Header

Search
Center for Reproducible Science and Research Synthesis

Explaining variation and heterogeneity in replication projects and in real world evidence emulations

The difference in effect estimates, hazard ratios, observed in the RCT and the pooled RWE depending on whether it was a close emulation or not. The horizontal line represents a scenario of no difference between RWE and RCT estimates.

 

Using methodology from meta-analysis we aim quantifying and explaining the amount of heterogeneity in two different use cases. First, we use the results from RCT-DUPLICATE (https://www.rctduplicate.org/) to understand whether design differences explain variation in results between randomized trials and their non-randomized emulations. In the second case-study, we use the large-scale replication projects to further investage the methods to explain heterogeneity between the original and replication results.

Project Lead

Rachel Heyard

CRS Collaborators

Leonhard Held

External Collaborators

Sebastian Schneeweiss, Shirley V Wang

Publications

Heyard, R., Held, L., Schneeweiss, S., and Wang, S. V. (2023). Design differences explain variation in results between randomized trials and their non-randomized emulations. Preprint. https://doi.org/10.1101/2023.07.13.23292601

Köppe, J., Micheloud, C., Erdmann, S., Heyard, R., & Held, L. (2025). Assessing the replicability of RCTs in RWE emulations. BMC Medical Research Methodology, 25(1), 141. https://doi.org/10.1186/s12874-025-02589-z

Heyard, R., Held, L. (2025). Meta-regression to explain shrinkage and heterogeneity in large-scale replication projects. PLOS ONE https://doi.org/10.1371/journal.pone.0327799