Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Transportability of aggregate trial results to an external environment in causally interpretable meta-analysis

Tran Trong Khoi Le, Marie-Felicia Béclin, Sivem Afach, Tat-Thang Vo

TL;DR

A novel method for incorporating AD into causal meta-analysis of IPD studies that can overcome the challenge of analytically intractable integration and modeling the trial membership as a function of baseline covariates is proposed.

Abstract

In evidence synthesis, multilevel modeling approaches (MMAs) are commonly employed to combine aggregate data (AD) and individual participant data (IPD). These approaches rely on an aggregate outcome model that is ideally obtained by integrating the prespecified individual- level outcome model over the covariate distribution observed in each eligible study. In non- linear settings, such an integration may however be analytically intractable and requires ap- proximations. In this paper, we propose a novel method for incorporating AD into causal meta-analysis of IPD studies that can overcome this challenge. Rather than relying on an ag- gregate outcome model that is difficult to be correctly formulated, we propose modeling the trial membership as a function of baseline covariates. This model allows one to estimate the individual-level outcome model in each AD study by leveraging IPD available in other trials, and then to transport the treatment effects estimated from both AD and IPD trials to an external target population, even when only aggregate covariate data are available for that population. Unlike previous proposals, we do not require pseudo-IPD to be generated from the aggregate data, which helps minimize bias due to incomplete information on the covariate distribution in each AD trial and in the target population.

Transportability of aggregate trial results to an external environment in causally interpretable meta-analysis

TL;DR

A novel method for incorporating AD into causal meta-analysis of IPD studies that can overcome the challenge of analytically intractable integration and modeling the trial membership as a function of baseline covariates is proposed.

Abstract

In evidence synthesis, multilevel modeling approaches (MMAs) are commonly employed to combine aggregate data (AD) and individual participant data (IPD). These approaches rely on an aggregate outcome model that is ideally obtained by integrating the prespecified individual- level outcome model over the covariate distribution observed in each eligible study. In non- linear settings, such an integration may however be analytically intractable and requires ap- proximations. In this paper, we propose a novel method for incorporating AD into causal meta-analysis of IPD studies that can overcome this challenge. Rather than relying on an ag- gregate outcome model that is difficult to be correctly formulated, we propose modeling the trial membership as a function of baseline covariates. This model allows one to estimate the individual-level outcome model in each AD study by leveraging IPD available in other trials, and then to transport the treatment effects estimated from both AD and IPD trials to an external target population, even when only aggregate covariate data are available for that population. Unlike previous proposals, we do not require pseudo-IPD to be generated from the aggregate data, which helps minimize bias due to incomplete information on the covariate distribution in each AD trial and in the target population.
Paper Structure (18 sections, 1 theorem, 31 equations, 4 tables)

This paper contains 18 sections, 1 theorem, 31 equations, 4 tables.

Table of Contents

  1. Introduction
  2. Estimating conditional and marginal treatment effect under treatment version heterogeneity and limited access to data
  3. Causal estimands and identification
  4. Stage 1: estimating conditional treatment effects
  5. Stage 2: transporting treatment effects to the target population
  6. Stage 3: Meta-analyzing transported treatment effects
  7. Asymptotic inference
  8. A simulation study
  9. Simulation setup
  10. Estimation strategies
  11. Analysis
  12. Simulation results
  13. Finite-sample behaviors under model mispecification and poor overlap
  14. Data analysis
  15. Discussion
  16. ...and 3 more sections

Key Result

Proposition 1

The matrix $A(\bm\psi_0)$ only involves terms of the form $\mathsf{E}\{I(S=k) \bm g(\bm O,\bm\psi_0)\}$, for $k=Z+1,\ldots,K$. The matrix $B(\bm\psi_0)$ involves terms of the same form, as well as $\mathsf{E}(\bm L\bm L^\top|S=0)$, $\mathsf{E}(\bm L\bm L^\top|S=j)$ and $\mathsf{E}(\bm L Y|x,S=j)$, w

Theorems & Definitions (3)

  • Remark 1
  • Remark 2
  • Proposition 1