Contextualizing Uncertainty in EU HTA by Coming to Terms With Imperfect Data: Can We Use Novel Analytic Methods to Quantify the Data Imperfections and Biases?

PURPOSE:

The cross-European Health Technology Assessment Regulation is now a reality. EU methodological guidelines published the minimum acceptable methodological standards to guide sponsors on designing the quantitative evidence synthesis to support comparative effectiveness results. The acceptability bar for these analyses has been set so high as to almost exclusively require findings from randomized trials with high quality, internal validity and statistical precision. In circumstances where such evidence generation is unfeasible, such as in rare diseases or complex evidence pathways, this leaves limited room for non-randomized evidence and detailed exploration of uncertainty in joint clinical assessments (JCAs). Statistical uncertainty was over-emphasized over the importance of the “shifted null hypothesis”.

DESCRIPTION:

This workshop will involve a role play, interactive, fictional EU JCA submission and assessment. The discussion leader will consider predictive PICOs for a hypothetical technology in a targeted population with high unmet need and complex evidence generation needs (5 minutes). Dr. Hernandez will take the industry perspective exploring the challenges in producing strong evidence for such technologies. Dr Kent will discuss the role of non-randomised studies in the context of JCAs and present as an example results of the Q-BASEL study, an assessment of quantitative bias analysis methods for external control arms in lung cancer. Dr Pedder will discuss whether bias-adjustment and other information sharing synthesis methods can be used to strengthen inference where high quality randomized evidence is limited. (15 minutes each). Then, the audience will help inform the company’s submission evidence package and hypothetically exploring what would be the impact in JCA if bias adjustment methods were explored. An HTAG discussion will be mirrored with emphasis on issues around internal, external validity and precision of comparative effects. An interactive discussion including live polling questions will conclude the workshop (10 minutes). Statisticians, epidemiologists, payers, and Industry and patient representatives will benefit.