Overall Survival Adjusted for Subsequent Treatments: A Review of Health Technology Assessments
Author(s)
Mariëtte Strydom, MA1, Maximilian Schlueter, MSc1, Pauline Herscu, MSc2, Evelina Bertranou, MSc3.
1IQVIA, London, United Kingdom, 2Daiichi Sankyo EU GmbH, Munich, Germany, 3AstraZeneca, Cambridge, United Kingdom.
1IQVIA, London, United Kingdom, 2Daiichi Sankyo EU GmbH, Munich, Germany, 3AstraZeneca, Cambridge, United Kingdom.
OBJECTIVES: While treatment switching is a common feature of randomised controlled trials, particularly in oncology, it can bias the true treatment effect on overall survival (OS). Until recently, statistical methods to adjust for treatment switching focused on instances where patients from the control arm crossed over to the experimental treatment arm. The objective of this study was to investigate how Health Technology Assessment bodies (HTAB) perceive adjustment methods controlling for treatment switching more broadly, including the case of subsequent therapies.
METHODS: HTAB methods guides were reviewed to identify guidance on adjusting OS data to correct for subsequent treatment confounding. HTA appraisals in solid tumours from 13 HTAB were screened, and 6 case studies selected. HTAB perceptions and critiques of adjustment methods for treatment switching were analysed qualitatively.
RESULTS: HTAB guidance on OS adjustment is limited and heterogenous: of the four HTAB with published guidance on this issue, three (NICE, NoMA, PBAC) consider adjustments as informative, while one (IQWiG) rejects all adjustment methods as too uncertain. The case studies confirmed that IQWiG/G-BA disagreed with all uses of OS adjustment for treatment switching, while other HTAB (NICE, PBAC, CDA, NoMA, TLV, Medicinradet) accepted it in some cases. The three most frequently submitted methods were Rank Preserving Structural Failure Time (RPSFT), Inverse Probability of Censoring Weights (IPCW), and Two-Stage Estimation (TSE). In some cases, HTAB differed in the adjustment methods they accepted; for some methods, HTAB acceptance differed case-by-case. Adjusted OS had a positive impact on 11 of 61 reviewed HTAs, and a negative impact on none.
CONCLUSIONS: Adjusting OS data to account for subsequent treatments can overcome limitations of confounded intention-to-treat analyses. Although there are cases where HTAB accepted OS adjustment, there is no methodological consensus across HTAB. Joint Clinical Assessment provides an opportunity to develop a methodological framework and increase consistency.
METHODS: HTAB methods guides were reviewed to identify guidance on adjusting OS data to correct for subsequent treatment confounding. HTA appraisals in solid tumours from 13 HTAB were screened, and 6 case studies selected. HTAB perceptions and critiques of adjustment methods for treatment switching were analysed qualitatively.
RESULTS: HTAB guidance on OS adjustment is limited and heterogenous: of the four HTAB with published guidance on this issue, three (NICE, NoMA, PBAC) consider adjustments as informative, while one (IQWiG) rejects all adjustment methods as too uncertain. The case studies confirmed that IQWiG/G-BA disagreed with all uses of OS adjustment for treatment switching, while other HTAB (NICE, PBAC, CDA, NoMA, TLV, Medicinradet) accepted it in some cases. The three most frequently submitted methods were Rank Preserving Structural Failure Time (RPSFT), Inverse Probability of Censoring Weights (IPCW), and Two-Stage Estimation (TSE). In some cases, HTAB differed in the adjustment methods they accepted; for some methods, HTAB acceptance differed case-by-case. Adjusted OS had a positive impact on 11 of 61 reviewed HTAs, and a negative impact on none.
CONCLUSIONS: Adjusting OS data to account for subsequent treatments can overcome limitations of confounded intention-to-treat analyses. Although there are cases where HTAB accepted OS adjustment, there is no methodological consensus across HTAB. Joint Clinical Assessment provides an opportunity to develop a methodological framework and increase consistency.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
HTA262
Topic
Clinical Outcomes, Economic Evaluation, Health Technology Assessment
Topic Subcategory
Value Frameworks & Dossier Format
Disease
No Additional Disease & Conditions/Specialized Treatment Areas, Oncology