The Nonexistence of Interpersonal Utility Scales

Abstract

Utility scales, as elicited by the usual methods, are personal and cannot be averaged across individuals. Unfortunately, the Maximum Expected Utility principle calls for such averaging whenever medical decisions affect several patients--as they generally do because of budget constraints--and whenever a patient's scale is uncertain, such as for a comatose patient. Interpretation of therapeutic trials is particularly problematic when examined in this light, as are psychiatric decisions involving mental incompetence. To overcome this deficiency, a supplementary equal-right-to-treatment principle seems necessary, but the proposals examined here clash with the patient's right to choose his own utility scale for valuation of prospective treatment outcomes. Perhaps the basic assumption that personal suffering cannot be measured on an interpersonal scale is too radical, but the counterproposal involves too many assumptions to appear convincing. These issues have received remarkably little attention, if any, in the medical decision making literature.