The Second Generation Acceptability Curve: A Novel Visualization Approach to Cost-Effectiveness Analysis

Abstract

The standard cost-effectiveness acceptability curve is a tool often employed in cost-effectiveness analysis. Such a curve is calculated by finding bootstrapped p-values for incremental net monetary benefits, and mapping them over a range of acceptable resources. With the recent development of the second generation p-value comes an opportunity to innovate upon the often used cost-effectiveness acceptability curve. The second generation acceptability curve is constructed by comparing bootstrapped incremental net monetary benefit confidence intervals to pre-specified null hypothesis intervals. This comparison creates second generation p-values over a range of acceptable resources. We describe and elaborate upon the process of constructing a second generation acceptability curve, as well as explore methods one could take to depict the curve. The means of interpreting and understanding the second generation acceptability curve are also discussed. Various sample studies are performed in order to explore the properties that a second generation acceptability curves possess, from changes in costs or effectiveness, to differences in other variables like null hypothesis interval length and sample size. The standard and second generation curves are compared in a couple examples. Each sample study is discussed and commented upon. The second generation curve is also employed in an applied example, following a paper that looks at how cancer data sets are handled by cost-effectiveness analysis methods, and how they manage in handling uncertainty and inconclusiveness. Ultimately, the second generation acceptability curve is shown to provide improvements to the standard cost-effectiveness acceptability curve in its ability to depict regions of approximate equivalence and inconclusiveness.