Tjeerd van Staa is a physician and clinical epidemiologist with training in Pharmaco-epidemiology from McGill University (MSc) and Utrecht University (PhD) and training in Medical Ethics from Kings College (MA). He is Professor of Health eResearch at the Health eResearch Centre, University of Manchester. One of his current research interests is Efficient Trials, which aims to harness advanced health informatics and electronic health records to improve clinical trials. Another interest is the Learning HealthCare System in which routinely collected data are used to feedback actionable information to clinicians and patients. Recent work has focused on measuring the uncertainty and lack of generalisability of risk prediction models that use routinely collected data (such as electronic health records from primary care). He has published over 260 peer-reviewed articles and is a well recognised speaker in the field of pragmatic trials and clinical epidemiology.
van Staa is also a Research Fellow at the Alan Turing Institute, where his work focuses on the quantification of uncertainty in models that predict clinical risks for individual patients using routinely collected data such as electronic health records (EHRs). Predictive modelling with EHRs is anticipated to drive personalised medicine and improve healthcare quality. However, EHRs typically contain information that is recorded at the point of care by multiple clinicians and varied sites, giving potential for major biases, and the quality and completeness of the records is variable. van Staa and his group aim to understand the impact on the accuracy on predicted individual risks due to these factors. In addition, they are investigating the generalisability of transportability of risk prediction models across multiple heterogeneous clinical sites. The ultimate aim of this work is to implement approaches to minimise uncertainty in risk prediction models.