Introduction: The GRADE working group has recently suggested a rigorous framework for clinical practice guidelines (CPG) addressing diagnostic tests and test strategies based on the impact of alternative approaches on patient-important outcomes. The framework mandates explicit evidence summaries, ratings of the quality of evidence, and specifying recommendations as strong or weak.
Objectives: To test the feasibility and performance of the GRADE approach, we applied this framework to well-researched issues in the diagnoses of deep venous thrombosis (DVT).
Methods: A 16-member panel with interest in thromboembolism and CPG development identified pertinent clinical questions. Our search for relevant studies included existing CPG and systematic reviews. We summarized the data in form of evidence tables and developed recommendations including, when needed, a formal consensus process.
Results and conclusions: We provide three groups of recommendations for clinicians practicing in settings with access to different types of D-dimer tests -- highly sensitive, moderately sensitive, and no availability of D-dimer. We consider the use of clinical prediction rules in guiding the diagnostic process, the potential for negative D-dimer or venous ultrasound (US) to rule out disease, and the role of follow-up testing (US following positive D-dimer result, D-dimer following negative US, and serial US) depending on the probability of DVT at the start of diagnostic process. We recommend the following: that clinicians without access to a highly or moderately sensitive D-dimer test rely on US to guide DVT diagnosis; that those with access use the highly sensitive D-dimer to determine, in patients with low or moderate probability of DVT (by the Wells rule) whether US is needed; that in patients with low pre-test probability (pre-TP) and a negative D-dimer (either highly or moderately sensitive) they follow patients without further testing; that in patients with high pre-test probability they perform a compression ultrasound without D-dimer testing.