OBJECTIVES: NICE have previously recommended nilotinib, but not dasatinib, as a first-line (TA251) and second-line (TA241) treatment in chronic phase (CP) CML. Within these appraisals, different methods were used to estimate overall survival (OS). Bosutinib, a potent dual Src/Abl TKI, is undergoing a NICE appraisal in second-line or later CML (ID495). The objective is to review the OS methods used and assess their impact on cost-effectiveness and recommendations in CML appraisals. METHODS: We identify the methodologies used for estimating OS in TA241 and TA251 and investigate the impact of using these to estimate OS for the TKIs, including bosutinib, in relapsed/refractory CML. Finally, we consider the implications of the various methodologies on the cost-effectiveness results and recommendations in previous and future NICE assessments of TKIs for CML. RESULTS: The base-case in TA241 used a surrogate relationship between response (MCyR) and OS to derive OS estimates for nilotinib and dasatinib of 13.0 and 13.4 years respectively (second-line CP). Using the same approach for bosutinib gives an OS of 12.8 years (third-line CP). The base-case in TA251 used a cumulative approach, where OS is equal to the duration of treatments in the pathway. If this method is applied to the TKIs in second-line CML (TA241), OS is reduced to approximately 9.4 and 10.1 years for nilotinib and dasatinib respectively (second-line CP). Similarly, bosutinib OS (third-line CP) is also reduced using this method and a substantial increase in the ICER is seen. CONCLUSIONS: There are methodological inconsistencies in NICE’s assessments of TKIs for CML. Applying the OS methodology from TA251 to TA241 may have led to nilotinib not being recommended for routine use in the NHS. The impact of new methodologies on previous appraisal results and recommendations should be considered when assessing the validity of a new approach.