OBJECTIVES: Survival outcomes from partitioned-survival models are typically associated with uncertainty, particularly in oncology where projections of long-term survival estimates heavily influence the cost effectiveness of products. However, no guidance exists on how best to address the uncertainty of survival estimates. We investigated alternative techniques for addressing survival uncertainty, along with recommendations for implementation into economic modelling in order to facilitate decision making. METHODS: A range of techniques used to report survival uncertainty were sourced from those applied in previous submissions to the National Institute for Health and Care Excellence (NICE) as well as original methods developed for this study. The methods for demonstrating uncertainty included one-way sensitivity analysis, probabilistic sensitivity analysis (PSA), and uncertainty ellipses. In addition, the probabilistic uncertainty surrounding survival parameters alone was also tested. These techniques were then applied to a simplified replication of the three-state partitioned-survival model used to inform the technology appraisal of pixantrone for third/fourth-line treatment of aggressive non-Hodgkin’s lymphoma (TA306). RESULTS: Variation of individual curve fit parameters in isolation demonstrated relatively large uncertainty on model results, but failed to clearly establish the uncertainty around the survival of patients due to the intrinsic correlation that exists between related parameters. Broader methods that explored “whole curve” uncertainty appeared to reduce overall uncertainty in model results (leaving them in line with clinical results), while also presenting easily communicable differences in likely survival. CONCLUSIONS: We recommend analysts use techniques that illustrate uncertainty in survival estimates to account for the correlation between relevant parameters, as this accurately reflects the uncertainty inherent in parameters. The methods that meet these criteria include the use of ‘whole curve’ techniques, PSA, and PSA of survival parameters in isolation.