A Systematic Literature Review of the Mitigation Strategies to Overcome the Risk for Confounding of Overall Survival in Trials With Crossover/Treatment Switching Evaluated in Health Technology Assessments of Treatments for Select Tumors in Oncology
Author(s)
Anandaroop Dasgupta, PhD1, Ankita Kaushik, PhD2, Jens Grueger, PhD3, Barinder Singh, MPharm4, Tomasz Burzykowski, PhD5, Sumeet Attri, MPharm4.
1Director, Gilead Sciences, San Francisco, CA, USA, 2Gilead Sciences Inc., Foster City, CA, USA, 3The CHOICE Institute, School of Pharmacy, University of Washington, Seattle, WA, USA, 4Pharmacoevidence, London, United Kingdom, 5IDDI Louvain-la-Neuve, Belgium, Belgium.
1Director, Gilead Sciences, San Francisco, CA, USA, 2Gilead Sciences Inc., Foster City, CA, USA, 3The CHOICE Institute, School of Pharmacy, University of Washington, Seattle, WA, USA, 4Pharmacoevidence, London, United Kingdom, 5IDDI Louvain-la-Neuve, Belgium, Belgium.
OBJECTIVES: Switching from control to experimental arm often confounds overall survival (OS) estimates. This systematic literature review (SLR) assessed how heath technology assessment (HTA) bodies have evaluated and critiqued mitigation approaches to address treatment switching impacting OS, and acceptance of unaffected endpoints [e.g. progression-free survival (PFS)] in clinical trials studying advanced/metastatic cancer of lung (non-small cell, NSCLC), breast, prostate, renal, colorectal, and gastrointestinal stromal tumors.
METHODS: Published reports (2013-2024) of Institute for Clinical and Economic Review, and HTA bodies of the UK (NICE), France (HAS), Germany (IQWiG), Australia (PBAC) and Canada (CDA) were retrieved from respective websites. Search strategies were augmented with information from clinical practice guidelines (CPG) and product labels (PL) from regulatory agencies (Spain, Italy and US).
RESULTS: Of 1,653 records screened, 162 product submissions permitting treatment switches were included (HTA: 147; PL: 11; CPG: 4). Mitigation strategies were reported in 63 submissions, primarily in NSCLC (n=31) and first-line setting (n=39). Seven reported surrogate endpoints (PFS, NICE: 5, CDA: 1; disease-free survival: NICE:1), while 11 included counterfactual validation (NICE: 4, CDA: 1, HAS: 5, IQWiG: 1). Rank preserving structural failure time (RPSFT), iterative parametric estimation, inverse probability of censoring weighting (IPCW), 2-stage methods were the statistical approaches used leveraging trial data with the RPSFT method being most common (n=41). Mitigation methods were prespecified in six submissions. IPCW-adjusted results were submitted post-hoc based on ERG recommendation (TA432). Surrogates were accepted by NICE, CDA, and PBAC; counterfactuals by NICE, CDA, and HAS. IQWiG accepted counterfactual validation derived from trial data, however PFS as surrogate was rejected.
CONCLUSIONS: Considerable variability exists in HTA adoption of mitigation strategies to address impact of treatment switching on OS in oncology trials. Since cross-over is an ethical approach to trial design for patients with life threatening illness, it is important to align on approaches to inform HTA decision making.
METHODS: Published reports (2013-2024) of Institute for Clinical and Economic Review, and HTA bodies of the UK (NICE), France (HAS), Germany (IQWiG), Australia (PBAC) and Canada (CDA) were retrieved from respective websites. Search strategies were augmented with information from clinical practice guidelines (CPG) and product labels (PL) from regulatory agencies (Spain, Italy and US).
RESULTS: Of 1,653 records screened, 162 product submissions permitting treatment switches were included (HTA: 147; PL: 11; CPG: 4). Mitigation strategies were reported in 63 submissions, primarily in NSCLC (n=31) and first-line setting (n=39). Seven reported surrogate endpoints (PFS, NICE: 5, CDA: 1; disease-free survival: NICE:1), while 11 included counterfactual validation (NICE: 4, CDA: 1, HAS: 5, IQWiG: 1). Rank preserving structural failure time (RPSFT), iterative parametric estimation, inverse probability of censoring weighting (IPCW), 2-stage methods were the statistical approaches used leveraging trial data with the RPSFT method being most common (n=41). Mitigation methods were prespecified in six submissions. IPCW-adjusted results were submitted post-hoc based on ERG recommendation (TA432). Surrogates were accepted by NICE, CDA, and PBAC; counterfactuals by NICE, CDA, and HAS. IQWiG accepted counterfactual validation derived from trial data, however PFS as surrogate was rejected.
CONCLUSIONS: Considerable variability exists in HTA adoption of mitigation strategies to address impact of treatment switching on OS in oncology trials. Since cross-over is an ethical approach to trial design for patients with life threatening illness, it is important to align on approaches to inform HTA decision making.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
HTA16
Topic
Health Technology Assessment
Topic Subcategory
Systems & Structure
Disease
Oncology