OBJECTIVES: To utilize data on a large population-based cohort of elderly non-Hodgkin’s lymphoma (NHL) patients treated with chemotherapy to measure the cost-effectiveness (as measured as cost per life-year saved) of white blood cell growth factor (CSF) use in a real-world setting METHODS: We identified 13,203 NHL patients from the SEER-Medicare database diagnosed from 1992 to 2002 who received chemotherapy within 12 months of diagnosis. Patients were followed from initial chemotherapy date until death or end of study period (October 31, 2006). Effectiveness of CSF use (primary and secondary prophylaxis) was measured as improved overall survival. Costs were estimated by summing reimbursement amounts derived from claims. Cost-effectiveness was estimated by modeling the joint influence of CSF use on costs and effectiveness using a propensity-score net monetary benefit approach. RESULTS: Primary prophylactic CSF use was cost-effective at lower willingness to pay thresholds, whereas at higher thresholds, not providing prophylactic CSF became the cost-effective strategy. For secondary prophylactic CSF use following neutropenia, fever, and/or infection, the opposite trend was observed. For low willingness to pay thresholds (less than $20,000 per life year gained), not administering CSF was the cost-effective strategy, while CSF use became cost-effective as willingness to pay increased (from $100,000+ per life year gained). CONCLUSIONS: To our knowledge this is the first population-based study to empirically measure the cost-effectiveness of CSF among cancer patients treated with chemotherapy. Results suggest that CSF use as primary or secondary prophylaxis may be cost-effective depending on society’s willingness to pay for improvements in outcomes.