Population-adjusted indirect comparisons for NICE submissions (NICE DSU TSD 18)
This method helps compare treatments when direct head-to-head trials are not available, especially for NICE appraisals. It adjusts for differences in patient populations across studies using individual patient data, making comparisons more relevant to the target population.
At a glance
Use when
Direct comparative evidence is unavailable and population differences across trials are expected to bias standard indirect comparisons; individual patient data are available for at least one trial; conducting submissions to agencies like NICE requiring robust indirect evidence
Avoid when
Individual patient data are unavailable or of poor quality; effect-modifying factors are unknown or unmeasured; unanchored comparisons are the only option and no strong external evidence supports similarity assumptions
Inputs
Individual patient data from at least one trial, aggregate data from external trials, specification of effect-modifying variables, definition of target population
Outputs
Adjusted treatment effect estimates (e.g., hazard ratios, mean differences) in the target population, measures of uncertainty, assessment of methodological assumptions
How it works
Population-adjusted indirect comparison methods, such as Matching-Adjusted Indirect Comparison (MAIC) and Simulated Treatment Comparison (STC), are used when aggregate data network meta-analysis assumptions are violated due to cross-trial differences in effect-modifying variables. These methods use individual patient data (IPD) from one or more trials to reweight or simulate outcomes in a target population, enabling adjusted indirect comparisons. The approach distinguishes between anchored (using a common comparator) and unanchored (no common comparator) comparisons, with the latter relying on stronger, often untestable assumptions. The method is applied in health technology appraisal contexts, particularly for NICE submissions, to generate more valid estimates of relative treatment effects in specific populations.
- HTA domains
- Clinical Effectiveness, Costs & Economic Evaluation, Patient and Social Aspects
- Assumptions
- Effect-modifying variables are known and measured; individual patient data are representative; correct specification of weighting or simulation models; for unanchored comparisons, similarity of trial populations in absence of a common comparator
- Strengths
- Enables more valid indirect comparisons when population differences exist; improves relevance to target population; uses available IPD to reduce bias from confounding by population differences; supports decision-making in absence of head-to-head trials
- Limitations
- Relies on availability and quality of individual patient data; unanchored comparisons involve strong, often unverifiable assumptions; results sensitive to model specification; limited empirical validation through simulation studies
- Also known as
- NICE DSU TSD 18, Population-adjusted indirect comparison, MAIC, STC, Matching-Adjusted Indirect Comparison, Simulated Treatment Comparison
Questions this answers
- › How can treatments be compared when no direct trial evidence exists?
- › How can differences in patient populations across trials be accounted for in indirect comparisons?
- › What are the assumptions behind population-adjusted indirect comparisons?
- › When should MAIC or STC be used in HTA submissions?
- › What distinguishes anchored from unanchored indirect comparisons?
- › How can results from population-adjusted methods be made transparent and consistent for decision-making?
References & sources
Similar by meaning
Beta record. Generated from the primary source via AI extraction and independent audit, pending final human review.