Cost-Minimization Analysis of Once-Weekly Versus Thrice-Weekly Epoetin Alfa for Chemotherapy-Related Anemia

BACKGROUND: For individuals with chemotherapy-related anemia, the clinical effectiveness of epoetin alfa (EPO) dosed once weekly ([QW], 40,000 units per dose) has been demonstrated to be indistinguishable from that observed with thrice-weekly dosing ([TIW], 10,000 units per dose). Whether the advantage of less-frequent administration justifies the higher EPO dosage used in the weekly regimen in terms of overall cost of care is unknown. OBJECTIVES: To conduct a cost-minimization analysis comparing QW and TIW EPO dosing from a societal perspective. METHODS: Direct and indirect medical cost data were calculated for a 16-week period for 2 large, prospective, multicenter, community-based studies. Costs measured included EPO, transfusions, laboratory tests, office visits, and opportunity cost of patient time. RESULTS: The average total costs in 2002 (first half) dollars were nearly equivalent across the 2 groups (QW: $9,204; 95% confidence interval [CI], $9,057-$9,350. TIW: $9,265; 95% CI, $9,083-$9,447. P=0.60). QW incurred mean drug acquisition costs that were 23% higher (QW: $6,725; 95% CI, $6,611-$6,838. TIW: $5,474; 95% CI, $5,350-$5,598. Pless than0.001). However, QW patients can avoid the resource use and time cost associated with 2 additional office visits incurred each week (QW: $592 [$583-$600]; TIW: $1,709 [$1,678-$1,740]; Pless than0.001). Transfusion and laboratory test costs were slightly higher in the TIW group (QW: $1,888 [$1,837-$1,940]; TIW: $2,082 [$2,020-$2,144]; Pless than0.001). CONCLUSIONS: Total anemia treatment costs over a 16-week period with EPO QW were similar to those of TIW dosing. In the absence of cost differences between regimens, the noneconomic advantages of less-frequent dosing intervals should make weekly dosing increasingly attractive to patients, clinicians, and payers.

ss Methods

Data Sources
Two large, open-label, community-based clinical studies evaluating, respectively, TIW (n = 2,198) and QW (n = 2,856) EPO therapies were used as the primary data sources (Table 2). 4,6 The patients in each study were anemic (hemoglobin ≤11 g/dL) and received cisplatin-based or non-cisplatin-based chemotherapy, with or without radiation therapy. These patients presented with a variety of tumor types. The primary difference between the trials was the protocol for EPO administration.
In the TIW study, patients initially received a dose of 10,000 units 3 times per week. If a patient' s hemoglobin concentration did not increase by at least 1 g/dL by week 4, the dosage was increased to 20,000 units TIW. In the QW study, patients were initially given a dose of 40,000 units per week. If a patient did not experience an increase in hemoglobin concentration of 1 g/dL or greater by week 4, the dose was increased to 60,000 units per week. In both trials, treatment was discontinued if, after 8 weeks, change in hemoglobin concentration was less than 1 g/dL. The U.S. Food and Drug Administration-labeled doses for EPO are 150 units per kg TIW, with escalation at week 8 to 300 units per kg TIW in the event of nonresponse, and 40,000 units QW, with escalation at week 4 to 60,000 units QW in the event of nonresponse. The 40,000 QW regimen is commonly used. 11 Patients with at least 1 recorded dose of EPO, a nonmissing baseline hemoglobin reading of less than or equal to 11 g/dL as required by the protocol, and sufficient data to evaluate hemoglobin response (i.e., at least 1 hemoglobin reading after the baseline) were included in this analysis. Comparison of the end points of each trial demonstrates similar hemoglobin and quality-of-life responses for the TIW and QW EPO dosing schedules. 4,6 While the proportion of patients receiving transfusions differed between the 2 trials (31% of TIW patients received transfusions compared with 25% of QW patients), more patients in the TIW trial were transfused before the start of EPO therapy (29% of TIW patients compared with 22% of QW patients) 4,5 The proportions of patients with a dose escalation were similar between the 2 studies: 36.7% of TIW patients received a dose of 20,000 units or greater, and 33.4% of QW patients received a dose of 60,000 units or greater at any time. Response rates were also similar between the 2 trials, with 65.8% of TIW patients and 68% of QW patients experiencing a 2 g/dL rise in hemoglobin, or a hemoglobin level that is greater than or equal to 12 g/dL at some point in the study in the absence of a transfusion. 4,5 Patients in the 2 trials withdrew and experienced adverse events at similar rates (Tables 3 and 4).

