OBJECTIVES: Temporary health states (states with a duration of less than one year) are common and include many infectious diseases, short-term treatments, and diagnostic procedures. Valuation of these states requires special consideration because the health state is transitory and standard methods ignore the influence of duration on preferences. Inaccurate assessments could introduce bias into cost-utility ratios. There is no “gold standard” valuation method. The aim of this study is to review and critique temporary health state valuation methods and identify areas for future research. METHODS: We reviewed the literature and evaluated preference-based temporary health state valuation methods according to five criteria: (1) Consistency with quality-adjusted life year theory; (2) Ease of use; (3) Relevance to temporary health state-specific domains; (4) Sensitivity to health state duration; and (5) Extent of bias. Our goal was to provide a critical assessment of methods that could be used to obtain values for use in cost-utility analyses. RESULTS: We identified six temporary health state valuation methods. Methods modified standard approaches by prorating utilities, using a chained approach, or trading-off waiting time or sleep instead of death in a time trade-off. These modifications capture the effect of duration better than standard methods. The strength of methods varied. No method was well tested for validity and reliability with respect to temporary health states. CONCLUSION: The literature on temporary health state valuation methods is sparse and inadequate. Our critique did not identify a method that is appropriate for valuation of all temporary health states. Selection of the most appropriate method should depend on the duration of and type of temporary health state being considered. Further research should focus on the validity, reliability and feasibility of valuation under different circumstances. Utility values obtained using temporary health state methods should be compared to those using standard methods to quantify biases.