To address controversies in the applications of cost-effectiveness analysis, we investigate the principles underlying the technique and discuss the implications for the evaluation of medical interventions. Using a standard von Neumann-Morgenstern utility framework, we show how a cost-effectiveness criterion can be derived to guide resource allocation decisions, and how it varies with age, gender, income level, and risk aversion. Although cost-effectiveness analysis can be a useful and powerful tool for resource allocation decisions, a uniform cost-effectiveness criterion that is applied to a heterogeneous population level is unlikely to yield Pareto-optimal resource allocations.