Presentation (CTI)

OBJECTIVES: To understand discrepancies between manufacturer and public analyses regarding quality of life (QoL) assessment in oncology-related Japanese health technology assessment (HTA) evaluations.
METHODS: Publicly available HTA reports with a “completed” status on the C2H website between April 2019 and May 2026 were reviewed. Differences in QoL assumptions, data sources, and utility estimates between manufacturer-submitted and public analyses were assessed for oncology indications. Minutes from the MHLW Cost-Effectiveness Evaluation Expert Committee (Chuikyo) were also reviewed for additional discussions related to QoL assessment.
RESULTS: Among 19 oncology indications (13 products) reviewed, company-submitted QoL values were accepted without major modification in 9 indications, although several included important comments, scenario analyses, or methodological concerns. In 4 indications, the original data sources were accepted but assumptions were modified, resulting in revised utility estimates. In 5 indications, data sources were rejected and alternative utilities were adopted. Public analyses consistently prioritized directly measured trial-based utilities, Japanese preference-based values, and state-specific utility approaches, while critically assessing both data sources and application methods. However, inconsistencies were observed across assessments. For example, unanchored MAIC-based QoL comparisons were accepted in some cases despite acknowledged uncertainty, whereas similar approaches had previously been considered excessively uncertain and unsuitable for evaluation. Different standards were also applied to alternative utility sources, including mapped utilities, vignette studies, and external HTA-derived values. Acceptance of QoL inputs appeared partly dependent on whether alternative evidence could be identified through external HTA reports or literature searches by the assessment team. Overall, QoL estimates were often adjusted downward to avoid overestimating treatment benefit.
CONCLUSIONS: Public analyses demonstrated consistent preferences for directly measured data, Japanese population utility values, and rejection of treatment-specific QoL adjustments. However, in evidence-limited settings, such as rare cancers or heavily pretreated populations, methods associated with greater uncertainty may be considered acceptable on a case-by-case basis.