Kaplan-Meier (KM) survival analysis forms a cornerstone of the oncology modeller’s toolkit. KM time-to-event data is often utilised by decision analysts to inform a range of oncology model inputs. These may include the probability of a patient dying, progressing or ceasing treatment over the course of a model’s time horizon (typically key determinants of the incremental QALY and cost associated with an intervention). As KM data is often incomplete, extrapolation beyond the period observed is generally required. Extrapolation of KM data in several recent NICE technology appraisals has moved away from the mechanical application of statistical functions on the basis of  ‘best-statistical fit’ as determined by Akaike and Bayesian Information Criterion  (a method used by manufacturers in a range of previous NICE appraisals) to the use of extrapolation informed by consideration of cumulative hazard plots. OBJECTIVES: To demonstrate the use of cumulative hazard plots, their interpretation and their ability to inform resource allocation and research decisions in health care. METHODS: A series of blinded case studies of oncology pharmaceutical interventions assessed in NICE’s STA program are presented. These case studies demonstrate the importance of appropriately considering cumulative hazard plots when assessing the cost-effectiveness of pharmaceutical technologies in oncology. RESULTS: Appropriate interpretation of cumulative hazard plots can 1) help inform rational, rather than mechanical, extrapolation of survival data; 2) can assist an analyst in explaining the discordance between statistical functions found to be the ‘best-fit’ by traditional test-based approaches and subsequent poor face validity when compared to observed data; and 3) can be utilised to generate hypotheses that may be tested in future clinical research (particularly the assessment of the effectiveness and cost-effectiveness of restricted/extended durations of treatment). CONCLUSIONS: Measured consideration of cumulative hazard plots from both RCT and observational data is essential to inform appropriate decision making in oncology.