Xintera
Making evidence work
Constructing the estimand
A successful development starts with the construction of a suitable estimand (ICH E9 R1). An estimand has five attributes: the treatment itself, the treated population, the endpoint, the summary measure, and anticipated intercurrent events.
Intercurrent events are events that occur after randomization and can affect the interpretation of the endpoint and summary measure, such as treatment discontinuation, treatment switch, death or loss to follow-up.
The impact of these events need to be discussed with regulators so as to agree on their interpretation for the analysis.
Defining appropriate endpoints
Suitable trial endpoints are identified based on review of current regulatory guidance, prior approvals, Health Technology Assessment appraisals and academic guidelines.
Clinical, anatomical and patient-reported endpoints are then evaluated for their suitability and acceptability and relevance vis-à-vis future stakeholders: regulators, payers, healthcare professionals, patients and caregivers.
Evidence exploration and expansion
Pivotal clinical trials tend to be powered for the primary endpoint only. All too often, underpowered secondary endpoints are dismissed as "non-significant", even when the results show clinically relevant effect sizes. We believe these are opportunities for further analysis and research.
Modern clinical trials yield rich datasets that can contain information to help identify treatment response covariates and clinical subgroups that are more or less likely to benefit from treatment.
These analyses should be used to generate hypotheses for future research in Phase 4 or real-world settings.
Evidence consolidation
At later stages in the product life cycle, real-world evidence and randomized trial data can be combined and compared by means of systematic literature reviews that may use meta-regression and/or individual participant data.
This allows to confirm that the effectiveness observed in rigorously controlled trials translates well into clinical practice.
These analyses can then be leveraged in network meta-analyses vs competitor products.