OBJECTIVES: Previous cost-effectiveness analyses of targeted therapies in metastatic renal cell carcinoma (mRCC) have been based on randomised trials and evaluate just one single treatment-line. The aim of this study was to estimate the real-world cost-effectiveness of sequential first- and second-line treatments for patients with mRCC using a patient-level simulation (PLS) model. METHODS: Based on patient-level data from a Dutch population-based registry, a PLS model was developed that comprised entities (i.e. patients with mRCC), attributes assigned to the entities (i.e. prognostic factors), and events (i.e. second-line treatment or death). Patients were repeatedly simulated from the model and time-to-event was estimated using a lognormal distribution. A separate sampling process was used to determine which type of event occurred. Time to death following second-line treatment was modelled using a Weibull distribution. Lifetime healthcare costs were modelled using patient-level data from the registry. RESULTS: In current daily practice, 50% (341/686) of patients did not receive any targeted therapy and 42% (291/686) received sunitinib as first-line therapy. In the second line, 31% (33/107) were treated with sorafenib and 31% (33/107) with everolimus. Mean overall survival (OS) was 13.6 months and mean costs were €69,622 for all patients. In a strategy where all patients are treated according to clinical guidelines, mean OS was 15.2 months and costs were €91,059. This meant an increase in OS (1.6 months) and costs (€21,437) compared to current practice, with an incremental cost-effectiveness ratio of €159,107 per life-year gained. Probabilistic sensitivity analyses showed the robustness of these results. CONCLUSIONS: A complete disease model and real-world data are essential in estimating real-world cost-effectiveness. Our PLS model allows comparisons between treatment strategies spanning multiple treatment lines, which will ultimately help to reveal the optimal strategy. For example, guidelines-based treatment appears to increase both OS and costs compared to current daily practice.