Projects

Bayesian variable/model selection techniques can be used to find a prediction model for mortality at the intensive care unit (ICU) by assessing the effect of hospital-acquired pneumonia (HAP) and other possible variables (age, patient characteristics at day of admission, patient characteristics during ICU stay) on mortality at the ICU.

For more information and to participate please visit the project page.

Project lead: Leonhard Held

We develop novel clinical trial designs for the development of new antibiotic drugs. A particular focus is on the incorporation of historical information to reduce the number of patients needed in current trials.

For more information and to participate please visit the project page.

Project lead: Leonhard Held

The project aims to examine the relationship between CD4+ and CD8+ during HIV infection and the prognostic value of CD8+ additional to CD4+ for HIV disease progression. This relationship will be examined in treatment naive patients as well as for patients after starting a highly active antiretroviral therapy (HAART). The influence of different treatment regimens on the CD4+ and CD8+ counts will be investigated. In order to answer these questions, an methodological framework will be developed which extends and applies existing methods for longitudinal data.

For more information and to participate please visit the project page.

Project lead: Leonhard Held

This research proposal aims to develop novel statistical methodology for objective Bayesian model selection in generalized regression models. There is now a large literature on automatic and objective Bayesian model selection for the linear model, which unburden the statistician from eliciting manually the parameter priors for all models in the absence of substantive prior information (Berger and Pericchi, 2001). The g-prior, usually attributed to Zellner (1986) but already used by Copas (1983), is the standard choice for the regression coefficients. However, for generalized linear models and further extensions, there are computational and conceptual problems with the g-prior approach. Similarly, research on the appropriate prior distribution on the model space and the selection of the “best” model has been done mainly in the linear model, e. g. Scott and Berger (2006, 2010) and Barbieri and Berger (2004).
We will fill these gaps and will extend the scope of objective Bayesian model selection to generalized regression models.

For more information and to participate please visit the project page.

Project lead: Leonhard Held
Funding: Swiss National Science Foundation

This research project aims to develop novel statistical methodology for both retro- and prospective analysis of space-time data on infectious disease incidence. The new techniques will be applied in the particular context of space-time surveillance data, but important parts of the methodology can be used in a wider context.

Project lead: Leonhard Held
Funding: Swiss National Science Foundation

During infectious disease outbreaks such as the current coronavirus disease (COVID-19) pandemic, modern surveillance systems continuously produce detailed data on reported disease incidence. Typically, these data are available at various geographic resolutions and stratified by age and sex, leading to high-dimensional count time series. Statistical modelling approaches which can handle the heterogeneities and interdependencies in such data are a valuable tool to inform public health decision makers about disease dynamics, to evaluate the effect of intervention measures, and to provide probabilistic forecasts of disease spread. Important factors which need to be taken into account are social contact patterns, mechanisms of geographic spread, and possible underreporting, all of which can vary across regions, age groups, and time. The endemic-epidemic (in the following: EE) framework is an established flexible modelling framework for multivariate infectious disease surveillance counts. A robust, free, and easy-to-use implementation is provided in several R packages. To our knowledge, this is the only readily available implementation of a sophisticated and general model framework for age-stratified spatio-temporal surveillance data. In the past, the EE framework has mainly been used for seasonal diseases, but there is a clear need for general and well-implemented multivariate modelling tools also for acute outbreak situations like the current COVID-19 pandemic. The goal of this project is to extend the EE framework to further improve its applicability in such contexts. Specifically the extensions aim to better address the following aspects:

• Assessing the impact of underreporting due to asymptomatic and prodromal carriage and insufficient levels of testing
• Determining the role of different age groups and their contact patterns in transmission
• Providing impact estimates of control and mitigation strategies such as travel restriction and other social distancing policies.

This project will provide evidence to improve public health response and aid in decisions on optimal social control strategies, particularly when to initiate travel restrictions and social distancing measures, and improve situational awareness. For further information please visit the project page.

Project lead: Leonhard Held
Funding: Swiss National Science Foundation