¹MS-Register der DMSG, Bundeverband e.V., MS Forschungs- und Projektentwicklungs-gGmbH, Hannover, Germany
²Hochschule Hannover, Hannover, Germany
³Leibniz Universität Hannover, Hannover, Germany

Text

Central Statistical Monitoring (CSM) is increasingly recognized for its role in enhancing data quality and compliance in clinical trials [1], [2]. The literature contains a wealth of research that utilizes basic statistical methods for the CSM benefit. However, these methods include various shortcomings and limitations. This research presents a general framework for the application of CSM using comparisons of the average of individual centers to the Grand Mean (GM). This approach can be applied across various data types with appropriate statistical models. A common output is generated for a straightforward interpretation to ease the assessment decision process. Using Real-World Data (RWD) from the German Multiple Sclerosis Register (GMSR) the models’ performance were validated in practice. The practical application of these methods on GMSR data highlights their utility in identifying centers that deviate from the GM.

In previous research studies, we have investigated various data types with their adequate statistical models to ensure their robustness and applicability for a CSM methodology [3], [4], [5]. Specifically, we ran various balanced and unbalanced simulation studies with numerous scenarios relevant to clinical trials to test whether the statistical models advocated would be appropriate to control the familywise error rate (FWER) to 5%. Generalized linear models (GLM) and Bayesian GLM models were used for binomial and count data. To account for overdispersion and the presence of zero counts (zero inflation) Bias-reduced GLM and a negative binomial GLM were additionally tested for count data outcomes respectively. Furthermore, a non-parametric method was used for ordinal outcomes whereas for a time-to-event outcomes Weibull and Cox proportional hazards models were used.

Each of these studies provided insights into the strengths and limitations of the respective models, allowing us to refine our approach and ensure its reliability across diverse data types commonly encountered in clinical trials. The control of FWER by the statistical models varied according to the settings of a specific simulation in the data generation process. Thus, a guideline is needed to choose a proper model when running GM comparisons. In this work we present a general guideline embedded within an R Shiny app that includes assistance for the user on the model requirements when applied on the supplied outcome. Furthermore, the user can utilize the app to run GM comparisons on their dataset. The user could also generate graphs that include simultaneous confidence intervals for contrasts of center means with the GM. Finally, the guideline provides the potential extensions to the statistical models that can be applied to run the GM comparison when additional covariates should be considered.

The authors declare that they have no competing interests.

The authors declare that an ethics committee vote is not required.

Literatur

[1] Niangoran S, Barbieri A, Badje A, Journot V, Kouame G, Marcy O, et al. A New Centralized Statistical Monitoring Method for Detecting Atypical Distribution of Qualitative Variables in Multicenter Randomized Controlled Trials. Stat Biopharm Res. 2024;17(3):479–487. DOI: 10.1080/19466315.2024.2404631
[2] EMA. Reflection paper risk based quality management in clinical trials. 2013 [cited 2025 Feb 4]. Available from: https://www.ema.europa.eu/en/documents/scientific-guideline/reflection-paper-risk-based-quality-management-clinical-trials_en.pdf
[3] Fneish F, Ellenberger D, Frahm N, Stahmann A, Fortwengel G, Schaarschmidt F. Application of Statistical Methods for Central Statistical Monitoring and Implementations on the German Multiple Sclerosis Registry. Ther Innov Regul Sci. 2023 Nov 1;57(6):1217–28.
[4] Fneish F, Ellenberger D, Frahm N, Stahmann A, Schaarschmidt F. Appropriate statistical model for count data in central statistical monitoring and application on German Multiple Sclerosis Registry. In: Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie, editor. 68. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie e. V. (GMDS). Heilbronn, 17.-21.09.2023. Düsseldorf: German Medical Science GMS Publishing House; 2023. DocAbstr. 164. DOI: 10.3205/23gmds160
[5] Fneish F, Ellenberger D, Frahm N, Stahmann A, Fortwengel G, Schaarschmidt F. Central Statistical Monitoring for time-to-event Endpoints and Application on Data from the German Multiple Sclerosis Registry. In: Gesundheit – gemeinsam. Kooperationstagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie (GMDS), Deutschen Gesellschaft für Sozialmedizin und Prävention (DGSMP), Deutschen Gesellschaft für Epidemiologie (DGEpi), Deutschen Gesellschaft für Medizinische Soziologie (DGMS) und der Deutschen Gesellschaft für Public Health (DGPH). Dresden, 08.-13.09.2024. Düsseldorf: German Medical Science GMS Publishing House; 2024. DocAbstr. 184. DOI: 10.3205/24gmds141