You are here: Home Projects Current Projects Modeling nosocomial infections …

Modeling nosocomial infections using individual patient data

Funding: 36 months, since 2011

Principal Investigator: PD Dr. sc. hum. Martin Wolkewitz (IMBI)


Four main components are required to understand a specific hospital-acquired infection (HAI) and construct a suitable model: knowledge about the biological transmission dynamics, epidemiological information on the individual patient level, demographic information for the population and the inclusion of time-dependent data. We aim to develop and apply an innovative statistical model that is able to incorporate all four components. Detailed individual patient data of a prospective cohort are available from 537 different intensive care units (ICUs) with 119,699 admissions from an international surveillance network. It includes data on several micro-organisms, resistance, anti-microbial treatment, daily records of invasive devices and potential risk factors. The major aim after developing the model is then to assess the importance of each component because effective intervention to prevent HAI requires an understanding of themost important route of acquisition. Furthermore, we expect to gain an insight about the transmission dynamics under ordinary conditions in ICUs. Patient-individual risk factors will be analysed across many ICU’s while respecting epidemic conditions; internal will be opposed to external factors. These results can then be extrapolated to other ICU settings. Finally, this work will carefully study different time scales and compare different designs to study risk factors for HAI including the popular nested case-control design.