Data driven multi level models of infectious diseases

NWO-program on Computational Life Sciences: Data driven multi level models of infectious diseases.

Antibiotic resistance of micro organisms is an emerging global health care problem, driven by the spread of micro organisms and resistance genes and fuelled by the very use of antibiotics. Data on prevalence and incidence of resistant micro organisms are abundant. But how to base sound conclusions, e.g. concerning the effectiveness of various potential control measures, on these data ? Traditional multivariate risk factor analyses neglect that cross-transmission creates dependence among individuals (or, in a different jargon, leads to nonlinearity at the population level). Hence they may miss the main point. Our group focuses on a mechanistic description of pathogen transmission within and between patients in a hospital. Our aim is to determine from the available data the relative importance of the various routes leading to detectable colonisation. The quantification of transmission routes is a prerequisite for the quantification of the effects of control measures. Our work centers around three tenets:

given the low number of patients in hospital settings, the model should be formulated as a Markov chain;

hosts should be subdivided in ”compartments”, each of which can either be colonised or ”free”;

probabilistic consideration of unobservable events (transmission) should not thwart the deterministic bookkeeping of known facts (such as admission, discharge and the outcome of tests).

We plan to develop, along these lines, a flexible methodology, taking ongoing clinical studies as guiding principal, challenge and test case, all in one.