The introduction of Joint Clinical Assessments (JCAs) is changing the context in which indirect comparative evidence must be generated and defended. With multiple PICO questions potentially requiring evidence across different comparators and populations, direct head-to-head data may not always be available or sufficiently aligned with the JCA decision problem.

Matching-adjusted indirect comparisons (MAICs) can help address some of these evidence gaps, but their value depends on more than technical execution. In the JCA environment, the central challenge is demonstrating that the comparison is methodologically appropriate, clinically justified, and sufficiently transparent for assessors to understand its assumptions and remaining uncertainty.

Here, we discuss MAIC, its role and what makes MAIC credible in JCA, the importance of early planning, and more.

What is MAIC?

Matching-adjusted indirect comparison (MAIC) is a type of indirect treatment comparison methodology developed to address cross-trial differences in patient characteristics in the absence of head-to-head evidence between competing interventions.

In situations where individual patient data (IPD) are available for only one study, while only aggregate-level data are available for the comparator study, MAIC applies a reweighting approach to align the baseline characteristics of the IPD population with those reported in the comparator trial. By adjusting for imbalances in observed baseline characteristics that may affect the comparison, MAIC seeks to reduce bias arising from differences between study populations. The variables that need to be adjusted for differ according to the evidence structure: in an anchored MAIC, treatment-effect modifiers are of primary importance, whereas an unanchored MAIC requires adjustment for both prognostic factors and treatment-effect modifiers and relies on substantially stronger assumptions.

As a result, MAIC has become a widely used population-adjustment method in HTA submissions when conventional anchored indirect comparisons or network meta-analysis may not adequately address important differences between trial populations.

Where does MAIC stand today?

Over the last decade, MAIC has become an established population-adjustment approach in health technology assessment (HTA) agencies, particularly when important differences exist between trial populations. NICE has developed specific methodological guidance for population-adjusted indirect comparisons, including MAIC, and MAIC analyses have also been evaluated by other HTA agencies, including CADTH/CDA.

This growing acceptance reflects the increasing recognition that indirect evidence is often unavoidable in rapidly evolving therapeutic areas. However, their credibility depends on appropriate covariate selection, sufficient population overlap, and the availability of relevant characteristics across studies. Importantly, MAIC cannot adjust for unobserved or unreported differences between populations.

JCA and its implications for MAIC

The introduction of JCAs under the EU HTA Regulation places greater emphasis on generating comparative evidence across the PICO questions relevant to EU Member States. While direct head-to-head randomized evidence remains the preferred source of comparative effectiveness information, the European HTA Coordination Group recognizes that indirect comparisons are often necessary when direct evidence is unavailable.

The 2024 Methodological Guideline for Direct and Indirect Comparisons outlines the conditions under which indirect treatment comparisons may be considered and highlights the importance of assessing exchangeability across studies, including similarity, homogeneity, and consistency assumptions.

Population adjustment should not be viewed as an automatic solution when these assumptions are challenged; the choice of method and its underlying assumptions must be justified based on the available evidence.

Conducting an MAIC for JCA

For JCAs, the key question is not simply whether an MAIC can be conducted, but whether it provides a credible comparison for the PICO being assessed. Sponsors considering MAIC should therefore provide a robust and transparent rationale for its use. It is important to highlight that the same level of rigor should be carried out for identifying relevant evidence and comparators and treatment effect modifiers. However, the identifications of treatment-effect modifiers, and prognostic factors in the case of unanchored MAICs, should be done holistically, encompassing clinical evidence, biological plausibility, and expert inputs rather than just focusing on statistical significance alone.

The weighting procedure should be clearly described, including the variables used for matching and the resulting effective sample size (ESS). A substantial reduction in ESS may indicate poor population overlap and signal instability in the resulting estimates. In addition, sensitivity analyses should be performed to evaluate the robustness of the findings to alternative assumptions and variable selections.

The JCA guidance places considerable emphasis on transparency, reproducibility, and justification of all methodological choices involved in evidence synthesis.

The challenges of presenting MAIC to JCA

Achieving balance after weighting does not, by itself, establish that an MAIC is unbiased. Assessors must consider whether:

• The studies are sufficiently comparable
• Relevant treatment-effect modifiers have been identified and adjusted for, and
• Important differences remain after weighting.

This is particularly important for unanchored MAICs, which require adjustment for both prognostic factors and treatment-effect modifiers and rely on the strong assumption that all relevant factors have been accounted for. Limited population overlap can further reduce the effective sample size and precision of the resulting estimates.

These limitations should be clearly characterized when presenting MAIC evidence within JCA.

Consider MAIC early in evidence planning

Within JCA, MAIC should be considered early in evidence-generation planning rather than as a late-stage solution to gaps in the evidence network. Sponsors should assess whether relevant covariates are collected and consistently reported across trials, whether sufficient population overlap exists, and whether the assumptions required for population adjustment can be adequately supported.

The question therefore shifts from “Can we conduct a MAIC?” to “Can we demonstrate that this MAIC provides a credible comparison for the JCA PICO?”

This has implications for trial design, comparator selection, systematic literature review, and early ITC feasibility assessment. JCA methodological guidance places particular emphasis on the assumptions underlying indirect comparisons, making issues such as limited population overlap, loss of precision following weighting, and residual bias from unmeasured or unreported treatment-effect modifiers especially important. Sponsors should therefore provide a clear rationale for the use of MAIC, test key assumptions through appropriate sensitivity and scenario analyses, and transparently communicate the remaining uncertainty.

Key takeaways

MAICs continue to remain an important option in generating comparative effectiveness evidence when direct comparative trials are unavailable. However, within the JCA environment, credibility of MAIC results depends not only on the technical execution of the analysis but also on the strength of the clinical justification, assessment of exchangeability, and the transparency of assumptions and uncertainty. The focus therefore shifts from simply conducting a MAIC to demonstrating that it is appropriate and credible for the JCA decision problem.

Interested in learning more?

Our new white paper, “Comparative Effectiveness in Motion: Guidance Development and the Growing Acceptance of Novel Methodologies in Health Technology Assessment,” explores how novel indirect treatment comparison (ITC) methods continue to evolve alongside the demand for complex evidence generation strategies to demonstrate the comparative effectiveness of new health technologies. Download your copy today!

Share Via