Statistical Modeling to Adjust for Time Trends in Adaptive Platform Trials Utilizing Non‐Concurrent Controls
Biometrical Journal · 2025
DOI: 10.1002/bimj.70059 ↗Kurz gefasst
Fachartikel in der Zeitschrift Biometrical Journal, 2025. Verfasst von Pavla Krotka und 4 weiteren Autor:innen. Beteiligte Einrichtung: Center for Medical Data Science, Medical University of Vienna, Vienna, Austria. Hauptthemen laut MEDLINE: Models, Statistical, Biometry, Clinical Trials as Topic. Die Arbeit wurde bisher 7-mal zitiert (Crossref) und nennt 37 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Austrian Science Fund, Swedish Research Council, Deutsche Forschungsgemeinschaft.
Abstract (Originalsprache)
Utilizing non‐concurrent control (NCC) data in the analysis of late‐entering arms in platform trials has recently received considerable attention. While incorporating NCC can lead to increased power and lower sample sizes, it might introduce bias to the effect estimators if temporal drifts are present. Aiming to mitigate this potential bias, we propose various frequentist model‐based approaches that leverage the NCC, while adjusting for time. One of the currently available models incorporates time as a categorical fixed effect, separating the trial duration into periods, defined as time intervals bounded by any arm entering or leaving the platform. In this work, we propose two extensions of this model. First, we consider an alternative definition of time by dividing the trial into fixed‐length calendar time intervals. Second, we propose alternative model‐based time adjustments. Specifically, we investigate adjusting for random effects and employing splines to model time with a polynomial function. We evaluate the performance of the proposed approaches in a simulation study and illustrate their use through a case study. We show that adjusting for time via a spline function controls the type I error in trials with a sufficiently smooth time trend pattern and may lead to power gains compared to the standard fixed effect model. However, the fixed effect model with period adjustment is the most robust model for arbitrary time trends, provided that the trend is equal across all arms. Especially, in trials with sudden changes in the time trend, the period‐adjustment model is preferred if NCCs are included.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Center for Medical Data Science, Medical University of Vienna, Vienna, Austria.
- Department of Statistics and Operations Research, Universitat Politècnica de Catalunya, Barcelona, Spain.
- Department of Pharmacy, Uppsala University, Uppsala, Sweden.
- Department of Neurology, LMU University Hospital, Ludwig-Maximilians-Universität (LMU) München, Munich, Germany.
- German Center for Neurodegenerative Diseases (DZNE), Munich, Germany.
- Munich Cluster for Systems Neurology (SyNergy), Munich, Germany.
ORCID-Profile: Pavla Krotka, Marta Bofill Roig
Förderung
- Austrian Science Fund
- Swedish Research Council
- Deutsche Forschungsgemeinschaft
- European Joint Programme on Rare Diseases
- Vetenskapsrådet
Zitieren
APA
Krotka, P., Posch, M., Gewily, M., Höglinger, G., & Bofill Roig, M. (2025). Statistical Modeling to Adjust for Time Trends in Adaptive Platform Trials Utilizing Non‐Concurrent Controls. Biometrical Journal. https://doi.org/10.1002/bimj.70059
Vancouver
Krotka P, Posch M, Gewily M, Höglinger G, Bofill Roig M. Statistical Modeling to Adjust for Time Trends in Adaptive Platform Trials Utilizing Non‐Concurrent Controls. Biometrical Journal. 2025. doi:10.1002/bimj.70059
BibTeX
@article{Krotka2025,
author = {Krotka, Pavla and Posch, Martin and Gewily, Mohamed and Höglinger, Günter and Bofill Roig, Marta},
title = {Statistical Modeling to Adjust for Time Trends in Adaptive Platform Trials Utilizing Non‐Concurrent Controls},
journal = {Biometrical Journal},
year = {2025},
doi = {10.1002/bimj.70059}
}
EUVIMED