The research interests of the Statistical Modeling of Secondary and Registry Data Unit are statistical methods to estimate the use, effects and risks of medications based on secondary data, and to estimate the progress as well as the survival of cancer. The unit has extensive statistical expertise for studies based on the pharmacoepidemiological research database that has been established at the BIPS and on the Bremen Cancer Registry. Methodological development is dedicated to the specifics of this database, i.e. the longitudinal data structure, the huge size of the population, the multitude of available covariates, the problem of unmeasured confounders and limitations, as well as the strengths of secondary data sources. Therefore, the main research focus is on methods for adjustment for confounders and analyses of linked data sources such as:

• Two-phase designs,
• Instrumental variable methods to account for unmeasured confounding,
• Marginal structural models to adjust for time-dependent confounders,
• New statistical methods for studies of comparative effectiveness and safety of medications.

The unit supports the department of Clinical Epidemiology regarding the design, conduct, and analysis of pharmacoepidemiological studies and contributes towards the development and maintenance of the pharmacoepidemiological research database and the Bremen Cancer Registry. Furthermore, the unit is involved in the Competence Network of Secondary and Registry Data of the German National Cohort.

Bianca Kollhorst
Head of Unit

Contact:
Dr. Bianca Kollhorst
Tel: +49 (0)421 218-56980
Fax: +49 (0)421 218-56941

Email:
kollhorst(at)leibniz-bips.de

• Statistical Modeling of Primary Data
• Statistical Methods for Genetics and Molecular Epidemiology
• IT, Data Management and Medical Documentation
• Statistical Methods for Causal Inference Research Group
• Emmy Noether Junior Research Group: Beyond Prediction - Statistical Inference with Machine Learning