In 2007 I started working on a PhD project on diagnosing venous thrombo-embolism in a primary care setting, at the Julius Center for Health Sciences and Primary Care – University Medical Centre of Utrecht, Utrecht University, the Netherlands. My supervisors at this project were professor K.G.M. Moons and professor A.W. Hoes. I obtained my PhD in 2011 (thesis: “Strategies in suspected venous thrombo-embolism in primary care”).
After obtaining my PhD, I continued working as an academic general practitioner (formal position: assistant professor) at the Julius Center, combining research with clinical work.
My main research interests are transferring and further improving evidence on anticoagulant related disorders to a primary care domain. The key question for general practice usually is: "Should this patient be referred for (or treated with) anticoagulant treatment, given the patients' risk of having (or developing) a thrombotic event?" Obviously, this often is a risk-benefit analysis between the risk of a thrombotic event and the risk of a bleeding event. For primary care, these questions include:
(1) Improving evidence on applying diagnostic strategies and disease detection in suspected venous thrombo-embolism – this is deep vein thrombosis and pulmonary embolism – in primary care, with a particular focus on predicton models in the elderly. Especially in an elderly population diagnosing venous thrombo-embolism implies that the general practitioner has to face the possibility of both overdiagnosis (referring to secondary care too easily when no disease is present), but also underdiagnosis, or missing a diagnosis.
(2) Prognostic models in patients with atrial fibrillation (such as CHADS2 and CHA2DS2-VASc) and venous thrombo-embolism in primary care. Patients with AF are often very old, and AF in fact can even be seen as part of a cardiovascular geriatric syndrom, including numerous comorbidities such as heart failure, myocardial infarction, renal failure and TIA / stroke. As such, I strongly believe that the focus on predicting multiple outcomes in these patients. Accordingly, this also influences the optimal choice of anticoagulants (either vitamin K antagonists or direct oral anticoagulants). In the case of venous thrombo-embolism, there is still ongoing debate on what patients are actually at increased risk of recurrent events warranting prolonged anticoagulant treatment.
With all this, I have an interest in the use of big data and individual patient data meta-analysis (IPD-MA). IPD-MA is a very exciting approach that enables a robust evaluation of subgroup effects. Subgroup effects in prediction research (both diagnosis and prognosis) mainly relate to the relative impact of co-morbidity / multi-morbidity on the performance of diagnostic tests or models. As an example, this could answer questions like: “What is the impact of chronic obstructive pulmonary disease (COPD) presence on the value of signs and symptoms in suspected heart failure?” Alternatively, IPD-MA also allows for the development of (multinomial) models that can predict multiple outcomes simultaneously.
Additionally, I am very enthusiastic on the emerging field of biomarker research. Combinations of different biomarkers can be added to diagnostic or prognostic models in order to arrive at an optimized prediction. Biomarkers as such can be truly helpful for primary care, given that they can provide added value (information) to the clinical picture of signs and symptoms. This information can fundamentally change the subsequent management of patients in primary care: this is earlier, delayed or even no referral to secondary care facilities. One of the best current examples is the added value of D-dimer testing in diagnosing VTE, yet new and promising biomarkers for endothelial dysfunction, hypercoagulability and systemic inflammation are discovered almost on a daily basis.
Potsdamer Platz, visited at ISTH conference Berlin 2017