|
Cost-benefit analysis of prophylactic granulocyte colony-stimulating factor during CHOP antineoplastic therapy for non-Hodgkin's lymphoma |
Dranitsaris G, Altmayer C, Quirt I |
|
|
Record Status This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn. Health technology Granulocyte colony-stimulating factor (G-CSF) during cyclophosphamide, doxorubicin, vincristine and prednisone (CHOP) therapy for non-Hodgkin's lymphoma.
Economic study type Cost-effectiveness analysis.
Study population Patients with non-Hodgkin's lymphoma.
Setting Hospital. The economic study was carried out in Toronto, Canada.
Dates to which data relate The data for the effectiveness study were obtained from a study published in 1995. The resource use data were obtained from information collected from 1985 to 1995. The prices used were from 1995.
Source of effectiveness data Effectiveness data were derived from a single study.
Link between effectiveness and cost data The costing was not based on the same patient sample as that used in the effectiveness analysis. The costing was undertaken retrospectively.
Study sample No power calculations were reported. The number of patients included in the study was 20, which represented a total of 50 antineoplastic therapy cycles with G-CSF and 48 cycles without it, with CHOP (or alternatingCHOP/cyclophosphamide, vincristine and prednisone (CVP)). The patient demographic and disease characteristics were reported as being similar to the authors' experience in their institution.
Study design The study design was not clearly reported. The study was conducted in a single centre and the follow up covered the study period (three and a half months, approximately).
Analysis of effectiveness It was not reported whether the analysis was based on intention to treat or on treatment completers only. The primary health outcomes used were as follows: incidence of febrile events, neutropenia and antineoplastic therapy delays. Presence of neutropenia was defined as an absolute neutrophil count (ANC) below 1 X 10?9 /L.
Effectiveness results In patients treated with G-CSF, the incidence of neutropenia was 12%, whereas in the control group the corresponding figure was 60% (absolute risk reductionof 48%). The number of cycle delays represented a proportion of 2% in the intervention and 45% in the comparator groups (absolute risk reduction of 43%). The incidence of febrile episodes was 2% in the intervention whereas it was 12% in the comparator (this result was reported to be "nonstatistically significant").
Clinical conclusions No statistically significant difference between G-CSE recipients and controls was found in the incidence of febrile episodes. The authors stated that "It is probable that the clinical trial may have come under the influence of a type II error".
Modelling The probabilities associated with risk of events (neutropenia) were used to calculate the expected costs. These included costs associated with testing, hospitalisation, travel costs and indirect costs. The model was used in order to make up for the lack of a prospective economic evaluation alongside the clinical trial.
Measure of benefits used in the economic analysis The measure of benefits used in the economic analysis was number of cycles needed (incremental benefits) in order to avoid one case of either neutropenia of fever. This measurement was calculated as the inverse of the clinical benefit (absolute risk reduction with intervention relative to comparator) reported in the effectiveness analysis section.
Direct costs The resource quantities and cost items were reported. The costs measured were as follows: operational costs and costs associated with complications. The boundary adopted was society. The estimation of unit cost for febrile neutropenia was based on actual data for 24 randomly selected patients from 85 cases with CHOP-induced febrile neutropenia treated in the authors' institution. Theacquisition cost of G-CSF and the outpatient dispensing fee were obtained from the outpatient pharmacy of the same institution. The cost of antibacterials, inclusive of personnel and supplies, as well as laboratory tests and drug monitoring costs, were obtained from statistics held at the corresponding departments of that institution. The daily cost of hospitalisation was obtained from the Ontario Hospital Association.1995 prices were used.
Statistical analysis of costs The mean cost of febrile neutropenia was reported with a 95% confidence interval (CI), which was used later in the sensitivity analysis.
Indirect Costs The quantity of resources use was reported (days off work). The costs measured were those associated with productivity losses. The boundary adopted was the patient. The estimation of quantities was based on the assumption that patients with ANC above 1.0 X 10?9 /L would be able to continue regular employment throughout a course of CHOP. Therefore for the 21-day period between antineoplastic drug administration, the patient will be able to work 14 days (once one day is eliminated due to the CHOP administration). This assumption was supplemented by actual data on the proportion of working patients from the 24 randomly selected (advanced-NHL) patients attending the authors' institution (from a database search for the years 1985-1995). The unit costs (wages) were obtained from Statistics Canada, 1995.
