Pavla Krotka, a graduate of the Master's degree program in Data Science at the University of Vienna, has received an award from the Austrian Statistical Society for her master's thesis. Her thesis entitled “Model-based Adjustments for Non-concurrent Comparisons in Platform Trials” was written at Medical University of Vienna, at the Institute of Medical Statistics, under the supervision of Martin Posch and Marta Bofill Roig and was awarded the prize for the best master's thesis in applied statistics.
Platform trials can significantly improve the efficiency of drug development as they, compared to separate randomized clinical trials, allow for the joint use of resources. These trials evaluate the efficacy of multiple treatment arms and allow for later integration of new experimental treatment arms as they become available. The efficacy of treatments is usually evaluated by comparison with a common control arm. For treatment arms that are included in the ongoing study later on, the control data are divided into concurrent and non-concurrent controls. Non-concurrent controls refer to control patients who were randomized before the respective treatment arm became part of the platform trial. Although the use of non-concurrent controls can reduce the required sample size and increase statistical power, there is a risk of biased effect estimates in the presence of time trends, which can lead to false conclusions about the efficacy of new treatments.
In her master's thesis, Pavla Krotka examines statistical methods for the inclusion of non-concurrent controls. After an overview of current methods from the literature, she proposes various extensions of the frequentist modeling strategy. Subsequently, results of a simulation study are presented in which the properties of the proposed methods are evaluated in different scenarios. Moreover, the R package “NCC” was developed as part of the work, which implements the investigated methods and can thus help statisticians planning and evaluating platform trials to decide whether the use of non-simultaneous controls is appropriate.
Every year, the Austrian Statistical Society awards prizes for the best theses in theoretical and applied statistics. This prize is intended to promote young scientists. The awards ceremony took place on October 3, 2024 in Linz, Austria.
References
- Krotka, Pavla, et al. “Statistical modeling to adjust for time trends in adaptive platform trials utilizing non-concurrent controls.” arXiv:2403.14348 (2024).
- Krotka, Pavla, et al. “NCC: An R-package for analysis and simulation of platform trials with non-concurrent controls.” SoftwareX 23 (2023): 101437.
- Krotka, Pavla, et al. “NCC”. https://cran.r-project.org/package=NCC