Measuring uncertainty in economic evaluations : a case study in liver transplantation
It is important to account for all sources of uncertainty when evaluating the clinical or cost-effectiveness of health care technologies. Therefore, this thesis takes as its basis a cost-effectiveness study in liver transplantation and identifies two previously unexplored issues that can arise in clinical and cost-effectiveness studies. A literature review of studies evaluating the effectiveness, costs or cost-effectiveness of solid organ transplantation confirmed that these issues were important and relevant to other transplantation studies. The first issue concerns the selection of an appropriate method for estimating mean study costs in the presence of incomplete (censored) data. Twelve techniques were identified and their accuracy was compared across artificially created mechanisms and levels of censoring. Lin's method with known cost histories and short interval lengths is recommended for accurately estimating mean costs and their uncertainty. It is assumed that these findings are generalisable to any solid organ transplant study where censoring is an issue. The second issue explored in this thesis relates to methods for measuring uncertainty around survival, HRQL and cost estimates derived from prognostic models in the absence of observed data. Probabilistic sensitivity analysis is recommended for measuring prognostic model parameter uncertainty and estimating individual patient outcomes and their uncertainties, as it is able to incorporate the additional uncertainty from using prognostic models to estimate control group outcomes. This thesis shows the quantitative importance of these issues and the methodological guidance offered should enable decision makers to have more confidence in clinical and cost-effectiveness estimates. Providing decision makers with a fuller estimate of the uncertainty around clinical and cost effectiveness estimates will aid them in decisions about the necessity of conducting further research in to the clinical or cost-effectiveness of health care technologies.