The field of biostatistics is concerned with the collection and evaluation of data in the different life sciences. Specifically in drug research, biostatistics covers the statistical trial design and analysis of studies on the efficacy and safety of medicinal products. A biostatistical methodology provides efficient plans and analysis strategies that allow extrapolating valid conclusions to larger populations by way of samples, assesses these, and quantifies the uncertainty of the statements made on this basis.

Head of the Research group

PD Dr. Norbert Benda

Phone: +49-(0)228-99-307-3514
E-Mail: norbert.benda@bfarm.de

Curriculum Vitae

Publications

Projects

Treatment effect estimation for neuropsychiatric disease trials in the presence of non-compliance

Ann-Kristin Leuchs, Jörg Zinserling, Gabriele Schlosser-Weber, Markus Neuhäuser*, Norbert Benda

Aims:
In clinical studies concerned with the development of new medicinal products for treating neuropsychiatric diseases, estimation of the treatment effect may be substantially impeded by the great number of patients discontinuing or switching therapy in the course of the study. Considering the general framework, some of these patients could be followed up until the end of the study. Incorporating this data appropriately into the statistical analysis would allow assessing the treatment effect under real conditions. This is to be distinguished from the “ideal“ treatment effect, in which case it is assumed that all patients have taken the medicine as prescribed.

Methods:
Considering potentially different patterns of discontinuing or switching treatment, and including the data collected after discontinuation or switch, various statistical methods for the estimation of treatment effects are developed and compared according to their validity and efficiency.

Outcome and conclusions:
Our previous analyses show that the choice of a valid analysis strategy depends essentially on differentiating between the “ideal” treatment effect if all patients have taken the medication as prescribed and the treatment effect achieved in actually occurring treatments. Estimation of treatment effect on the basis of the actual treatment process should include the data obtained after discontinuation or switch. For an overall picture of treatment effects, it should generally be attempted that after discontinuation/switch as many patients as possible are followed up until the end of the study.

References:

1. Leuchs AK, Zinserling J, Schlosser-Weber G, Berres M, Neuhäuser M, Benda N (2013) Estimation of the treatment effect in the presence of non-compliance and missing data. Stat Med (in press)
2. Mallinckrodt C (2013) Preventing and treating missing data in longitudinal clinical trials: a practical guide, Cambridge University Press
3. Panel on Handling Missing Data in Clinical Trials, Committee on National Statistics and Division of Behavioral and Social Sciences and Education, National Research Council (2010) The prevention and treatment of missing data in clinical trials, National Academies Press

*Hochschule Koblenz, RheinAhrCampus Remagen

Surrogate endpoint validation for count data with application in drug development for Multiple Sclerosis

Dorothee Wirtz, Markus Neuhäuser*, Jörg Zinserling, Norbert Benda

Aims:
Long lasting clinical studies and high patient numbers are often necessary to detect treatment effects in relevant endpoints likewise survival (oncology). One possibility to keep studies shorter is the use of surrogate endpoints in phase II studies measurable at an earlier point of time and reflecting the relevant endpoints.
In pharmaceutical studies on multiple sclerosis the number of lesions in the CNS is referred to as surrogate for relapse rates within a defined period of time, even though the correlation is low at patient level. A higher correlation was claimed at study level by some authors (Sormani et al. [2]).

Methods:
In this context of count data, we evaluate various methods for the validation of surrogate endpoints at study level using mixed models and simplified hierarchical models. These methods have been derived from those developed for normally distributed data by Burzykowski et al. [1] and Tibaldi et al. [3].

Outcome and conclusions:
The correct evaluation of surrogate endpoints is essential for drawing reliable conclusions from surrogates to the actually relevant endpoints. The results will also be useful for a better assessment of multiple sclerosis studies presented in authorisation procedures.

References:

1. Burzykowski T, Molenberghs G, BuyseM (2005) Evaluation of surrogate endpoints. Springer, Berlin Heidelberg New York Tokyo
2. Sormani MP, Bonzano L, Roccatagliata L et al (2009) Magnetic resonance imaging as a potential surrogate for relapses in multiple sclerosis: a meta-analytic approach. Ann Neurol 65:268–275
3. Tibaldi FS, Abrahantes JC, Molenberghs G et al (2003) Simplified hierarchical linear models for the evaluation of surrogate endpoints. J Stat Comput Simul 73(9):643–658

*Hochschule Koblenz, RheinAhrCampus Remagen