Evidence Generation & Comparative Effectiveness
Generate Evidence that Stands Up to Scrutiny
The strength of an EU JCA assessment depends on the quality, relevance and credibility of the evidence behind it. As direct head-to-head evidence is not always available for every PICO, sponsors increasingly need to understand earlier whether indirect evidence, population-adjusted methods or real-world data can credibly address the decision problem.
Cytel combines evidence synthesis, comparative effectiveness, RWE and biostatistics to help sponsors assess those questions early and develop an evidence strategy that is both rigorous and realistic.
Talk to an EU JCA expertEvidence Generation Starts with the Decision Problem
Cytel does not begin with a preferred statistical method. We begin with the clinical and policy question. Our teams assess the PICO, available evidence, study designs, populations, comparators, endpoints, and relevant treatment-effect modifiers before determining which analytical approach is appropriate.
This may involve systematic literature reviews, effect-modifier and prognostic-factor reviews, evidence mapping, ITC feasibility assessments, network meta-analysis, MAIC, RWE or supplementary statistical analyses.
The objective is not to maximize the number of analyses performed. It is to develop the strongest evidence package that can credibly answer the question.
Comparative Effectiveness Under EU JCA
Where direct comparative evidence exists, it remains the preferred basis for comparative effectiveness assessment. When it does not, the analytical pathway becomes more complex.
Cytel evaluates whether a connected evidence network exists, whether study populations and treatment regimens are sufficiently comparable, whether key effect modifiers are available and whether assumptions such as exchangeability, homogeneity and consistency can reasonably be supported.
Where appropriate, anchored indirect treatment comparisons or NMA may provide the strongest solution. Where networks are disconnected or population differences require adjustment, approaches such as MAIC, STC or ML-NMR may be considered.
The choice should always follow the evidence structure — not the other way around.
MAIC: From Technical Possibility to Credible Evidence
Matching-adjusted indirect comparison (MAIC) is a population-adjustment method developed to address differences between trial populations when individual patient data are available for one study and only aggregate data are available for another.
MAIC reweighs the IPD population so selected baseline characteristics align more closely with the comparator population. In anchored MAIC, treatment-effect modifiers are particularly important; unanchored MAIC requires adjustment for both prognostic factors and treatment-effect modifiers and consequently relies on substantially stronger assumptions.
MAIC has become an established population-adjustment approach in HTA, but its credibility depends on much more than achieving covariate balance. The relevant variables need to be clinically justified, population overlap needs to be adequate, and the impact of weighting on effective sample size needs to be understood.
The question is no longer “Can we conduct a MAIC?”
It is “Can we demonstrate that this MAIC provides a credible comparison for the JCA PICO?”
This distinction is particularly important in JCA. A technically successful analysis does not automatically mean that the resulting estimate is credible. Assessors need to understand the comparability of the studies, the variables that were adjusted for, the assumptions that remain, and the uncertainty around the resulting estimate.
Cytel therefore assesses MAIC feasibility early, including evidence availability, treatment-effect modifiers, prognostic factors, population overlap, weighting and effective sample size, and the sensitivity and scenario analyses needed to characterize residual uncertainty.
RWE and External Control Arms
For single-arm studies and other evidence-constrained development programmes, real-world evidence and external comparators may become part of the broader evidence strategy.
However, the availability of RWD does not automatically make an external control arm appropriate for JCA. Sponsors need to consider whether the data source represents the relevant population, whether treatment and outcome definitions can be aligned, whether confounding can be addressed and whether residual uncertainty can be transparently characterised.
This creates an opportunity for Cytel to bring RWE and comparative-effectiveness expertise together. The right question is not whether an ECA, MAIC, or another method can technically be produced, but which approach provides the most credible evidence for the decision problem.
FAQs
Can MAIC be used for EU JCA?
It can be considered where the evidence structure supports it, but its suitability must be justified for the specific PICO and the assumptions underlying the analysis.
What is the difference between anchored and unanchored MAIC?
Anchored MAIC uses a common comparator and focuses primarily on treatment-effect modifiers. Unanchored MAIC has no common comparator and requires adjustment for both treatment-effect modifiers and prognostic factors, resulting in materially stronger assumptions.
What does effective sample size tell us?
Effective sample size indicates how much information remains after reweighting. A substantial reduction can indicate limited overlap between populations and may result in less precise or unstable estimates.
Can MAIC remove all cross-trial bias?
No. MAIC can only adjust for characteristics that are measured and available across the studies. Unobserved or unreported differences cannot be eliminated.
We help sponsors prepare for JCAs
Preparing for an upcoming EU JCA? Talk to Cytel’s experts about building an integrated evidence strategy.