Biostatistics
(Bio)statistics have shown an exponential growth over the last three decades thanks to the development of powerful computer hard- and software, making it possible to tackle increasingly complex questions. This has led to the development of a rich variety of techniques, such as: survival analysis of clustered data, repeated measurements analyses, models dealing with missing data, meta-analytic techniques with multivariate outcomes, Bayesian methods, et cetera.
The objective of the Biostatistics Department is to improve the methodological quality of the research of clinical and epidemiological colleagues within and outside Erasmus MC. For the benefit of a range of research projects, the department develops statistical models and software tools. This department therefore has an indirect but important effect on the scientific community. The research carried out in the Department of Biostatistics focuses on:
Joint modelling of longitudinal and survival processes
With these novel statistical models researchers can work on better analyses of longitudinal studies where there are missing-not-at-random mechanisms and analyse survival studies when dealing with time-dependent covariates.
New models for analysing growth curves
The department develops models for multivariate growth curves.
New statistical techniques developed for analysing longitudinal genome-wide association studies (GWAS)
These techniques are expected to speed up the GWAS computations by a factor of about 500.
New Bayesian statistical techniques
These techniques are used to analyse complex hierarchical and longitudinal studies, which include blood bank data, ophthalmologic longitudinal data, and Bayesian variable selection techniques.
Models and algorithms for high-volume “omics” data
This research concerns haplotype estimation, analysis of sequencing data, proteomics, metabolomics, construction of biological networks from gene expression data and the classification and prediction with gene expression and methylation data.
Multistate Markov models for modelling longitudinal and time-to-event data
This includes developing a modelling framework, based on Markov models, which are used to estimate progression rates of chronic diseases, using various types of observational data.
For more information, faculty member profiles, and publications, visit the Department of Biostatistics website .