e17509 Background: As patient-centered care becomes more prominent, a better understanding of patient preferences and tradeoffs amongst treatment alternatives and outcomes is needed. This study used a discrete choice experiment to examine the preferences and willingness to pay for prophylactic G-CSF to decrease the incidence of chemotherapy (CT)-induced febrile neutropenia in breast cancer patients who previously received CT. Methods: An online survey was developed with 16 paired treatment choice scenarios comparing 3 alternative G-CSF options (11 versus 1 or 6 versus 1 injections per CT cycle) with a follow-up “no treatment” option. Each scenario had 4 attributes: risk of disruption to CT schedule due to neutropenia, risk of infection requiring hospitalization, frequency of G-CSF administration, and total out-of-pocket (OOP) cost for G-CSF during a CT cycle. Patients’ preferences and willingness to pay (as OOP cost) were estimated using logistic regression. Results: Patients’ (n = 296) preferred G-CSF options with the lowest OOP costs, the fewest injections, and improved outcomes (lowest risk of disruption to CT schedule and lowest risk of infection requiring hospitalization). In the context of this discrete choice experiment, OOP costs and risk of disruption to CT schedule were the most important attributes to patients; risk of infection requiring hospitalization and frequency of G-CSF administration affected patients’ choice of G-CSF option to a smaller but similar degree. Patients were willing to pay OOP $1,076 per cycle to reduce the risk of disrupting the CT schedule from high to low, $884 per cycle to reduce the risk of developing an infection requiring hospitalization from 24% (high) to 7% (low), and $851 and $667 per cycle to decrease the number of G-CSF injections per cycle from 11 to 1 and 6 to 1, respectively. Conclusions: With a current focus on patient-centered approaches in decision-making, physicians need to consider patient preferences when making decisions about therapy, including supportive care agents.