OBJECTIVES: This study tested a method to measure the variability of data among countries, and to assess the generalisability of cost evaluation results. METHODS: The first step of the method consisted in identifying, within cost evaluations, all the factors potentially responsible for variability among locations. The second step consisted in selecting, among all potential transferability factors, the final transferability factors which generated variability, impacted on outcomes of economic evaluation, and were both measurable and distinguishable from other factors. The third step was the identification of transferability areas as sets of homogeneous final transferability factors. Both the Euclidean metric and Principal Components Analysis were used in the fourth step to explore the generalisability of the results. The method was applied to a high-cost innovation in oncology currently under development in five European countries, carbon ion radiotherapy. RESULTS: 138 potential factors identified, 116 final factors were analysed and regrouped into 9 areas. The Principal Components Analysis between countries 1 and 2, and between 4 and 5, showed proximity between the costs of equipment and the cost of buildings. A large variation was observed using the Euclidean metric between countries 1 and 2, especially for working time, and between countries 4 et 5 for the use of personnel resources. On the opposite, a low distance was observed between countries 2 and 4 for treatment capacity, and between countries 2 and 3 for technology availability and costs of personnel. CONCLUSION: Using this method we could assess the generalisability of the cost of carbon ion radiotherapy, and we identified factors and areas that limited this generalisability. This study also showed the necessity to integrate recommendations in order to increase the generalisability of economic evaluations in health care.