The Statistical Abyss: Real-World Evidence for Health Technology Assessment
Author(s)
Martin Scott, BSc, MSc, Kerry Mueller, B.Sc. M.Sc..
Numerus, Reutlingen, Germany.
Numerus, Reutlingen, Germany.
OBJECTIVES: To highlight the methodological challenges and strategic considerations in using real-world evidence (RWE) for health technology assessment (HTA) in the European Union, particularly under the Joint Clinical Assessment (JCA) framework.
METHODS: This conceptual analysis draws on regulatory guidance from the European Medicines Agency (EMA) and the EU HTA Coordination Group. It synthesizes key challenges in non-interventional study design, including selection bias, attrition, and confounding. Advanced statistical techniques such as propensity score adjustment, double robust methods, and marginal structural modelling are discussed, alongside the importance of early planning and interdisciplinary collaboration.
RESULTS: Non-randomised evidence presents a high risk of bias in estimating treatment effects. The EMA and JCA guidelines emphasise the need for rigorous design and analysis to mitigate these biases. Real-world data (RWD) often lacks completeness and consistency, complicating causal inference. Sensitivity analyses, including Quantitative Bias Assessment (QBA), are essential to validate findings. Strategic planning—such as developing RWE protocols and conducting analyses prior to PICO scope finalisation—is critical for successful dossier submission.
CONCLUSIONS: While randomised controlled trials remain the gold standard, the increasing use of RWE in regulatory and HTA contexts demands robust statistical approaches and careful planning. Statisticians play a pivotal role in guiding clinical and market access teams to ensure internal validity and scientific rigor. Given the implications for oncology and ATMPs under the EU HTA JCA since January 2025, proper planning and methodologically sound RWE strategies are more important than ever.
METHODS: This conceptual analysis draws on regulatory guidance from the European Medicines Agency (EMA) and the EU HTA Coordination Group. It synthesizes key challenges in non-interventional study design, including selection bias, attrition, and confounding. Advanced statistical techniques such as propensity score adjustment, double robust methods, and marginal structural modelling are discussed, alongside the importance of early planning and interdisciplinary collaboration.
RESULTS: Non-randomised evidence presents a high risk of bias in estimating treatment effects. The EMA and JCA guidelines emphasise the need for rigorous design and analysis to mitigate these biases. Real-world data (RWD) often lacks completeness and consistency, complicating causal inference. Sensitivity analyses, including Quantitative Bias Assessment (QBA), are essential to validate findings. Strategic planning—such as developing RWE protocols and conducting analyses prior to PICO scope finalisation—is critical for successful dossier submission.
CONCLUSIONS: While randomised controlled trials remain the gold standard, the increasing use of RWE in regulatory and HTA contexts demands robust statistical approaches and careful planning. Statisticians play a pivotal role in guiding clinical and market access teams to ensure internal validity and scientific rigor. Given the implications for oncology and ATMPs under the EU HTA JCA since January 2025, proper planning and methodologically sound RWE strategies are more important than ever.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
MSR202
Topic
Clinical Outcomes, Methodological & Statistical Research, Real World Data & Information Systems
Topic Subcategory
Confounding, Selection Bias Correction, Causal Inference
Disease
No Additional Disease & Conditions/Specialized Treatment Areas