Background : Hematopoietic growth factors (HGFs) have been shown to reduce the incidence of neutropenia and fever in patients receiving cancer chemotherapy. Purpose : This cost analysis was designed to determine the conditions in which use of HGFs in patients receiving cancer chemotherapy is cost-effective. Methods : We used a standard model based on decision theory; the model assumes that all patients experiencing neutropenia and fever will be hospitalized and treated with intravenous antibiotics. Data from a prospective, randomized clinical trial of granulocyte colony-stimulating factor in small-cell lung cancer treated with combination chemotherapy were used to determine baseline probabilities for control hospitalization risk and survival; proportional hospitalization risk with prophylactic HGF; and median durations of hospitalization and prophylactic HGF use. The model was analyzed by one-way and multivariate sensitivity analyses, with estimation of threshold values at which the expected cost is the same for either of two treatment options. One or more of the specific costs and durations and the probability for each group of threshold curves were varied in a sensitivity analysis that generated variable thresholds. Use of Monte Carlo analysis based on the available distributions of the main variables provided 90% confidence limits and an inference method for comparing decision options. Results : The expected excess cost per treatment cycle, based on hospitalization for neutropenic fever and/or HGF administration, was $5500 for no HGF, $4750 for prophylactic HGF, and $6875 for therapeutic HGF. Sensitivity analysis provided the following thresholds for no HGF versus prophylactic HGF: control risk of hospitalization, 0.40; risk of hospitalization with HGF as a proportion of control, 0.64; total daily cost of hospitalization, $727; total daily cost of HGF, $344; duration of hospitalization, 7.3 days; and duration of HGF use, 11.0 days. Multivariate analysis revealed that conditions favoring the use of HGF on a cost basis become greater (a) as risk of hospitalization, total daily hospital cost, and duration of hospitalization increase and ( b ) as the proportional risk of hospitalization with HGF, daily cost of HGF, and duration of HGF treatment decrease. Conclusions : The major determinants of total excess cost were the control risk of hospitalization the proportional reduction in risk with HGF, and the average daily hospital cost. Implications : Use of HGFs should be based on the risk of hospitalization for neutropenic fever and consideration of the patient population and institutional costs. [J Natl Cancer Inst 85: 488–493, 1993]