ABSTRACT
OBJECTIVES: the number of economic evaluations alongside multinational trials is increasing. Pooling of health effects of patients from different countries may be possible, but simply pooling costs does not give valid results and may not address national health policy issues properly. How should costs be handled in multinational economic evaluations? METHODS: a range of factors is discussed, that determine the (un)comparability of medical consumption and costs between countries; patient characteristics, epidemiology, patterns of medical practice, absolute and relative prices of medical services and cost estimation procedures. RESULTS: as an empirical example we investigated a clinical trial on breast cancer therapy in three European countries, focusing on the relationship between differences in medical practice and hospital medical consumption and costs. Correcting for differences in patient mix a multivariate analysis showed that differences in the number of laboratory procedures, day care treatments and outpatient visits between countries appear to be related to differences in medical practice. We calculated the costs of medical consumption as if all patients were treated in one country, correcting for these differences in medical practice. CONCLUSIONS: although the extent of the correction in this particular example was not very substantial, this approach seems a sensible way of analysis in order to investigate cost differences between countries in multinational trials. This provisional approach should be further developed and compared with alternative methods of correcting medical consumption and costs.