on self-report to monitor disparities in health care. It is thus critical to determine whether reporting biases associated with race affect the validity of these estimates. Furthermore, erroneous conclusions regarding the absence of racial/ethnic disparities in preventive care may result in missed opportunities to reduce racial/ethnic disparities in health such as racial/ethnic disparities in breast, colon, or prostate cancer mortality.
