WHEN TRIAL CONTEXT SHAPES VALUE: SAME EFFECT, DIFFERENT OUTCOMES...
Author(s)
David Dai-Wee Lee, MBBS, MCO1, Yoke Fui Wong, MBBS, MCO2, Nathorn Chaiyakunapruk, PharmD, PhD3, Junice Ng, PhD4, Chee Yoong Foo, PhD, MD5.
1Sunway Medical Centre, Subang Jaya, Malaysia, 2National Cancer Institute, Putrajaya, Malaysia, 3Department of Pharmacotherapy, College of Pharmacy, University of Utah, Salt Lake City, UT, USA, 4BeOne Medicines, Ltd, Singapore, Singapore, 5BeOne Medicines, Ltd, Petaling Jaya, Selangor, Malaysia.
1Sunway Medical Centre, Subang Jaya, Malaysia, 2National Cancer Institute, Putrajaya, Malaysia, 3Department of Pharmacotherapy, College of Pharmacy, University of Utah, Salt Lake City, UT, USA, 4BeOne Medicines, Ltd, Singapore, Singapore, 5BeOne Medicines, Ltd, Petaling Jaya, Selangor, Malaysia.
OBJECTIVES: In cross-trial comparisons, similar relative treatment effects can produce different absolute outcomes because trial context shapes baseline risk. These differences complicate interpretation, as observed outcomes may reflect population risk rather than true treatment effect. While indirect comparison methods estimate relative effects, absolute outcomes still depend on baseline risk. This study examined how similar relative effects can lead to different absolute outcomes across trials and explored implications for health technology assessment (HTA) decision-making.
METHODS: Published Kaplan-Meier curves from two Phase III trials (JUPITER-02 and RATIONALE-309) in recurrent/metastatic nasopharyngeal carcinoma were used as a case study. Curves were reconstructed over a common follow-up period. Control-arm hazard patterns were used to characterize baseline risk differences between trials. Survival patterns were examined across early (0-6 months), middle (7-12 months), and late (13-18 months) phases. Control-arm hazard ratios between trials were then applied to contextualize observed progression-free survival (PFS) differences.
RESULTS: Median PFS differed substantially between trials despite similar hazard ratios (0.52 vs 0.50). Median PFS was 21.4 months in the JUPITER-02 experimental arm and 9.6 months in RATIONALE-309. Control-arm hazard patterns suggested higher baseline progression risk in RATIONALE-309, particularly during the early phase. After applying control-arm hazard ratios, median PFS estimates aligned (>18 months in both trials). Similar convergence was observed for restricted mean survival time over 18 months.
CONCLUSIONS: In HTA, economic evaluations often rely on absolute outcomes anchored to trial populations. This study demonstrates that similar relative treatment effects may produce substantially different absolute outcomes due to variation in baseline risk and trial context. Such differences can materially influence estimates of survival benefit, cost-effectiveness, and budget impact, raising important questions regarding which trial context is most relevant for decision-making. Using a control-arm hazard-anchored framework, the findings highlight the importance of contextual interpretation when translating cross-trial evidence into HTA settings.
METHODS: Published Kaplan-Meier curves from two Phase III trials (JUPITER-02 and RATIONALE-309) in recurrent/metastatic nasopharyngeal carcinoma were used as a case study. Curves were reconstructed over a common follow-up period. Control-arm hazard patterns were used to characterize baseline risk differences between trials. Survival patterns were examined across early (0-6 months), middle (7-12 months), and late (13-18 months) phases. Control-arm hazard ratios between trials were then applied to contextualize observed progression-free survival (PFS) differences.
RESULTS: Median PFS differed substantially between trials despite similar hazard ratios (0.52 vs 0.50). Median PFS was 21.4 months in the JUPITER-02 experimental arm and 9.6 months in RATIONALE-309. Control-arm hazard patterns suggested higher baseline progression risk in RATIONALE-309, particularly during the early phase. After applying control-arm hazard ratios, median PFS estimates aligned (>18 months in both trials). Similar convergence was observed for restricted mean survival time over 18 months.
CONCLUSIONS: In HTA, economic evaluations often rely on absolute outcomes anchored to trial populations. This study demonstrates that similar relative treatment effects may produce substantially different absolute outcomes due to variation in baseline risk and trial context. Such differences can materially influence estimates of survival benefit, cost-effectiveness, and budget impact, raising important questions regarding which trial context is most relevant for decision-making. Using a control-arm hazard-anchored framework, the findings highlight the importance of contextual interpretation when translating cross-trial evidence into HTA settings.
Conference/Value in Health Info
2026-09, ISPOR Asia Pacific 2026, Bangkok, Thailand
Value in Health, Volume 55, Issue S1
Code
HTA4
Topic
Health Technology Assessment
Topic Subcategory
Decision & Deliberative Processes
Disease
No Additional Disease & Conditions/Specialized Treatment Areas, SDC: Oncology