Presentation (CTI)

OBJECTIVES: Fractional polynomial (FP) models offer a flexible approach to survival analysis in health technology assessment (HTA), enabling extrapolation of complex hazard functions and relaxing proportional hazards assumptions. While FP models are increasingly used in economic evaluations, there is little clarity on how HTA agencies perceive or guide their application. This exploratory review examines (1) the extent to which major HTA agencies have accepted FP models in recent submissions, and (2) whether any formal or informal guidance on FP use exists within public HTA methodology.
METHODS: We reviewed HTA submissions and guidance documents published between 2020 and 2025 across 13 national HTA agencies known to engage in survival modeling: NICE (UK), CADTH (Canada), HAS (France), PBAC (Australia), ZIN (Netherlands), TLV (Sweden), NCPE (Ireland), SMC (Scotland), REvalMed (Spain), AGENAS (Italy), MHLW (Japan), HIRA (South Korea), and Medicinrådet (Denmark). Public appraisal reports and official HTA websites were screened for evidence of FP model use or guidance. Findings were summarized in comparative tables.
RESULTS: FP models were identified in submissions to NICE, CADTH, HAS, NCPE, SMC, and Denmark’s Medicinrådet, primarily in oncology and where non-proportional hazards were present. No evidence of FP model use was found in submissions from ZIN, TLV, Spain, Japan, South Korea, or Italy. Only four agencies (NICE, CADTH, PBAC, and Denmark) offered formal or semi-formal methodological support for flexible survival models. Most others evaluated FP-based models on a case-by-case basis, without explicit endorsement.
CONCLUSIONS: While fractional polynomial survival models are emerging in HTA submissions, their use remains limited and uneven across jurisdictions. Methodological guidance is scarce, raising the need for harmonized recommendations as flexible modeling becomes more prevalent in survival extrapolation.