Economic evaluation in intensive care: The case of SDD

Abstract

The aim of this thesis was to examine the use of modelling techniques in the economic evaluation of selective decontamination of the digestive tract (SDD), used to prevent intensive care unit (ICU) acquired pneumonia. The need for evidence for the effectiveness and cost effectiveness of technologies used in intensive care was highlighted through an examination of the literature. The clinical and economic issues pertinent to ICU-acquired pneumonia and SDD were described. It was suggested that an economic evaluation of SDD was required. An evaluation using modelling techniques was proposed. A secondary economic evaluation of SDD was carried out, utilising a decision-analytic model and published clinical and economic evidence to derive cost/outcome ratios. This analysis showed that SDD could be a dominant therapy, but improved economic and long term outcome evidence was required to increase the robustness of conclusions. This thesis concentrated on improving the economic evidence. A national survey of SDD use provided information on clinical practice. A prospective observational study was carried out at two British ICUs to obtain evidence on the economic impact of ICU-acquired pneumonia. The impact of infection and confounding factors on resource use was handled quantitatively, using regression techniques. It was found that ICU-acquired pneumonia significantly increased length of ICU stay. These two sets of empirical data were used in a revised economic evaluation of SDD. SDD was found to be a dominant therapy at both centres. Uncertainty around cost/outcome ratios was considered to be decreased, or at least quantified, by this primary economic evidence. This thesis concludes that modelling has a place in economic evaluation in intensive care, if rigorous methods are used. It has also demonstrated that current, reliable and applicable economic evidence is a prerequisite to any economic evaluation, if it is to be included in the decision-making process.