OBJECTIVES: The Dutch policy measure on expensive inpatient medicines aims to ensure accessibility by relieving financial burden of hospitals. After three years, outcomes research influences decision-making on the continuation of additional funding. We explored the consequences of daily clinical practice variation for real-world pharmacoeconomics of bortezomib in relapsed or refractory multiple myeloma. METHODS: Our study included 139 multiple myeloma patients who progressed from first line therapy and received bortezomib outside of an RCT. Detailed case reports were retrospectively collected from medical records in 38% of all Dutch hospitals. Treatment variation and combinations of bortezomib were explored until sixth line therapy. RESULTS: All patients had at least two treatment lines, 66% received third line, 41% fourth line, 14% fifth line and 6% sixth line therapy. At least nine chemical agents were given in all lines as mono-therapy or in different combinations. No specific treatment order could be identified because of large variation in regimes and drug usage in different and reversed order. Moreover, guidelines have changed over the years and recommend earlier use of bortezomib; and lenalidomide, another very effective medicine, was increasingly used. In total, 72 patients received bortezomib, 30% as mono-therapy and 70% as combination therapy. Bortezomib was most often combined with dexamethasone (60%), but combinations with five other drugs were seen. Clinical guidelines state that differences in patient circumstances require professional discretion; this results in extensive variability in daily practice. CONCLUSIONS: Compared to an RCT, outcomes research of bortezomib is complicated by extensive treatment variation in daily clinical practice. This suggests that a standard pharmacoeconomic model comparing two treatment arms is not sufficient. Comprehensive modelling using different data sources is required to acquire a valid and precise (cost-) effectiveness measure of bortezomib in daily clinical practice.