Our society is continually faced with difficult decisions when allocating resources and containing costs. Cost-effectiveness analysis (CEA) is an economic study that compares the relative expenditures and outcomes of multiple strategies for performing the same task. The use of CEA in various fields, including economic, social, biomedical, and public health sciences, has been popular for many years. [43,895 articles found in PubMed using """"""""cost-effectiveness analysis""""""""]. Statistical methods for CEA have been extensively developed, and some measures have been widely adopted. However, it is clear that CEA remains controversial despite its long history of use and the efforts that have been made to understand various aspects of it. [For example, more than 20 methodology papers have been published on one statistical problem: deriving or comparing confidence intervals for the incremental cost-effectiveness ratio. This level of attention and publication is highly unusual in other statistical problems/settings.] Moreover, many journals have written publication policies on CEA. This is primarily due to the fact that: 1) different methods for CEA can yield different, often counterintuitive or inconsistent results;2) there is uncertainty and difficulty in interpreting the analysis that was performed;3) there is a lack of consensus on methods;and 4) there is a huge impact that CEA may have on decision making. As such, most of the existing analyses require great caution and care before, during and even after the analysis. We claim that the current standard approach for CEA is suboptimal and can be problematic, as it is based on only one analytical perspective and does not account for some important methodological issues. In this proposal, we intend to address these aspects of CEA and to develop methodologies and computer software in a unified framework. We wish to stress that multiple analytic methods should be chosen and evaluated together before one makes a conclusive statement about the cost-effectiveness of different strategies. The proposed methods do not compete with one another but are complementary, because they fulfill different tasks and would collectively provide a more comprehensive perspective for economic evaluations.