Cost-Minimization Analysis
If 2 interventions document the same clinical outcome, a comparative economic evaluation can be conducted as a cost-minimization analysis. 19,20 In both of the EPO-effectiveness studies, resource utilization data were reported for EPO, transfusions, laboratory tests, office visits, and opportunity cost of patient-time resources over the period during which patients remained in the studies. Costs were approximated by applying cost estimates of specific procedures from published literature to the utilization observed for each patient in each trial. Because the use of EPO was well defined in time and limited to patients with cancer who were undergoing chemotherapy, intervention boundaries were easily identified. The relatively short duration of the treatment period makes discounting costs unnecessary (mean treatment durations appear in Table 2; "discounting" refers to the practice in financial analysis of adjusting future cash flows to reflect the time value of money). The assumptions that were made regarding cost and utilization for both studies are summarized in Table 5. When cost assumptions gathered from the literature (e.g., cost of transfusion, EPO concentration,

Cost-Minimization Analysis of Once-Weekly Versus Thrice-Weekly Epoetin Alfa for Chemotherapy-Related Anemia
Epoetin Alfa Treatment Guidelines  21 The cumulative dose was calculated as the sum of the mean weekly EPO dose observed over the study period. The costs of syringes, alcohol swabs, and other  4,6 Note: These are the observations with at least 1 recorded dose of epoetin alfa, a nonmissing baseline hemoglobin value less than or equal to 11.0 g/dL as required by the protocol, and data to evaluate hemoglobin response. medical supplies were not considered because they are included in the outpatient procedure costs and would thus be counted twice.

Statistics
Costs were calculated for each patient in the 2 studies. These patient-specific costs were averaged across patients within each sample, and the means were compared using standard t tests; 95% confidence intervals (CIs) around the point estimates of the mean costs were also constructed. For all tests, P was considered statistically significant when it was less than 0.05. Sensitivity analyses were conducted by varying the values of cost drivers and testing for statistically significant differences in TIW and QW mean total costs. For each cost driver, the percentage change that led to a statistically significant difference between the total costs of TIW and QW was also identified. All analyses were performed using SAS software version 8.00 (SAS Institute, Cary, North Carolina).

Sensitivity Analysis
The sensitivity analysis in Table 6 shows that the total cost of the QW regimen remains close to the total cost of the TIW regimen under a range of assumptions. Higher-than-average fees for office visits or for the opportunity cost of time raise the cost of the TIW regimen above that of the QW regimen and eventually lead to the QW regimen' s being statistically less costly. Very low values for physician fees or the opportunity cost of time yield the opposite results; the TIW regimen becomes statistically less costly. Ten-percent changes in the cost of EPO or in the cost of blood do not affect the statistical results. Table  6 also shows the percentage changes in cost of drugs, physician fees, transfusion costs, opportunity costs, and visits per injection that make QW administration statistically more costly. An EPO cost increase of 27% results in statistically significant higher total costs for QW patients.
ss Discussion EPO-dosed TIW has been demonstrated previously to be cost effective for anemic cancer patients who are receiving chemotherapy with or without radiation therapy. 10 The need for 3 office visits and 3 injections (intravenous infusions) each week, however, is often not feasible for some cancer patients,  caregivers, and the health care delivery systems. Now that a QW regimen has been shown to be equivalent in effectiveness to the TIW regimen, one would expect a switch to the equally effective yet substantially simpler regimen. 4,6 Because the starting dose of the QW regimen necessitates 33% more of EPO per week, the clinical and logistical advantages of the QW regimen should also be assessed from the economic perspective. Our principal finding shows that the use of 1 EPO administration per week does not require any statistically significant incremental expense from a societal perspective when compared with a TIW regimen. These cost trade-offs need to be assessed individually. Further extension of dosing regimens to reduce potential medical visits, however, must have natural limits as to the increases in drug costs that are acceptable to achieve these reductions and still remain cost effective.

Adverse Events Reported in at Least 2% of Patients
The results described in this analysis are based on the actual resource use incurred by patients while they remained enrolled in the clinical studies. One advantage of communitybased trial data is the availability of actual utilization and effectiveness measures obtained under standard (usual care) practices rather than the efficacy data determined from randomized controlled trials. 15,16 Data for patients who discontinued study participation early (i.e., prior to the sixteenth treatment week) were evaluated up to the point of the patients' withdrawal, even though patients may have continued to incur costs and benefits. Nevertheless, the relative overall costs of patient treatment in TIW and QW are informative and demonstrate that the overall costs of QW dosing are similar to TIW dosing.
Several potential cost categories, not considered here because of limited data, may further increase the economic advantages of less-frequent dosing regimens. For example, caregiver time, which has been demonstrated to be an important cost driver in cancer care, was not included, even though patients may require assistance to travel to their providers' offices. 22 Intangible costs, such as avoiding 2 additional injections every week (and avoiding the medical office altogether), have not been included and would make the QW regimen even more advantageous. Overall, however, the categories that were not considered are unlikely to significantly affect the relative costs reported in this study.
One potential implication of this analysis is that EPO regimens with administrations less frequent than once weekly may result in relatively small reductions in total treatment costs. Total time costs represented 18% of the total costs for the TIW regimen and 6% for the QW regimen. Assuming that patients make office visits only when an EPO dose is scheduled, a once-  ss Conclusion

Cost and Utilization Assumptions
The cost-minimization analysis was based on data from 2 large, open-label, community-based clinical studies, which allowed the use of actual cost measures that would be incurred under customary care circumstances rather than those generally observed in a randomized controlled trial. Thus, the results obtained are likely to more accurately reflect actual clinical practice. In conjunction with earlier work establishing the similarity of clinical effectiveness, this research suggests that once-weekly EPO dosing for anemic cancer patients is a costeffective alternative to traditional thrice-weekly EPO dosing, with the potential for improved convenience.