Racial/Ethnic Disparities of Erythropoiesis Stimulating Agent Use Among the Insured Poor: Pre- and Post-Safety Advisories

Racial/ethnic variations in utilization of cancer care services and outcomes are well documented with African Americans (AAs) almost uniformly shown to have lower utilization rates than Whites. This issue is especially important with respect to use of the supportive care agents, erythropoiesis stimulating agents (ESAs) during time-periods where ESAs were perceived as being associated with high benefit-risk ratios as well as with low benefit-risk ratios. We evaluated ESA utilization among Medicaid patients in South Carolina with anemia and lung cancer (LC), non-Hodgkins lymphoma (NHL) or breast cancer (BC). Three time-points were selected; 2002 nearly 10 years after epoetin alfa received an indication for anemia associated with chemotherapy and the first full year following the approval of darbepoetin, 2006 the year prior to ESA safety notifications and 2009 the most recent full year of data following the implementation of a black box warning.

The analysis utilizes South Carolina Medicaid fee-for-service paid claims. AA and White patients diagnosed with LC=162.xx; NHL=201.xx and 202.xx; or females with BC=174.xx plus anemia (ICD-9-CM=284.xx, 285.xx) were evaluated. Utilization rates were calculated annually as a percent of patients within each racial group with at least one prescription claim or procedure code for an ESA (i.e. epoetin/darbepoetin). Fisher's exact test was used to test for utilization rate differences.

The number of SC Medicaid patients diagnosed with anemia and one of the three cancers ranged from 442 to 655 annually. The average age of the study population in 2002 (55.6 years), 2006 (55.3 years) and 2009 (57.7 years) did not vary significantly between AAs and Whites. Among all cancer patients, ESA utilization increased from 2002 peaking in 2006 (prior to safety notifications and CMS National Coverage Decision). Following publication of the black box warning, utilization rates decreased 60% for AAs and 67% for Whites, in 2009. Initially (2002) utilization rates were 15% lower among AAs than Whites. However, by 2006 utilization rates were 13% higher in AAs compared to Whites. Following the black box warning, utilization rates were 30% greater in AAs versus Whites among all cancer patients (LC, BC or NHL). More specifically, ESA utilization rates were 47% greater among AAs compared to Whites with BC, and 280% greater among AA versus White patients with NHL. In contrast, ESA utilization rates were 9% lower in AAs compared to Whites with LC. (Table)

Following the approval of darbepoetin in 2002, utilization of ESAs were lower among AA patients with LC, BC or NHL compared to Whites. However, by 2006 when these agents were perceived to have high benefit-risk profiles utilization rates were greater among anemic AA versus White cancer patients. The utilization rates remained higher for AAs during 2009 when ESAs were perceived to have lower benefit-risk profile. Our findings indicate that disparities in the utilization of ESA changed overtime. Counter to results reported in prior studies of racial/ethnic disparities, among the insured poor in South Carolina in recent years, rates of ESA utilization were greater for AAs versus Whites. Additional analyses will evaluate if these differences were associated with differences in outcomes and toxicity.

Bennett:Pfizer: Consultancy.