Cost-Effectiveness of interferon alfa-2b added to chemotherapy for high-tumor-burden follicular non-Hodgkin's lymphoma.

UNLABELLED Recent data from GELF (Groupe d'Etude des Lymphomes Folliculaires) have shown that the addition of interferon alfa-2b (IFN) to a doxorubicin-containing regimen (CHVP: cyclophosphamide, doxorubicin, teniposide and prednisone) prolongs both progression-free survival and overall survival in high-tumor-burden follicular non-Hodgkin's lymphoma. This gain must be weighed against the incremental toxicity and cost of IFN over CHVP alone and the objective here was, to determine the marginal cost-effectiveness of additive IFN in the specific setting of high-tumor-burden follicular non-Hodgkin's lymphoma. Meta-analysis of GELF trial results employing a Markov model was used with three health states: No Progression, Progressive Disease, and Death. Treatment response, survival and toxicity data are drawn from the GELF study. The current study is based on the final analysis of 242 patients (J Clin Oncol 1998;16:2332-2338), with a six year median follow-up for overall survival (median overall survival: not reached for CHVP + IFN vs 5.6 years for CHVP Only, p = 0.008). MEASUREMENTS Quality of life data (utilities) are taken from studies with similar dosing of IFN, from Q-TwiST (quality adjusted time without symptoms or toxicity) analysis of the GELF data and from a panel of experts gathered to develop treatment models for high-tumor-burden follicular non-Hodgkin's lymphoma. Costs and quality-adjusted years of life saved were discounted at 3% per annum. SETTING Costs determined for university medical centers in the United States. Results showed that, at the median cohort age of 52, IFN add 9.9 quality-adjusted months at an added cost of $13,900 (marginal cost-effectiveness of $16,900 per quality-adjusted life year, or QALY). A more complex, two-stage model approximates the actual cohort survival curves much better than a simple, one-stage model, but both models yield essentially the same marginal cost-effectiveness. Sensitivity analysis to quality of life on IFN shows marginal cost-effectiveness ranging from $15,200/QALY (no penalty for IFN) to $21,300/QALY (20% quality adjustment, greater than that reported). The model is quite insensitive to the probability of IFN toxicity. The model is moderately sensitive to the efficacy of IFN in delaying progression, particularly in the first 18 months (pProgI), but the marginal cost-effectiveness does not rise to $50,000/QALY until pProgI increases 220% from the baseline. Although the model is moderately sensitive to the cost of IFN (cIFN), marginal cost-effectiveness is below $50,000/QALY for values of cIFN below $2580/month (baseline cIFN = $850/month, corresponding to a marginal cost-effectiveness of $16,900/QALY in the baseline case). If the model is modified to reflect the 14% overall survival advantage at five years found in trials utilizing more intensive initial chemotherapy (including the GELF trial), then the marginal cost-effectiveness drops to $11,900/QALY in the baseline case. In condusion, based on data from the GELF study, low-dose interferon alfa-2b is cost-effective when added to CHVP therapy in the treatment of high-tumor-burden follicular non-Hodgkin's lymphoma. The analysis is robust: the model employs very conservative assumptions, and additive IFN remains cost-effective over wide ranges of variables in sensitivity analyses. The marginal cost-effectiveness is best expressed as being in the range of $12,000/QALY to $17,000/QALY in the baseline case. A simple Markov model can be used to describe treatment regimens with distinct periods of therapy.