Additive Multicriteria Decision Analysis Models Are Not Misleading Aids for Life-Critical Shared Decision Making

Abstract

The recent article by Kujawski, Triantaphyllou, and Yanase presents a detailed theoretical discussion of multiattribute value theory and a critical analysis of a single multicriteria decision analysis (MCDA) model that fails to meet a ‘‘reasonableness test’’ defined as ‘‘can a treatment that causes premature death trump a treatment that causes acceptable adverse effects?’’ The conclusion is that additive MCDA models are not suitable for use in shared decision making. The argument presented in the article does not justify this conclusion. The adequacy of a MCDA model depends on the criteria included, the alternatives considered, and how well it provides useful insights for the involved decision makers. It is impossible to conclude from an analysis of a single model that all MCDA models are inadequate. Moreover, the ‘‘faulty MCDA’’ model reviewed was not even created to guide shared decision making but rather to illustrate approaches to manage uncertainty within the MCDA framework. For this reason, it was deliberately simplified. I do not understand why the authors chose this particular model to serve as an exemplar of MCDA models for use in clinical decision making; clearly, it was never intended to be. A second major problem with the article is that it ignores the fact that decisions are made under conditions of uncertainty. When the outcomes of a decision are uncertain, a good decision can result in a poor outcome and vice versa. For this reason, the best measure of the quality of a decision is how it was made, not the decision outcome, as implied by the authors’ ‘‘reasonable test.’’ MCDA meets all of the criteria that define a quality decision making process in a framework that can be readily used by practitioners and patients. There is no evidence presented in the article that the linear qualityadjusted life-year (QALY) model does so also. It is unfortunate that the authors chose to present an article primarily intended to advocate a particular approach to decision making rather than one that presents a careful discussion of the pros and cons of both MCDA and the linear QALY framework. The latter approach could have resulted in a set of well-reasoned and constructive suggestions regarding how both methods could be integrated into a combined format that captures the strengths and overcomes the limitations of the individual methods.