Research

The Division of Biostatistics currently focuses on several research topics:

Statistical methods for clinical trials
This research area deals with innovative methods for clinical trial designs and evaluation strategies for clinical data. Motivated by our involvement in a multitude of clinical trials in all phases, we develop methods for design and analysis of clinical trials, both in the frequentist setting as well as in the Bayesian framework.

Predictive modeling with clinical and molecular data
One focus in oncology lies in the prediction of clinical endpoints from a set of candidate clinical and biological covariates to support evidence for individualized treatment recommendations. We extend standard prediction methods to include high-dimensional covariate data, and we investigate their prediction performance. We focus on time-to-event endpoints and also consider competing risks settings and multi-state modeling.

Design and analysis of dose-response experiments
As dose/concentration-response experiments are commonly performed at the DKFZ, we develop optimal design and analysis procedures for dose-response relationships, especially when combinations of substances are investigated. Dose-response experiments are particularly valuable to identify susceptibility of individual tumor samples to specific available drugs.

Distance correlation
Distance correlation is a powerful measure of dependence between random variables. We develop distance correlation methods for biomedical data including time-to-event settings and genomic measurements.

Miscellaneous statistical topics
We are in intense collaboration with many scientists and develop statistical methodology for emerging scientific question. Some examples of this are described here.