Excluding deep vein thrombosis in primary care: validation, updating, and implementation of a diagnostic rule.

Abstract

In primary care patients suspected of deep vein thrombosis (DVT), it is a challenge to discriminate the patients with DVT from those without DVT. The risk of missing the diagnosis (which may result in a potentially lethal pulmonary embolism) and the risk of unnecessary referral and treatment with a potential harmful therapy has to be balanced by the primary care physician. Referring all patients suspected of DVT may be inefficient and results in substantial patient burden. This thesis aims to optimize the safety and cost-effectiveness of the current diagnostic process for DVT in primary care. This is done by validation of a previously developed diagnostic strategy or decision rule for DVT in the primary care setting. Furthermore, we tested the diagnostic strategy in ‘high risk’ groups, prospectively implemented it in daily primary care, assessed its potential for improvement, and the role of (point-of-care) D-dimer testing in recognizing DVT. In Chapter 2 we describe the validation of the diagnostic strategy in a new cohort of primary care patients suspected of DVT. The new data revealed that the strategy appears to be a safe diagnostic tool for excluding DVT in patients suspected of DVT in primary care, leading to a substantial reduction of unnecessary patient referrals to secondary care and consequently of patient burden. This conclusion was confirmed in Chapter 3, where we found that the strategy can be applied to all types of primary care patients suspected of DVT, regardless of age, gender, and history of venous thromboembolism. In a large management study (n=1028), over 300 Dutch general practitioners actually applied the strategy to decide whether to refer patients suspected of DVT to secondary care. It was concluded that implementing the diagnostic rule in daily primary care practice reduces the number of patient referrals for ultrasound measurements by almost 50%, at the cost of an acceptably low risk of subsequent venous thromboembolic events in the non-referred patients (1.4%; Chapter 4). The data also showed that (1) the originally developed strategy does not need to be adjusted or modified to current and local circumstances as it still showed optimal safety and efficiency for excluding DVT in primary care (Chapter 5), and (2) the use of the rule is cost-effective as compared to hospital based strategies (Chapter 6). The proposed diagnostic strategy makes use of a so-called D-dimer test. D-dimers are degradation products of cross-linked fibrin generated during fibrinolysis, and can therefore be used as an indirect measure of thrombus formation. We determined the performance of two often used laboratory D-dimer assays (Chapter 7). Finally, we compared the accuracy and user-friendliness of five different so-called point-of-care D-dimer assays, and found that all had reasonable or good accuracy, but just a few are suitable for primary care practice (Chapter 8). This thesis concludes with a review discussing important methodological aspects of diagnostic research in general, illustrated with the diagnostic strategy for DVT diagnosis in primary care.