Development and Validation of Risk Prediction Algorithm (QThrombosis) to Estimate Future Risk of Venous Thromboembolism: Prospective Cohort Study

Venous thromboembolism (VTE) is a major cause of morbidity and mortality that is potentially preventable with anticoagulation and other interventions. Although a number of clinical decision rules have been presented in the literature, they are generally designed for use in evaluating patients with presentations concerning for acute VTE. This article, in contrast, is targeted at primary care providers and aimed to predict absolute risk of VTE at 1 and 5 years in asymptomatic patients. The authors used a large research database of data routinely collected from general practices in the United Kingdom to create a prospective cohort of patients aged 25–84 years with no history of VTE, not on oral anticoagulants, and without pregnancy in the preceding year. They randomly assigned patients to either a derivation cohort or a validation cohort. They then identified a broad group of risk factors for VTE based on previously published studies including age, body mass index, tobacco use, congestive heart failure, chronic renal failure, use of hormone replacement therapy, cancer, and others. The clinical outcome was diagnosis of VTE, including deep venous thrombosis or pulmonary embolism. After extensive analysis, the authors developed an algorithm (available as a calculator online at: http://www.qthrombosis.org) to predict the risk of developing VTE in asymptomatic patients, and used their validation cohort to test their model. The predicted and observed risk correlated well, but with a positive predictive value of only 2.0% for diagnosis of new VTE in 5 years among patients in the top decile of predicted risk.