Abstract:Population adjustment methods such as matching-adjusted indirect comparison (MAIC) are increasingly used to compare marginal treatment effects when there are cross-trial differences in effect modifiers and limited patient-level data. MAIC is sensitive to poor covariate overlap and cannot extrapolate beyond the observed covariate space. Current outcome regression-based alternatives can extrapolate but target a conditional treatment effect that is incompatible in the indirect comparison. When adjusting for covariates, one must integrate or average the conditional estimate over the population of interest to recover a compatible marginal treatment effect. We propose a marginalization method based on parametric G-computation that can be easily applied where the outcome regression is a generalized linear model or a Cox model. In addition, we introduce a novel general-purpose method based on multiple imputation, which we term multiple imputation marginalization (MIM) and is applicable to a wide range of models. Both methods can accommodate a Bayesian statistical framework, which naturally integrates the analysis into a probabilistic framework. A simulation study provides proof-of-principle for the methods and benchmarks their performance against MAIC and the conventional outcome regression. The marginalized outcome regression approaches achieve more precise and more accurate estimates than MAIC, particularly when covariate overlap is poor, and yield unbiased marginal treatment effect estimates under no failures of assumptions. Furthermore, the marginalized regression-adjusted estimates provide greater precision and accuracy than the conditional estimates produced by the conventional outcome regression, which are systematically biased because the measure of effect is non-collapsible.

A note on marginalization of regression parameters from mixed models of binary outcomes

Marginalization in nonlinear mixed-effects models with an application to dose-response analysis

Mixed‐Effect Hybrid Models for Longitudinal Data with Nonignorable Dropout

Marginally Interpretable Generalized Linear Mixed Models

Flexible marginalized models for bivariate longitudinal ordinal data

Marginalization of Regression-Adjusted Treatment Effects in Indirect Comparisons with Limited Patient-Level Data

Identifying Gene-environment interactions with robust marginal Bayesian variable selection

A collection of marginalized two-part random-effects models for analyzing medical expenditure panel data: Impact of the New Cooperative Medical Scheme on healthcare expenditures in China

Adverse Subpopulation Regression for Multivariate Outcomes with High-Dimensional Predictors

Approximate Marginal Likelihood Inference in Mixed Models for Grouped Data

A bias-reduced generalized estimating equation approach for proportional odds models with small-sample longitudinal ordinal data

Robust data integration from multiple external sources for generalized linear models with binary outcomes

A Fully Nonparametric Modelling Approach to Binary Regression

Population-calibrated multiple imputation for a binary/categorical covariate in categorical regression models

A two-step penalization and shrinkage approach for binary response data that is jointly separated and correlated: The effects of social networks on diarrheal disease

Exact gradient evaluation for adaptive quadrature approximate marginal likelihood in mixed models for grouped data

Model Averaging for Estimating Treatment Effects With Binary Responses

Robust model-based estimation for binary outcomes in genomics studies

Building Marginal Models for Multiple Ordinal Measurements

Regression-with-residuals Estimation of Marginal Effects: A Method of Adjusting for Treatment-induced Confounders that may also be Moderators

Mid-quantile regression for discrete responses