Development of predictive models for critically ill patients with acute kidney injury

Predictive models are widely used in intensive care units (ICU), due to the widespread implementation of electronic systems to collect patient data: demographic information, continuous monitoring of vital functions, information on medication administered, results from labatory analyses, etc. In this project, we will develop predictive modelling techniques to tackle a number of machine learning challenges related to data analysis within an ICU context. One of these challenges is dealing with time to event data (also called survival data). Survival data analysis has a sound statistical basis, however has been underexplored in the machine learning community. While these approaches have been mostly used to detect associations between covariates and survival time, nowadays there is a great interest in prognostic models and their application to personalized medicine. Physicians are interested in accurate prognostic tools that will inform them about the future prospect of a patient in order to adjust medical care. We will focus on data provided by the AZ Groeninge hospital, in the area of acute kidney injury.

Keywords:machine learning, survival analysis, intensive care unit, acute kidney injury, personalized medicine

Disciplines:Artificial intelligence, Cognitive science and intelligent systems, Scientific computing, Bioinformatics and computational biology, Public health care, Public health services, Laboratory medicine, Palliative care and end-of-life care, Regenerative medicine, Other basic sciences, Other health sciences, Nursing, Other paramedical sciences, Other translational sciences, Other medical and health sciences