The Unit "Statistical Methods in Pharmacoepidemiology" holds statistical expertise for studies based on the pharmacoepidemiological research database. The unit accompanies the development of the pharmacoepidemiological research database and provides support to the department "Clinical Epidemiology" regarding the design, conduct and analysis of pharmacoepidemiological studies. Our own methodological focus is on confounder adjustment (twophase designs (see below), propensity score and instrumental variable methods) and on signal detection methods for the identification of unknown adverse drug reactions. The unit also conducts safety and efficacy analyses for data safety monitoring boards in the course of clinical trials.
Twophase designs are of particular interest for the unit. Twophase designs combine information from two data sources: from a phase-one sample, where disease status, some characteristics and (proxy) exposure information is known for a large number of persons, and a stratified, second phase subsample with supplemental or exact exposure measurements. The statistical analysis makes use of both data sources. Within the project "Use of population data bases with double sampling for choice-based sample surveys and time-to-event data" which is funded by the German Research Foundation (DFG) we investigate the application of twophase designs in population based database studies.
Only currently running projects or those that ended less than a year ago will be shown. The entries are sorted alphabetically.