Impact of Clinical/Demographic Characteristics and Duration of Clinical Benefit on Average Weekly Doses of Darbepoetin Alfa and Epoetin Alfa in Oncology: Results from a Claims Analysis.

Background: Comparison of darbepoetin alfa (DA) and epoetin alfa (EA) has often relied on calculation of unadjusted average weekly doses (AWD). This study examined the impact of addressing differences in clinical/demographic characteristics and duration of clinical benefit (DCB) of last dose administered on AWD.

Methods: MarketScan claims data were used to build EA and DA treatment episodes among cancer patients with >= 1 EA or DA claim and continuously enrolled January–June 2005. AWDs were calculated based on 2 different DCB scenarios (DCB1 and DCB2) for the last dose administered in the episode. DCB1 assumed DA doses <=100mcg =7 days, 101–299mcg =14 days, and >=300mcg =21 days; and EA doses <35,000units =2 days and >=35,000units =7 days. DCB2 assumed DA doses <60mcg =2 days, 60–149mcg =7 days, 150–299mcg =14 days, and >=300mcg =21 days; EA doses <=15,000units =2 days, 15,001–35,000units =5 days, and >35,000units =7 days. Logistic regression examined if clinical/demographic characteristics (age, gender, cancer type, comorbidities, and metastases) influenced likelihood of DA vs EA use; ANOVA models estimated an adjusted AWD controlling for these clinical/demographic differences.

Results: 2,942 DA and 2,574 EA episodes were created. Younger patients, with more comorbidities, and advanced cancer were found to have increased likelihood of DA use (all p<0.001). The DCB analysis revealed that EA was more sensitive to DCB changes than DA. Switching from DCB1 to DCB2 increased AWD for DA from 94mcg to 105mcg, while EA AWD decreased from 44,044units to 35,496units. With ANOVA adjustment to equalize the populations, estimated AWD for DCB1 and DCB2, respectively, were 97mcg and 104mcg for DA and 41,902units and 34,973units for EA.

Conclusions: Equalizing clinical/demographic characteristics and addressing possible DCB differences for DA and EA can substantially change the average weekly doses calculated from claims data; unadjusted AWDs may be misleading. In addition, comparisons of dose using claims data must be interpreted with caution because such data lack information on patient clinical outcomes of erythropoiesis stimulating agents and are subject to bias including data recording errors and confounding by indication.