Presentation (CTI)

OBJECTIVES: The Systemic Anti-Cancer Therapy (SACT) dataset contains real-world evidence (RWE) reported by National Health Service (NHS) trusts in England. Using data from SACT, the aim of this study was to compare estimates of cost-effectiveness and severity classification for the National Institute for Health and Care Excellence (NICE) severity modifier using data accepted by the NICE appraisal committee versus RWE. This was explored through a case study of Technology Appraisal (TA) 975 (tisagenlecleucel for paediatric and young adults with relapsed or refractory B-cell acute lymphoblastic leukaemia).
METHODS: Kaplan-Meier overall survival (OS) and event-free survival (EFS) curves for tisagenlecleucel (intervention) and blinatumomab (comparator) were digitised. Scenario 1 used single arm studies from TA975; scenario 2 used RWE from SACT (OS only). Mixture-cure models were fitted to reconstructed pseudo individual patient data (IPD) and incorporated into a reconstructed partition survival model using naïve indirect comparisons. Incremental cost-effectiveness ratios (ICERs) and severity modifier weights were estimated for both scenarios. Uncertainty was explored through probabilistic sensitivity analysis and scenario analysis.
RESULTS: Compared with single arm studies, RWE suggested improved OS in both arms with a relatively larger improvement for the comparator. This reduced the modelled relative survival benefit of the intervention, narrowing estimated incremental QALYs and costs. The ICER increased by over 135%, from £31,257 (scenario 1) to £74,102 (scenario 2) per QALY gained, and the severity modifier reduced from 1.7 to 1.2.
CONCLUSIONS: Incorporating RWE for relative effects can substantially impact cost-effectiveness estimates and may affect adoption decisions. However, aggregate SACT data limited the ability to identify or adjust for confounding, and whether SACT data fully reflected the target population remains unclear. Decision makers should assess the suitability of SACT data for relative effects case-by-case, but presenting trial- and RWE-based cost-effectiveness estimates could reduce decision uncertainty. Further research should use IPD across a range of case studies.