Automated risk stratification for infectious complications in immunodeficiency [PhD project]
Diseases such as rheumatoid arthritis, multiple sclerosis, or inflammatory bowel disease used to be very debilitating but can now be treated successfully with drugs that target the immune system. As a side effect, however, these treatments can induce secondary immunodeficiency thereby increasing the patient’s risk for potentially fatal infections. We will investigate ways to detect these increased risks. This project is part of the Personalized Prevention and Care theme.
Healthcare professionals are insufficiently acquainted with the various forms of secondary immunodeficiency that these treatments can induce and the associated infectious risks for the patients. As a consequence, patients may develop infections that could have been prevented or detected at an earlier stage, if the treating physician had been alerted to the risks. These so-called ‘opportunistic’ infections cause additional morbidity and mortality, increased numbers of hospital admissions and higher prescription rates of antibiotics.
In this project, we aim to improve the timely detection of increased infectious risk in patients treated with immunomodulatory drugs by developing an actionable prediction model based on data in electronic health record systems.
Our collaboration with STZ hospitals and the Dutch Pharmacovigilance centre Lareb will provide us with the electronic health record data that are needed to achieve our goal and with information to identify risk profiles, frequencies of the various types of infections, and the countermeasures that are available once a patient is identified as being at an increased risk of serious infection.
This project will use natural language processing techniques to quantify and extract information from data from the unstructured, free-text fields in the electronic health record system, apply techniques from supervised machine learning to learn from historical patient data to predict infection risk, and examine how a human-in-the-loop approach can improve the predictive performance and the healthcare professionals’ acceptance of such a tool in the decision-making process.
The interdisciplinary team comprises the following members with complementary expertise:
Prof. dr. Esther de Vries is a medical doctor and immunologist, now endowed professor at Tranzo, Tilburg School for Social and Behavioral Sciences, Tilburg University, Coordinator Data Science at Jeroen Bosch Hospital, and Senior Researcher at Laboratory of Medical Microbiology and Immunology of the Elisabeth-Tweesteden Hospital. She has ample experience in healthcare research and is an immunodeficiency expert. Her research focuses on using existing data to enable timely recognition of aberrant patterns such as secondary immunodeficiency.
Dr. Bennett Kleinberg, Assistant Professor in Data Science, Department of Methodology and Statistics, Tilburg School for Social and Behavioral Sciences, Tilburg University, and Honorary Associate Professor at University College London. His research revolves around the use of text data for computational social science.
Dr. Katrijn van Deun, Associate Professor in Computational Statistics, Department of Methodology and Statistics, Tilburg School for Social and Behavioral Sciences, Tilburg University. Her research focuses on the development of novel data analysis tools for high-dimensional multi-view data.
The Herbert Simon Research Institute for Health, Well-being, and Adaptiveness is a research center devoted to carrying out excellent, state of the art research in order to contribute to healthy and resilient people. We have selected three themes, which involve the collaboration between various Departments and address actual themes in need of both fundamental and applied research.