Department of Medical Informatics

Mission statement

Our mission is to generate reliable evidence from health data to enable better health decisions and better care. We do this by developing and applying novel quantitative and computational methods in close collaboration with academic partners, physicians, patients, payers, regulators and industry, and with a strong commitment to open science and knowledge sharing.

Management structure and decision making

The Health Data Science group, led by Peter Rijnbeek, is developing methods to exploit the wealth of data in electronic health records and other health databases, with a special interest in characterisation, patient-level prediction, and causal inference. The Observational Data Analysis group, led by Katia Verhamme and Daniel Prieto- Alhambra, focuses on the use of observational data to address real-world challenges in the drug, vaccine, and medical device effectiveness and safety domains, with an emphasis on combining databases from different countries. The Education group, led by Peter Moorman, participates in the training of medical doctors, clinical technicians, and medical specialists.

The research strategy and the financial status of the department are discussed and decided upon in monthly steering group meetings attended by all senior researchers. Bi-weekly management meetings involve the Head of the Department, two senior researchers, and the Director of the Health Sciences theme. These meetings address the day-to-day running of the department including the appointment of personnel and dependencies with Erasmus MC and theme-related issues. The department head has the final decision on the hiring of new staff members and promotion of personnel to permanent positions or other scientific promotions based on a proposal of research line leads. Bi-monthly department meetings are attended by all personnel and are used to communicate to the whole department and provide a forum to discuss ongoing issues.

Research strategy

The research strategy of the department is to remove the impediments to generate reliable evidence from health data and enable better health decisions and better care. Our research is aimed at scaling up the evidence generation while preserving its quality, i.e. generate reliable results for many different types of questions, for many diseases, on data from many patients (including the special populations, such as children, pregnant women, and the elderly) from many international data sources.

To scale up the evidence generation we will reduce the time from research question till results dissemination by removing the obstacles that currently prevent the utilisation of existing data: we will improve the interoperability of the data, develop standardised analytics and build and operate a large data network. We need to become more pro- active instead of responsive, which requires a strong methodological focus to answer impactful clinical questions. Therefore, an important strategy of the department is to have complementary research lines that together can improve, implement, and execute the full evidence generation process from source data to the dissemination of the results. We do not work in silos, we collaborate and have been able to produce a large body of publications, tools, and new methods as a result.

The Health Data Science (HDS) group, led by Peter Rijnbeek, is developing methods to utilise the wealth of data in electronic health records and other health databases. This includes improving the syntactic and semantic interoperability of the data through standardisation to the Observational Medical Outcomes Partnership Common Data Model (OMOP CDM). This standardisation of the data enables the development of standardised analytics and methodological research for descriptive analyses, prognostic and diagnostic prediction, and causal inference. Overall, research as done by the HDS group advances methods and best practices in areas such as development and evaluation of predictive models, operationalisation of models in observational data networks, large-scale clinical characterization, responsible and explainable AI, federated learning, heterogeneity of treatment effects, and the analysis of unstructured clinical text.

The Observational Data Analysis (ODA) group, led by Katia Verhamme, focuses on the use of the methods and the analysis of observational data to address real-world challenges in the drug safety and drug effectiveness domains, with an emphasis on multi-database studies from different countries. Research within the ODA group also addresses disease characterisation and drug utilisation which provides insight into the burden of disease over time and across different healthcare settings. Overall, research as done by the ODA group generates evidence on personalised care which is instrumental to enhance a patient’s treatment response. In addition, Daniel Prieto-Alhambra focusses in close collaboration with the HDS group on methodological as well as clinical research into the risk-benefit of vaccines and medical devices, and on the development of guidance for the generation of reliable real-world evidence, derived from multinational data sources.

Access to data is fundamental for our work and, therefore, the department maintains a large data source with medical records obtained from general practitioners in the Netherlands and drives the standardisation of the Erasmus MC electronic health record data to the OMOP CDM. Furthermore, we lead groundbreaking European projects related to unlocking health data at scale. The HDS group has taken ownership of developing, maintaining, and operating a very large data network containing a rich set of data sources from all over Europe. This research strategy has proven to be, and will be, of significant value for our methodological work, and will generate direct impact on patient care via our clinical research in disease characterisation, drug utilisation, and risk- benefit of drugs, vaccines, and devices.

We demonstrate the value of our work by applying our methods and tools to answer questions from a wide range of stakeholders, including academic partners, physicians, patients, payers, regulators and industry. We are strongly committed to open science by publishing in open access journals and making all our analytical tools available in open source and disseminating all our results.

We take the responsibility to disseminate our knowledge through the teaching of medical students and the broader stakeholder community.