To incorporate measures of risk and uncertainty (variance, skewness and kurtosis of the distribution of treatment outcomes) into standard cost/effectiveness measures of economic value for medical treatments with uncertain (stochastic) outcomes. We use Taylor Series methods to introduce variance, skewness and kurtosis into measures of expected utility, incorporating previously-quantified measures of risk attitude, including risk-aversion, prudence and temperance. This approach uses standard economic models of expected utility maximization, building upon and extending current best-practices for cost-effectiveness evaluations. If treatments have uncertain (stochastic) effects, standard cost/effectiveness models relying only on mean clinical outcomes can produce significantly biased estimates of true economic value. Either upward or downward bias can occur, depending on how the new treatment alters variance, skewness, and kurtosis of the distribution of treatment outcomes (vs. existing treatments). In general, lower variance and positive shifts in skewness of health outcomes increase expected utility, while increases in variance and negative shifts in skewness reduce expected utility. This helps explain prior empirical findings of a “value of hope” for patients presented with treatments with right-skewed but uncertain benefits. We graphically demonstrate the degree of potential errors in technology evaluations that omit these stochastic components using plausible changes in relevant statistical parameters and risk preference parameters from the economic literature. Clinical trials regularly assess variance to determine the precision of estimated treatment effects. Practitioners can implement our new method with (at minimum) a single new data-type from clinical trials – skewness in treatment outcome distributions. Additionally, assessing kurtosis in treatment outcome distributions adds further to our method’s precision. Omitting information about variance, skewness and (potentially) kurtosis can seriously bias medical technology assessments when treatment outcomes have uncertain (stochastic) outcomes. Clinical trials may need to enlarge sample sizes to measure skewness and kurtosis with appropriate precision.