The Secondary Use of Longitudinal Critical Care Data

Electronic health records capture rich clinical information over time providing opportunities for both discovery of disease mechanism, and for understanding and improving quality of care. It is currently unclear how electronic health records (diverse, messy data) can be used to define disease and disease sub-types relevant for precision medicine. It is also unclear how molecular mechanisms map to clinical phenotypes. By tackling such problems we may enhance the effectiveness of current treatments and improve our chances of developing new ones.

Critical care offers a particular set of opportunities because the volume of EHR data collected per patients is particularly large, it is often numeric, and it is real time. At UCL Professor Mervyn Singer and Drs Steve Harris, David Brealey and Niall MacCallum have made major steps in 'unlocking' critical care data for research. They are leading on a multicentre critical care data project funded by the NIHR that currently comprises >20,000 patients and >22 million datapoints. But there is much to be done.