Drug Information Journal - March 2009 - (Page 185) PHARMACOECONOMICS 185 Sampling Uncertainty in the Cost-Effectiveness Evaluations and Guidance Given by the Existing Pharmacoeconomic Guidelines: A Question to Be Considered The objective of this study is to analyze the guidance given by the existing pharmacoeconomic guidelines up to the sampling uncertainty. The aim is to focus on the existing degree of vagueness of guidelines in order to facilitate the tasks of those who must base their choices on studies of this kind. After the grouping of methods in an appropriate way so that there would be a relative correspondence with the guidelines, a selection of 22 guidelines was made to study the precision given to sampling uncertainty and the preferred methods of handling this uncertainty. The methods of handling uncertainty have plainly evolved in recent years. However, the differentiations in the guidelines concerning this field are considerable. Consequently, the tasks of submitters, evaluators, and decision makers become very difficult. Vilelmine Carayanni, PhD Assistant Professor of Biometry, Highest Technological Educational Institute of Athens, Greece Key Words Pharmacoeconomic guidelines; Stochastic CE analyses; Sampling uncertainty Correspondence Address Vilelmine Carayanni, PhD, Department of Public Hygiene, Highest Technological Educational Institute of Athens, Saint Spyodonos St. 12210 Egaleo, Athens *Other methods for nonnormal data such as permutation and Monte Carlo tests that are taking their place in the group of bootstrap techniques, as well as nonparametric tests unsuitable for incremental CE measures, are not discussed in this article. INTRODUCTION Pharmacoeconomic analyses are being used increasingly as the basis for reimbursing the costs of new drugs. However, the analyses are complex to evaluate and little guidance is given to researchers on exactly how the assessment of the implications of uncertainty should be done. The problem is more serious for the stochastic evaluations, that is, evaluations exclusively using patient-level data. In the next section, the methods to handle the sampling uncertainty in cost effectiveness (CE) analysis are briefly discussed. In the section following that, a comparison is made among 22 pharmacoeconomic guidelines concerning the precision given to sampling uncertainty and the predicted methods of handling this uncertainty. A final section offers a discussion of the issues raised in the article. METHODS OF HANDLING S A M P L I N G U N C E RTA I N T Y The methods of handling sampling uncertainty are briefly presented in Table 1 and have been grouped as follows. First, the category of methods of confidence interval (CI) construction around incremental CE ratios (ICERs) concerns the cases when uncertainty covers the northeastern quadrant on the CE plane (one therapy is more effective and more costly than the alternative treatment). The normal theory methods comprise (a) methods based on the assumption of normality, including confidence box (1–3), Taylor’s theorem (4–7), confidence ellipse (4,7,8), and Fieller’s theorem (2,6,8–10); and (b) methods based on asymptotic normality, including Bonferroni’s method and the modified box (onesided Bonferroni test) (4,5). The methods used for nonnormal data* or when the differences in effectiveness are close to the origin include (a) resampling methods, such as nonparametric bootstrap and jackknife (4–1 Bayesian bootstrap (12), and parametric 1), bootstrap (13–14); and (b) transformations and angular transformations (1 1,15). In the majority of existing researches, Fieller’s and bootstrap methods outperform Taylor’s method, the confidence ellipse, and the box methods and produce reasonably accurate CI (2,8,9,16). In general, box methods and Bonferroni confidence intervals seem to present more disadvantages and perform less well. Second is the alternative decision method based on CE ratios: (a) net benefit (NB) apSubmitted for publication: January 14, 2008 Accepted for publication: September 9, 2008 Drug Information Journal, Vol. 43, pp. 1 85–194, 2009 • 0092-8615/2009 Printed in the USA. All rights reserved. Copyright © 2009 Drug Information Association, Inc.
For optimal viewing of this digital publication, please enable JavaScript and then refresh the page. If you would like to try to load the digital publication without using Flash Player detection, please click here.