Strategies for Efficient Computation of the Expected Value of Partial Perfect Information

Abstract

Expected value of information methods evaluate the potential health benefits that can be obtained from conducting new research to reduce uncertainty in the parameters of a cost-effectiveness analysis model, hence reducing decision uncertainty. Expected value of partial perfect information (EVPPI) provides an upper limit to the health gains that can be obtained from conducting a new study on a subset of parameters in the cost-effectiveness analysis and can therefore be used as a sensitivity analysis to identify parameters that most contribute to decision uncertainty and to help guide decisions around which types of study are of most value to prioritize for funding. A common general approach is to use nested Monte Carlo simulation to obtain an estimate of EVPPI. This approach is computationally intensive, can lead to significant sampling bias if an inadequate number of inner samples are obtained, and incorrect results can be obtained if correlations between parameters are not dealt with appropriately. In this article, we set out a range of methods for estimating EVPPI that avoid the need for nested simulation: reparameterization of the net benefit function, Taylor series approximations, and restricted cubic spline estimation of conditional expectations. For each method, we set out the generalized functional form that net benefit must take for the method to be valid. By specifying this functional form, our methods are able to focus on components of the model in which approximation is required, avoiding the complexities involved in developing statistical approximations for the model as a whole. Our methods also allow for any correlations that might exist between model parameters. We illustrate the methods using an example of fluid resuscitation in African children with severe malaria.

Facet joint injections for people with persistent non-specific low back pain (Facet Injection Study): a feasibility study for a randomised controlled trial

“…Bayesian evidence synthesis methods will be used to derive distributions for model parameters based on all available information. The model will be used to estimate the expected value of sample information of a trial of the protocolised intervention and management plan, using algorithms developed by Madan et al 157 and Strong et al 158 The expected value of sample information calculations will be performed for alternative study designs to determine the optimal design in terms of factors such as sample size, length of follow-up and study population.…”

Estimating the Expected Value of Sample Information Using the Probabilistic Sensitivity Analysis Sample

“…Computationally cheaper single-loop approaches are sometimes available, but these rely either on the model being of a certain form 7,20,21 or on assumptions of Normality of the mean incremental net benefits. 1 Fast single-loop methods also exist for computing partial EVPI for single parameters, 22,23 but these have not yet been extended to the computation of EVSI.…”

Analytic approaches for research priority-setting: issues, challenges and the way forward

“…Although there exists a fair amount of literature looking into methodological aspects of analytic approaches -especially in relation to VoI [54][55][56][57][58][59][60] -less focus has been given on uncertainties and challenges that may be particularly pertinent to potential users of the methods. Such issues are discussed below.…”

“…Nonetheless, there is now a steady stream of research looking into ways of reducing the burden of undertaking VoI analyses. Such research has led to establishing situations in which VoI can be conducted with minimal modeling, as well as 'shortcuts' and efficient computational methods in calculating VoI measures [60,73,75,76,82,83]. This, coupled with explicit efforts to increase research capacity [84,85], imply that challenges related to time and expertise requirements are now less prohibiting than before.…”