Which Marginal Estimator to Apply for External Comparator Studies: The ATT , ATE , ATU , or ATO ?
Pharmacoepidemiology and Drug Safety · 2026-11 · 35 (11)
DOI: 10.1002/pds.70523 ↗Kurz gefasst
Fachartikel in der Zeitschrift Pharmacoepidemiology and Drug Safety, online erschienen am 15.11.2026. Verfasst von Gerd Rippin. Beteiligte Einrichtung: IQVIA, Frankfurt, Germany. Hauptthemen laut MEDLINE: Research Design, Drug Approval, Clinical Trials as Topic. Die Arbeit nennt 29 Literaturangaben.
Abstract (Originalsprache)
Purpose
External Comparator (EC) studies are increasingly applied for drug approval and payer submissions. However, there is no methodological guidance yet which marginal estimator should be applied. This publication provides considerations about advantages and disadvantages when selecting a primary marginal estimator for EC studies specifically.
Methods
The Average Treatment Effect (ATE) standardizes the estimated treatment effect according to the target population as represented by the pooled sample of the Single‐Arm Trial (SAT) and the EC. The Average Treatment Effect in the Treated (ATT) uses the SAT population instead, while the Average Treatment Effect in the Untreated (ATU) estimates a treatment effect as expected in the EC population. The Average Treatment Effect in the Overlap Population (ATO) targets a population which focuses on internal validity by weighting patients with about equal chances to receive either treatment most. A detailed list of advantages and disadvantages is compiled for these marginal estimators.
Results
We argue that the complementary strengths of the ATE, ATT, ATU and ATO should be leveraged by providing all of them. However, one has to be pre‐specified as primary in the study protocol. While the ATT and ATE have been favored traditionally, the ATO and ATU are reasonable options as well. We especially highlight the option of selecting the ATU due to (1) its unbiasedness in the typical case of missing EC baseline variables, and (2) closer representativeness. Because the ATO is less straightforward to interpret, we also derive a new visual interpretation approach and show how the ATO interpretation is related to the ones of ATE, ATT, and ATU.
Conclusions
All 4 marginal estimators are recommended to be presented, at least for the primary endpoint. The derived list of advantages and disadvantages enables researchers to make informed decisions about which one to select as the primary estimator.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Autor:innen und Institutionen
- IQVIA, Frankfurt, Germany.
ORCID-Profile: Gerd Rippin
Zitieren
APA
Rippin, G. (2026). Which Marginal Estimator to Apply for External Comparator Studies: The ATT , ATE , ATU , or ATO ?. Pharmacoepidemiology and Drug Safety, 35(11). https://doi.org/10.1002/pds.70523
Vancouver
Rippin G. Which Marginal Estimator to Apply for External Comparator Studies: The ATT , ATE , ATU , or ATO ?. Pharmacoepidemiology and Drug Safety. 2026;35(11). doi:10.1002/pds.70523
BibTeX
@article{Rippin2026,
author = {Rippin, Gerd},
title = {Which Marginal Estimator to Apply for External Comparator Studies: The ATT , ATE , ATU , or ATO ?},
journal = {Pharmacoepidemiology and Drug Safety},
year = {2026},
volume = {35},
number = {11},
doi = {10.1002/pds.70523}
}
EUVIMED