Use of a Clinical Prediction Score in Patients with Suspected Deep Venous Thrombosis: Two Steps Forward, One Step Back?

Since antiquity, a measure of a physician's skill and experience has been the ability to predict at the bedside whether a disease is present or absent. In the current era, standardized clinical prediction scores (also known as clinical decision rules) can level the playing field so that physicians, irrespective of experience, can obtain an estimate of disease likelihood (1, 2). Clinical prediction scores can guide clinical management (for example, Ottawa ankle/knee rules), estimate adverse outcome risk (for example, Detsky score), and determine prognosis (for example, Acute Physiology and Chronic Health Evaluation [APACHE] score). The increasing availability of electronic medical records facilitates the use of clinical prediction scores: After a physician enters the chief complaint or clinical scenario, the computer can display the appropriate clinical prediction score, calculate the score, display the corresponding pretest probability, and even calculate the post-test probabilities for each diagnostic test under consideration (3).

The clinician can choose from several clinical prediction tools to estimate the likelihood of lower-limb deep venous thrombosis (DVT) (4-6). Since 70% to 80% of patients with suspected DVT do not actually have it (7), the main objective of a prediction score is to identify patients at low risk for DVT, in whom diagnostic imaging (venous ultrasonography) can be safely deferred or avoided. The best-known prediction score for DVT is that of Wells and associates, which has undergone several iterations since its first publication in 1995 (8-10). The authors simplified the original 12-point version (8) into a 9-point score that classified patients as having a low (≤0 point), moderate (1 to 2 points), or high (≥3 points) risk for DVT, but was not applicable to patients with a history of DVT (9). The most recent, 10-point, version applies to all patients with suspected DVT—including those with previous …