Using explicit decision rules to manage issues of justice, risk, and ethics in decision analysis: when is it not rational to maximize expected utility?

Concepts of justice, risk, and ethics can be merged with decision analysis by requiring the analyst to specify explicity a decision rule or sequence of rules. Decision rules are categorized by whether they consider: 1) aspects of outcome distributions beyond central tendencies; 2) probabilities as well as utilities of outcomes; and 3) means as well as ends. This formulation suggests that distribution-based decision rules could address both risk (for an individual) and justice (for the population). Rational choice under risk if choices are one-time only (vs. repeated events) or if one branch contains unlikely but disastrous outcomes might ignore probability information. Incorporating risk attitude into decision rules rather than utilities could facilitate use of multiattribute approaches to measuring outcomes. Certain ethical concerns could be addressed by prior specification of rules for allowing particular branches. Examples, including selection of polio vaccine strategies, are discussed, and theoretical and practical implications of a decision rule approach noted.