Objectives: It has long been suggested that, whereas the results of clinical studies of pharmaceuticals are generalizable from one jurisdiction to another, the results of economic evaluations are location dependent. There has been, however, little study of the causes of variation, whether differences in study results among countries are systematic, or whether they are important for decision making.
Methods: A literature search was conducted to identify economic evaluations of pharmaceuticals conducted in two or more European countries. The studies identified were then classified by methodological type and analyzed to assess their level of variability and to identify the main causes of variation. Assessments were also made of the extent to which differences in study results among countries were systematic and whether they would lead to a different decision, assuming a range of values of the threshold willingness-to-pay for a life-year or quality-adjusted life-year (QALY).
Results: In total 46 intercountry drug comparisons were identified, 29 in multicountry studies and 17 in comparable single country studies that were considered to be sufficiently similar in terms of methodology. The type of study (i.e., trial-based or modeling study) had some impact on variability, but the most important factor was the extent of variation across countries in effectiveness, resource use or unit costs, allowed by the researcher's chosen methodology. There were few systematic differences in study results among countries, so a decision maker in country B, on seeing a recent economic evaluation of a new drug in country A, would have little basis on which to predict whether the drug, if evaluated, would be more or less cost-effective in his or her country. Given the extent of variation in cost-effectiveness estimates among countries, the importance of this for decision making depends on decision makers' thresholds in willingness-to-pay for a QALY or life-year. If a cost-effectiveness threshold (i.e., willingness-to-pay) for a life-year or QALY of dollar 50,000 were assumed, the same conclusion regarding cost-effectiveness would be reached in most cases.
Conclusion: This review shows that cost-effectiveness results for pharmaceuticals vary from country to country in Western Europe and that these variations are not systematic. In addition, constraints imposed by analysts may reduce apparent variability in the estimates. The lessons for inferring generalizability are not straightforward, although the implications of variation for decision making depend critically on the cost-effectiveness thresholds applying in Western Europe.