In recent years it has become fashionable to make comparisons (in 'league tables' or rankings) between health care interventions in terms of their relative cost-effectiveness, in cost per life-year or cost per quality-adjusted life-year gained. However, concerns have been raised about the unthinking use of league tables and some authors have questioned the theoretical basis of their construction. In this paper a recently-reported league table is scrutinized and the important methodological features of the source studies identified. These include the choice of discount rate, the method of estimating utility values for health states, the range of costs and consequences considered and the choice of comparison programme. Several recommendations are made for improvements, both in the methodology of economic evaluation studies and in the construction and use of league tables. It is concluded that, for league tables to be useful, decision makers should be able to assess the relevance and reliability of the evidence in their own setting. Fuller reporting of methods and results by the authors of economic evaluation studies would greatly assist in the appropriate construction and use of league tables.