Health economists have developed a robust and useful set of tools for presenting the results of economic evaluation to support health care decision making, including the incremental cost-effectiveness ratio, the efficiency frontier on the cost-effectiveness plane, and the cost-effectiveness acceptability curve (CEAC). These metrics and graphical analyses are produced to support a specific decision rule, which is that the decision-maker will provide the option with the greatest mean net benefit and reject all others.
While these metrics have been in general very successful in communicating the results of economic evaluation to decision makers, the underlying decision rule may not be appropriate to all settings, particularly those where several technologies are under appraisal.
There are some contexts, particularly in MTA, where regulators may not want to approve a single technology, but rather to encourage competition or to allow clinicians to exercise a degree of choice among several therapies. For example, standard decision rules take the prices of the technologies as given. Yet there are health industries where future prices of technologies may be influenced by the degree of competition in the market. In these circumstances, approving (or even encouraging) a single option may blunt market forces in the longer term, leading to higher prices and, possibly, less innovation.
In such contexts, health technology appraisal may be aimed more at providing high quality structured information to lower level decision makers (clinicians, local purchasers etc) than at making a definitive decision on their behalf. For example, regulators may wish to rank technologies from highest to lowest net benefit.
Appropriate metrics in these circumstances might include rank probabilities, cumulative rank probabilities and Surface Under the Cumulative Ranking. The paper shows how to calculate these measures from the results of Monte-Carlo simulations and argues that such metrics can complement traditional tools such as the cost-effectiveness acceptability curve. Rank probabilities are often used to show the results of network meta-analysis, but until now have not been used for economic evaluation. Furthermore, the paper goes on to show how presenting a wider range of rank metrics alongside the CEAC can improve communication between analysts and decision makers, for example, to intuitively explain why the option with the highest mean net benefit is not always the option most likely to have highest net benefit. The approach is illustrated using a case study.