Currency Canadian dollars (Can$). The reported conversion rate was Can$1 = US$ 0.75, October 1995.
Sensitivity analysis The parameters assessed in the sensitivity analysis were G-CSF frequency (from a median number of doses of 11 to 9) and dosage (60% reduction from 5 to 2 micrograms/kg per day), G-CSF administration continuity (G-CSF given only on days 3, 5, 7, 9, 11 and 13, with 6 doses in each treatment cycle) and duration (G-CSF given for 14 days at 5 micrograms/kg per day). In addition, the 95% confidence interval (CI) for the cost of febrile neutropenia (extreme values) was incorporated in the analysis. One-way simple sensitivity analyses were performed.
Estimated benefits used in the economic analysis In patients treated with G-CSF, the incidence of neutropenia was 12%, whereas in the control group the corresponding figure was 60% (absolute risk reductionof 48%). The number of cycle delays represented a proportion of 2% in the intervention and 45% in the comparator groups (absolute risk reduction of 43%). The incidence of febrile episodes was 2% in the intervention whereas it was 12% in the comparator (this result was reported to be "nonstatistically significant").
Cost results The net societal cost for the intervention, relative to the comparator, was Can$1,275, at 1995 prices.
Synthesis of costs and benefits The costs and benefits were not combined. However, one can obtain estimates of (incremental) costs per additional incidence of neutropenia and/or number of cycle delay averted with intervention, at 1995 prices. These figures turned out to be Can$1,550 and Can$2,958, respectively. The sensitivity analysis showed that the 'frequency', 'dosage' and 'continuity' scenarios included in the sensitivity analysis resulted in net savings associated with intervention, relative to the comparator, of Can$1,349, Can$ 6,564 and Can$ 5,260, respectively. The upper and lower limits of the 95% CI of cost of neutropenia, when substitutedin the model, yielded net (positive) cost results of Can$227 and Can$2,287, respectively. These figures had associated cost-effectiveness ratios in terms of neutropenia and delay averted of Can$454 and Can$527.9, respectively for the upper limit, and Can$4,574 and Can$5,318.6, respectively for the lower limit.
Authors' conclusions Given the avoidance of lost productivity, the study results suggest that using G-CSF in patients receiving CHOP antineoplastic therapy is almost cost neutral. The use of G-CSF at 2 micrograms/kg per day or 5 micrograms/kg per day every second day is "economically attractive to society due to its cost saving implications".
CRD COMMENTARY - Selection of comparators The reason for the choice of the comparator is clear.
Validity of estimate of measure of benefit The estimate of measure of benefit may not be internally valid given the small sample size.
Validity of estimate of costs The resource quantities were reported separately from the prices. However the costing was undertaken on a different patient sample from that used in the effectiveness analysis. In addition, the cost savings associated with febrile episodes averted were included in the cost and final analyses, despite the fact that they had been found to be statistically non significant.
Other issues Theauthors' conclusions may not be justified, given the uncertainties in the data. The results are not generalisable to older, non-employed patients. Some comparisons were made with other studies in support of the study findings.
Bibliographic details Dranitsaris G, Altmayer C, Quirt I. Cost-benefit analysis of prophylactic granulocyte colony-stimulating factor during CHOP antineoplastic therapy for non-Hodgkin's lymphoma. PharmacoEconomics 1997; 11(6): 566-577 Indexing Status Subject indexing assigned by NLM MeSH Antineoplastic Combined Chemotherapy Protocols /adverse effects /economics /therapeutic use; Cost-Benefit Analysis; Cyclophosphamide /adverse effects /economics /therapeutic use; Doxorubicin /adverse effects /economics /therapeutic use; Female; Granulocyte Colony-Stimulating Factor /economics /therapeutic use; Humans; Lymphoma, Non-Hodgkin /complications /drug therapy /economics; Male; Middle Aged; Prednisolone /adverse effects /economics /therapeutic use; Recombinant Proteins; Vincristine /adverse effects /economics /therapeutic use AccessionNumber 21997008252 Date bibliographic record published 31/03/1999 Date abstract record published 31/03/1999 |
|
|
|