OBJECTIVES: Relative dose intensity (RDI) is generally defined as a patient’s actual received chemotherapy dose over time period versus the intended dose over time where RDI<85% can be considered sub-optimal. Published cost-effectiveness analyses of febrile neutropenia (FN) prophylaxis modeled RDI assuming a fixed percentage of patients achieve low RDI. The objective of this study was to model the reasons for low RDI explicitly and evaluate the effects on results. METHODS: A lifetime Markov model of FN prophylaxis in the treatment of stage II breast cancer using TAC (FN risk ≥20%) and AC-T (FN risk <20%) chemotherapy was developed. Three FN prophylaxis strategies were compared: primary prophylaxis with pegfilgrastim (PPP), or with 6-day filgrastim (6D-PPF), and no prophylaxis (NP). RDI is modeled under two different scenarios: “fixed rate” where fixed proportion of patients receives low RDI and “flexible rate” where low RDI is function of chemotherapy-specific discontinuation and dose delay/reduction due to FN or other adverse events.  Total costs and quality adjusted life years (QALYS) are accumulated in the model and results are reported in terms of incremental cost-effectiveness ratios (ICERs). All model parameters are estimated from published literature or expert opinion. RESULTS: In the TAC fixed rate scenario, results demonstrate that PPP dominates (less costly and more effective) both 6D-PPF and NP. In TAC flexible scenario, PPP also dominates both 6D-PPF and NP.  In AC-T there is more variation in results; In the fixed rate scenario, PPP weakly dominates 6D-PPF, and the ICER versus NP is $106,700 per QALY gained, in the AC-T flexible scenario the ICER is $54,800 per QALY gained versus NP. CONCLUSIONS: Explicitly modeling discontinuations and dose delay/reductions can substantially affect model ICERS and subsequent willingess-to-pay interpretation, especially in lower-risk regimens. Furthermore, explicit modeling permits the flexibility to examine the effects of each factor alone.