(Vienna, 02 June 2021) An interdisciplinary team from MedUni Vienna has taken First Place in the Datathon competition run by the European Society of Intensive Care Medicine (ESICM). The multidisciplinary group comprising critical care doctors and data scientists developed an algorithm to optimise steroid treatment in sepsis and came first out of a total of 16 teams in the competition.
The purpose of the Datathon was to answer clinically relevant questions using Big Data analysis. The dissemination of electronic health records and advances in monitoring devices have made it possible to collect increasingly large quantities of data, especially in critical care.
Razvan Bologheanu, Oliver Kimberger, Thomas Hamp, Mathias Maleczek, Christoph Dibiasi and Sebastian Zeiner (all from anaesthesia), Lorenz Kapral and Daniel Laxar (both from LBI for Digital Health and Patient Safety) and Asan Agibetov (CeMSIIS) developed an algorithm to optimise steroid treatment in sepsis. The use of steroids in sepsis is still a controversial subject after 70 years (unproven benefit) and the scientists have shown that the algorithm learns from retrospective data (keyword: Reinfocement learning) which patients are likely to benefit from the treatment.
The project is still in its early stages, the algorithm is being further optimised and externally validated.
The opportunities arising from this form of Big Data are numerous and diverse and are leading to an upsurge in research funding and commercial interest. Indeed, their analysis with machine learning and predictive modelling could provide new insights and improve patient care. A Datathon for critical care is an event where multidisciplinary teams and international experts join forces to investigate clinical questions using large datasets of electronic health records.